c< C I I
< < t<
» CC cc CCC C< ccC cc
CT CC r< cc c<
0 fC c 0 (C c 0 (C(
0 C C
% cc .
^ CC c
*% C c v
CC
•<( , cCC ,
' rrccc t
, ccc - <arc m
- r CCC
< C fr
( c c
•r * i
* c c
< c r
■a cc
c <
■Mi'.
)\ $c ( cccc
<cc ^
CCC Q ((
, C i C (d
- .' <(C.
A Cf( ( i cc S^ C < c CC(
C: C?rr ! C^ ' 0 ^-- %^$f c i C - r
>' c C C.<H£ C C
;c ^::-ciE3?r^Z33r:-cz^?r2cv^7
J SCIENTIFIC LIBRARY,
•|UnitedStatesPatentOfficeJ
1(3'
. - C< ! «C < C CC '
CT(ccc
CE C C CC
C c cacl
c, c
C C . c < . fc c C CCC c
i 9v S c
k C- C^ c c <. f
f C^ C C (. r
r rc^c: c c c ' ,<>-. c c
c C ^
|
c |
cc |
|
( |
c |
|
c |
cc |
|
V |
cc |
|
c |
u |
|
C |
« |
|
c |
ft |
|
C |
Cf |
|
Cr |
■ c c
■ <
■ c
• • cc <"■ - <
C c.
C« i •0 ft. c X(cc< <
CCC Cv '
&y« c
^Ccc u
CCCC c CCCC (c
CCCOa:
c c CC f ( < CC c^'c'
CC Sm~ r r
V-C Cc
5 Sec
C cc
izC cc
c'C-
c c
CCsC
cC- c
c < c ,• i1-
c ^~ c.< , c ■- c c 1
c f,V Cc C(
S ( c<
rC c C ^
cccc ■ ccc.
' c c '
.CCC c
(CCc
c ccr c
>> CCC c C
CC'€c< cc or
CC : C: C ( ( ^> f
c < c c-x o
CC > r ■ C (<^ C
•CI C a"r £K c^((->. < cc • t <c
l<% CCcC
=(' cc . CCC •< ' (C c a
c - « cc - « <
< CC -< C7 c (
< CTc ' iCJ1 ( ( cc c , cc: (
=5' cc , CCC
V,S c<^cc
l( CC c - C C ' cc < c <c <
ff ( <c c ; C <G < c CJ
%■ " ■ v c c
^- Cc c c
-% ( CC . CC <.*->-.<<<_ £
«$
C ( <
CCC*
c -(5c- c
- c <«: X ■ C
3: «
4fyC Ch -CCCl
0 &&f
ccc ccwccc k ^ <X cc cc <t
S£--cc c ' <c <
> >-> C C CC -r
r £ & £ f-CC ft
v ccc C( cc cc c *r-rr cc CC cc «r cc V 5£ 1
«C^Cc£/v«fc<xV# 1
< c c c;
c c C c
ccccc
CCCCC
C C Cv ■ ■«■ C < c C<
CCc
«t ^C: c (C<( <rcr«t^ < ceo- <c<o
<c cc or c c jCC cc <3C Q '
CC XI. ^C (■'■:
Ct <^C « < C
cc cc <:< c (M. c
* cc <cc -
•. <: c. cc c
• c c<
CI IL «
-aP '-c" C. C "cc ^
C«tC^(,t Cv<C <C m
< <cc cc k*« c y
-c c<rC.cc^--,C"fr ,
1c ;
4%cf*V
c c f <: ff V c
«7C< a <
^CC RiT
- <C art. ^ 5 ^cf
SCCc ' c <
C -<C <<CC
< < cm:
< CO C (g <
cr«; <ftc
C C r^ ^ ??
cc occccc
C CC CCCCC
C a c r CTcc
: <xc c <- cc
cc cc cc
«c cc c< ccc cccc c
CCC c c c c < c c c «
, ,- <- c ®C-Tc
cc c r ^cc «!i
,^<< fee cc CCCJCC <
^ -cc cc ccaO c
-V-C ; cc c c . <i< cc<: <
>f^c <x CC cC CC <
J&CC CX CCCCC
^ cC C < CC'-CC c
MCi ' cc ~C c <c: -CC C
■ <■ < <~~ < c < . ■'« <" <•
rc <<'-<c C'
; < Cc c < <r,
^""<: . cc'4 c cCi
i; c cc <3r«C( Cv
r ■ ' _ C" <£ ^< '<" < C'l
,,.</ <^ c ccc «rc
? :&<< c< cc Cii.^
(< O % c cc </
cC c< CC c^ <
" «• <c < < <: Cv«
«:; 0")t C> f< <-r«
<ccccc<e
<: < cc cc CC
f C c5
CC'CC^ <cccccc cr
cccc m <~
< c c < c .1
<c: CC cc CCCCC C imrj <-<■. ■« C <Sf*. CCCCC c r«r<
CCCC CC CCCCC cccc cc c < <-<-<■ cc<:c c «c:<
ccCC cCCC CCCCC <<:
C3TSCC C < CCCC CC
C:<:c ccc ccdfc:r '-C C ;
c CC ClCC <0CC :CC^
<: C<
<U <g c <1 <c c
^: <«t .. c € ^ <@C C c
c <«r:c cc
c cc cc <c c- CC <Xi C
c/ cc cCC= c f< c:cc c < C Cf
c ^,< <rr .
ct:cC ccf 1 c C CCCC c CCCCC
c C> ( c CC (
' cccc
or <
#3?
cc c
?5 c C
C <c c
<- C cCJ <2T ^C*< C'CCc CcCCC cC ccccc <c .jp- ccc tec <<<( c
^ C< < CC r c CCCCC('( ( C. ccc Cc <$ Cc c
^cc ( (, c <--r ,, , CC C c CC CCc - <« </ - C C C
"^ V C^ <^ ( f r
^Scv-^^S^i
>/V- ^ «' c c .- ^rr( <^< c c
SFc r j^s < c ■
KC C.CC' -v C< C. «c> C c 4
c «.« C. <■
... CC
cc
m
aCC < CC
-
< cc <<r
JOURNAL
OF THE
Boston. Cleveland. Minneapolis. St. Louis.
Montana. St. Paul. Denver.
Virginia. Pacific Coast.
CONTENTS AND INDEX.
VOLUME XVII. July to December, 1896.
H 9 9 9 4
PUBLISHED BY
THE BOARD OF MANAGERS OF THE ASSOCIATION OF ENGINEERING SOCIETIES.
John C. Trautwine, Jr., Secretary, 257 S. Fourth Street, Philadelphia.
h
/
JA,^8 1897
CONTENTS.
VOL. XVII, July-December, 1896.
For alphabetical index, see page v.
No. 1. JULY.
PAGE
Principles Governing the Design of Foundations for Tall Buildings. Randell
Hunt 1
Discussion. Mr. Wagoner, Mr. Hunt, G. W. Percy, Prof. Marx, Prof.
Wing, Prof. Soule, Mr. Leonard, Mr. Curtis 19
Locomotive Counterbalancing. G. R. Henderson 27
Riveted Joints. Joseph R. Worcester 33
Discussion. Edward S. Shaiv, John C. Moses, James E. Howard, Gaetano
Lanza, J. P. Snow, Joseph R. Worcester 45
A Low Crib Dam Across Eock River. ./. W. Woermann . . 54
Proceedings.
No. 2. AUGUST.
Experiments on Vitrified Paving Brick. F. F. Harrington 65
The Conditions Necessary for Equality of Velocity in Particles Settling through
Liquids. Luther Wagoner 73
Proceedings.
No. 3. SEPTEMBER.
Water Supply and Sewerage as Affected by the Lower Vegetable Organisms.
By the late Clarence O.Arey, C.E., M.D 77
The Testing of Coals. Arthur Window 84
Methods and Results of Stadia Surveying. F. B. Maltby 90
Discussion. J. L. Van Ornum, B. H. Colby, J. A. Ockerson 107
Proceedings.
No. 4. OCTOBER.
The Historical Development of Stone Bridges. Prof. George F. Sivain • . ■ 117
Proceedings.
iv ASSOCIATION OF ENGINEERING SOCIETIES.
No. 5. NOVEMBER.
PAGK
Solar Work in Land Surveying. J. D. Varney 145
Boiler Efficiency, Capacity, and Smokelessness. with Low-Grade Fuels. M '/'/-
liam H. Bryan 1 57
Gas Producers, and the Mechanical Handling of Fuel for Same. 0. L. Saun- ders 169
Proceedings.
No. 6. DECEMBER.
Refrigeration. As Applied to Dwellings, Hotels, Hospitals, Business Houses
and Public Institutions. Alfred Siebert 1T5
Recent Practice in Railroad Signalling. George W. Blodgett 182
Discussion. Prof. C. Frank Allen, E. K. Turner, George F. Sampson,
Mr, Turner 194
The Galveston Harbor Works. W. J. Sherman 197
Structural Strength of Ships and Improved Arrangements for Repairing them
Without Diminution of their Strength. Joseph B. Oldham 209
A Few Points of Engineering Interest Observed on a Short Trip Abroad.
Francis W. Blackford 221
Proceedings.
INDEX.
VOL. XVII, July-December, 1896.
The six numbers were dated as follows:
No. 1, July. No. 3, September. No. 5, November.
No. 2, August. No. 4, October. No. 6, December.
Abbreviations. — P = Paper ; = D Discussion; I Illustrated. Names of authors of papers, etc., are printed in italics.
A PAGE
Few Points of Engineering Interest Observed on a Short Trip Abroad.
Francis W. Blackford P., I., Dec, 221
Arty, Clarence 0. Water Supply and Sewerage as Affected by the Lower
Vegetable Organisms P., Sept., 77
Arey, Clarence O. A Memoir Procs., Oct., 15
Dlackford, Francis W. A Few Points of Engineering Interest Observed on
a Short Trip Abroad ... P., I., Dec, 221
Blodgett, Geo. W. Eecent Practice in Eailroad Signalling. . P., I., Dec, 182
Boiler Efficiency, Capacity, and Smokelessness, with Low-Grade Fuels. Wm.
H. Bryan P., Nov., 157
Brick. Experiments on Vitrified Paving . F. F. Harrington. P., I., Aug., 65
Bridges. Historical Development of Stone . Prof. Geo. F. Swain.
P., I., Oct. 117
Bryan, Wm. H. Boiler Efficiency, Capacity and Smokelessness, with Low- Grade Fuels P., Nov., 157
Buildings. Principles Governing the Design of Foundations for Tall ■.
Randell Hunt P., I., July, 1
P '
VJoals. Testing of . Arthur Winslow P., Sept., 84
Conditions Necessary for Equality of Velocity in Particles Settling through
Liquids. Luther Wagoner P., I., Aug, 73
Counterbalancing. Locomotive . G. R. Henderson. . . . P., I., July, 27
Crib Dam. A Low Across Rock River. /. W. Woermann. P., I., July, 54
J_yam. A Low Crib Across Rock River. J. W. Woermann. P., I., July, 54
Hifficiency. Boiler , Capacity, and Smokelessness, with Low-Grade
Fuels. Wm. H. Bryan P., Nov., 157
Engineering. A Few Points of Interest Observed on a Short Trip
Abroad. Francis W. Blackford • P., L, Dec, 'I'll
Experiments on Vitrified Paving Brick. F. F. Harrington. ■ • P., I., Aug., 65
(v)
vi ASSOCIATION OF ENGINEERING SOCIETIES.
JF oundations for Tall Buildings. Principles Governing the Design of .
Randell Hunt P., I., July. 1
Fuel. Gas Producers, and the Mechanical Handling of for Same. C.
L. Saunders P., Nov., 169
Fuels. Boiler Efficiency, Capacity, and Smokelessness, with Low-Grade .
Wm. H. Bryan P., Nov., 157
Vjalveston Harbor Works. W. J. Sherman P. I., Dec, 197
Gas Producers, and the Mechanical Handling of Fuel for Same. C. L.
Saunders P., Nov., 169
H,
Larbor Works. Galveston . W. J. Sherman P., I., Dec, 197
Harrington, F. F. Experiments on Vitrified Paving Brick. . . P., I., Aug., 65
Henderson, G. R. Locomotive Counterbalancing P., I., July, 27
Historical Development of Stone Bridges. Prof. Geo. F. Swain. . P., I., Oct., 117
Hollaway, Josephus, Flavius. A Memoir . Procs., I., Oct., 15
Hint, Randell. Principles Governing the Design of Foundations for Tall
Buildings P., I., July, 1
Joints. Riveted . Joseph R. Worcester P., I., July, 33
Jjand Surveying. Solar Work in . J. D. Varney . . . . P., I., Nov., 145
Liquids. The Conditions Necessary for Equality of Velocity in Particles
Settling through . Luther Wagoner P., I., Aug., 73
Locomotive Counterbalancing. G.R.Henderson P., I., July, 27
Low Crib Dam Across Rock River. J. W. Wocrmann ... P., I., July, 54
Low-Grade Fuels. Boiler Efficiency, Capacity and Smokelessness, with .
William H. Bryan P., Nov., 157
iYlrt?%, F. B. Methods and Results of Stadia Surveying. . . . P., I., Sept., 90
Methods and Results of Stadia Surveying. F. B. Maltby. . . . P., I., Sept., 90
VJldham, Joseph R. Structural Strength of Ships and Improved Arrange- ments for Repairing them without Diminution of their Strength.
P., I., Dec, 209
Organisms. 'Vgater Supply and Sewerage as Affected by the Lower Vegetable
. Clarence 0. Arey, C.E., M.D P., Sept., 77
laving Brick. Experiments on Vitrified . F. F. Harrington.
P., I., Aug., 65 Principles Governing the Design of Foundations for Tall Buildings.
Randell Hunt P., L, July, 1
Producers. Gas , and the Mechanical Handling of Fuel for Same.
C. L. Saunders P., Nov., 169
Xtailroad Signalling. Recent Practice in . George W. Blodgett.
P., I., Dec, 182
Recent Practice in Railroad Signalling. Geo. W. Blodgett. . . .P., I., Dec, 182
Refrigeration. Albert Siebert P., Dec, 175
Riveted Joints. Joseph R. Worcester. P., I., July, 33
Rock River. A Low Crib Dam Across . J. W. Wocrmann. P., I., July, 54
INDEX.
s
aunders, C. L. Gas Producers, and the Mechanical Handling of Fuel for
Same P., Nov., 169
Sewerage. Water Supply and , as Affected by the Lower Vegetable
Organisms. Clarence 0. Arey, C. E., M.D P., Sept., 77
Sherman, W. J. Galveston Harbor Works P., I., Dec, 197
Ships. Structural Strength of , and Improved Arrangements for Repair- ing them without Diminution of their Strength. Joseph R. Oldham,
P., I., Dec, 209
Slebert, Albert. Refrigeration P., Dec, 175
Signalling. Recent Practice in Railroad . Geo. W. Blodgett. P., I., Dec, 182
Smokelessness. Boiler Efficiency, Capacity and , with Low-Grade Fuels.
William H. Bryan P., Nov., 157
Solar Work in Land Surveying. J. D. Vamey P., I., Nov., 145
Stadia Surveying. Methods and Results of . F. B. Mallby. . P., I., Sept., 90
Stone Bridges. Historical Development of . Prof. George F. Swain.
P., I., Oct., 117 Strength of Ships. Structural and Improved Arrangements for Repair- ing them without Diminution of their Strength. Joseph B. Oldham.
P., I., Dec, 209
Surveying. Methods and Results of Stadia . F. B. Maltby. . P., 1., Sept., 90
Surveying. Solar Work in Land . J. D. Vamey P., I., Nov., 145
Swain, George F. Historical Development of Stone Bridges . . .P., I., Oct., 117
T
all Buildings. Principles Governing the Design of Foundations for .
Randell Hunt P. I., July, 1
Testing of Coals. Arthur Winslow P., Sept., 84
V
arney, J. D. Solar Work in Land Surveying ?•>•"■•> Nov., 145
Vegetable Organisms. Water Supply and Sewerage as Affected by the Lower
. Clarence 0. Arey, C.E., M.D P., Sept., 77
Velocity. The Conditions Necessary for Equality of in Particles Set- tling through Liquids. Luther Wagoner '. . . P., I., Aug., 73
Vitrified Paving Brick. Experiments on . F. F. Harrington. P., I., Aug., 63
w
I V agoner, Luther. The Conditions Necessary for Equality of Velocity in
Particles Settling through Liquids . P., I., Aug., 73
Water Supply and Sewerage as Affected by the Lower Vegetable Organisms.
Clarence 0. Arey, C.E., M.D P., Sept., 77
Winslow, Arthur. Testing of Coals P., Sept., 84
Woermann, J. W. A Low Crib Dam Across Rock River P., I., July, 54
Worcester, Joseph R. Riveted Joints . . P., I., July, 33
Editors reprinting articles irom this journal are requested to credit both the Jouknal and the Society before which such articles were
Association
OF
Engineering Societies.
Organized 1881.
Vol. XVII. JULY, 1896. No- 1.
This Association is not responsible for the subject-matter contributed by any Society or for the state- ments or opinions of members of the Societies.
PRINCIPLES GOVERNING THE DESIGN OF FOUNDA- TIONS FOR TALL BUILDINGS.
By Randell Hunt, Member of the Technical Society of the Pacific Coast.
[Read before the Society, April 3, 1896.*] Copyright by Randell Hunt, 1896.
The modern tall building is in many respects a good personification of our national character. Utility first and foremost, adaptability to immediate surroundings, and a capacity for acquiring the " almighty dollar " with ease and certainty.
It is " all things to all men " — convenient to commerce, comfortable and healthy. Professional and business life within its walls is not wor- ried by the little annoyances which in ordinary surroundings are apt .to detract from the more serious pursuit of material happiness.
Originated in Chicago. — It is not surprising then that Chicago, most typical of all the cities of our national characteristics, should have been the place where these wonderful structures have received their great- est development. Nor was it cause for astonishment when it was discov- ered that the soft, muddy subsoil of this city precluded the use of the recognized methods of founding ordinary buildings; that a new and dis- tinctly original solution was applied to give a safe and permanent sup- port to the towering structures.
It can hardly be said that thjg result was achieved at once for any one building, for the true underlying principles of all foundation work were rather forced upon the builders and designers by a series of mis- haps, which demonstrated in a most practical manner that scientific prin- ciples are the only true and safe ones to follow.
* Manuscript received May 4, 1896. — Secretary, Ass'n of Eng. Socs.
2 ASSOCIATION OF ENGINEERING SOCIETIES.
• Necessity of providing for settlement under certain conditions. — If a particular soil will support but a limited weight with- out compression arid settlement, then one must make suitable provision for such change in the position of the base, which of necessity must occur in any structure founded upon it, and exerting a pressure beyond that amount it will carry without any yielding.
In the case of a building certain principles of construction have been recognized as necessary — when it is founded upon a compressible earth — to prevent unsightly cracks and sometimes dangerous results from occurring.
To keep the area of the base of the building so large that the press- ures transmitted to the earth will cause no settlement whatever is often regarded as impracticable, and in many localities it has been found that after a certain limited compression of the soil has taken place no further settlement need be apprehended.
This at least is, and has been, the argument in some places where many of the largest and most costly of our modern buildings are being erected.
Method of independent piers. — The method of founding large buildings upon independent piers is one now so common and so well understood by engineers and architects as to hardly call for any par- ticular explanation here. It is simply a recognition of the well-knjwn fact that if a beam is acted upon by two forces at or near its ends it tends to assume a curved form due to the unequal moments of the pressures transmitted to the beam from the reactions of the ground upon which it rests.
ffl/?J//M'l""M'"""'»M>Z.
Fig. 1
If this beam, as in the case of masonry connecting walls, is weak- ened by numerous openings, as windows or doors, one above the other, or by any other means, so as not to have sufficient strength in itself to transmit the pressures on its ends to the ground beneath throughout its entire length without deflection from reactions, then it will bend or crack on the lines of least resistance.
Uneven foundation pressures immaterial on rock. — In solid soils, or upon rock, or upon any perfectly unyielding foundation, it is immaterial, of course, how unevenly the pressures may be transmitted to it, and unless there is overloading to the crushing point, the princi- ple of independent piers is of no practical use, excepting from an eco- nomical and perhaps convenient point of view.
FOUNDATIONS FOE TALL BUILDINGS. 3
Damage from settlement. — The upward reaction between the points of great pressure in a compressible soil usually results in a build- ing being damaged by cracks extending from the base to the roof and following the line of windows from story to story.
Cooper Institute. — To illustrate briefly this common cause of failure in buildings, I have chosen as an example of the results which will happen from neglect of the principles iust explained the Cooper Institute of New York.
This building was founded in 1853 upon piers carried down 22 feet below the sidewalk, and resting upon a continuous masonry footing course 1 foot 4 inches thick. In 1885 the settlement and consequent cracking of the walls had become so great that it was considered dan- gerous, and extensive reconstruction of the foundations was entered upon.
An examination of the building shows at a glance that the chief weight of the walls is carried down the piers " A " and " B " to the foundations, and that the pier " C " between these two carries but a small weight in comparison. From bad proportioning of the footings of these main piers only a limited area was capable of receiving the full pressures, and as much as six tons per square foot was thrown on the foundation soil. Failure occurred by the continuous stone footing cracking across, and the main piers were shoved down into the over- loaded earth, tipping up the outer edges of the stones directly under them, which were not continuous but jointed in the center. At the same time the intermediate pier, with its light loading, remained without, settlement, and as the piers on either side sank the upward reaction was sufficient to cause a " vertical fraction at each side of every window from the third story down."
If all the footings of the piers had been properly proportioned, so as to have exerted a uniform pressure per square foot on the soil, and this had been well within its safe supporting power, there would have been no accident. But as this is at best a difficult thing to always secure, viz., a soil which does not compress at least a small amount, even with light loading, other methods of supporting the intermediate pier could have been adopted if unequal settlement was feared.
Characteristic construction of modern high buildings. — The modern high building consists, in most recent examples, of a steel skeleton frame from the foundations to the roof, in itself carrying all the weight, from story to story, of the masonry walls, partitions and floors. These walls are reduced to the least dimensions — to sustain them- selves only — being merely " curtain walls " in most cases.
The basements are usually required to have plenty of light and as much or more openings than the stories above. Therefore the method
ASSOCIATION OF ENGINEERING SOCIETIES.
fri.UNE orcKOUNDFi
^vr^r^
fc3
M MK M
±l
:l j:
J- , I , i , i , r
I 1
I I
1,1 , 1,1
£§,
Fig. 2. — Cooper Institute, New York.
FOUNDATIONS FOR TALL BUILDINGS. 5
of placing each column, or pair of columns, upon its own base, sepa- rated and independent from the next set, becomes a very convenient method of economizing space for light and available area for business purposes.
Masonry piers under columns. — If there is plenty of good, hard ground underneath the level of the basement floor, the area of base necessary to be placed under each steel column could be easily attained by excavating to sufficient depth, so that a masonry pedestal could be built, with only the usual safe slopes or offsets from the bedplate to the ground.
If the foundation soil was very hard and incompressible — capable of carrying large loads, say of five to six tons per square foot — then the ground area required for the foundation piers would be small, and consequently they need be of but limited depth in the ground. But if, on the other hand, the supporting power of the soil is very small and liable to compression, reducing the unit of safe weight which can be placed upon it to from one to two tons per square foot, then it becomes necessary to have a foundation base of large area. This can only be accomplished by making the masonry pier of considerable depth down in the ground, so as not to exceed the proper safe slope or projection of the footing courses.
Advantages and disadvantages op this construction. — There is no objection which can be urged to this method of securing a foundation if the soil is a suitable one for the practical carrying out of the construction.
In fact it has many distinct advantages over the usual shallow steel beam and concrete foundation.
Unfortunately the conditions of the foundation sites in many of our large cities, as well as extraneous conditions, such as neighboring buildings, ground, water, etc., make this method at times more or less impracticable.
Founding upon a bed of firm soil overlying softer ma- terial.— In Chicago, for instance, the peculiarities of the ground for- mation make it entirely out of the question to found in the above manner, for almost without exception the records of the borings of the foundation sites of the large buildings show a depth of only from twelve to fifteen feet of a very moderately firm soil overlying a much softer clay subsoil of considerable depth.
The use of this top crust of firm ground, without in any way cut- ting down into it, has been the object sought by most of the architects and engineers in founding their large structures.
Origin of concrete and steel beam foundations. — Hence, as is usually the case, necessity became the mother of invention, and
6 ASSOCIATION OF ENGINEERING SOCIETIES.
steel beams in shallow piles, placed in tiers at right angles to one another, took the place of more bulky masonry, and attained the same purpose without requiring' but very limited excavations.
Adopted by force of circumstances for peculiar conditions, yet the method of construction has shown itself to be more or less well adapted for other localities where these surroundings may not exist. .
A great change in the relative cost of materials has also been another factor in the use of these foundations. Steel beams are almost 100 per ceut. cheaper inmost to-day than fifteen years ago, and, in addi- tion, their properties and use'are better understood.
Mr. Bauman, an engineer of Chicago, called particular attention some years ago to the method of independent piers for foundations in compressible soils, and announced it as a scientific principle of construc- tion, which was a necessity in soils such as in Chicago. He simply explained a very ancient foundation method in use since the times of of the. Goths. It is certain, however, that the practice has followed on the lines indicated by him with more or less success.
Calculation of weights supported by independent piers. — If the lines of pressures on the ground area under one pier overlap those of a contiguous one it becomes necessary to make a single base for both piers. In this way there has developed a growing tendency in the later examples of large buildings to lessen the number of independent foundations by grouping several columns upon one base.
Undoubtedly independent piers, if very carefully proportioned to the exact weights carried upon their bases, as well as taking note of the friction upon their perimeters, offer a proper method of securing a foun- dation for tall buildings.
There are, however, certain principles not so easy to calculate in definite terms, which introduce more or less difficulty in arriving at a correct area to give the bases of columns transferring very unequal weights.
It is a known fact that large areas of soft soil will not support the same weight per unit of surface as more limited areas of the same soil.
It becomes necessary in designing the bearing area of the base of the foundations to take into consideration this fact if one is to feel per- fectly certain of an equal settlement of all the piers.
An equal allotment of weight to be supported per square foot of ground area, under small piers aswell as under much larger ones, will certainly result in an unevenness of settlement due to this principle just enunciated. /
Chicago practice is the result of experiment. — Before the era of extremely tall buildings — some twelve years ago — various methods of founding in that peculiar soil had been tried. The more common
FOUNDATIONS FOR TALL BUILDINGS. 7
practice of to-day is largely the result of actual experiment and success- ful precedent. Still one cannot but be impressed with the fact that there is not a uniformity of opinion that correct methods have yet been adopted. One is startled to know that buildings which have cost one and a quarter to one and a half millions of dollars, are expected to, and do, settle five to six inches duriug the first year or two of construction.*
The floating of such buildings upon a crust of soil but twelve to fif- teen feet thick, overlying a softer watery clay, shows a reliance upon future stability which is sublime, and entirely Chicagoan in its assurance.
Investigation of the direction of the ground pressures. — If reliance is placed upon the strength of a top layer of soil which overlies a weaker material, then an investigation should be made to deter- mine if the foundation pressures are distributed over a sufficient area of the lower soil to be within its safe bearing capacity.
It was at one time thought that the angle which the direction of the pressures through the ground made with the vertical was equal to the natural angle of repose of the material.
Experiments — in sand particularly — would seem to indicate that this is not correct, and that the angle is really about one-half the slope angle of the earth. t
Let the Fig. 3 represent the foundation of a column in a building resting upon sand. The angle of repose of ordinary moist sand is about 40°, in which case the angle of pressures becomes 20°; therefore the dimensions of the ground area which receives the original foundation pressures can be easily investigated at any depth.
Supposing a strata of soft clay, or other material, to exist, then by drawing to sufficiently large scale the figure herewith shown, one can easily measure off the dimensions receiving the pressures and see if the load per unit of area is within the safe pressures which is permissible on soft clay.
Let (P = the angle made by the direction of the pressures, and this is equal to one-half the natural angle of repose of the soil.
Then the width of base b', which receives the foundation pressures at a depth h, is :
V = b + 2 h tan 0,
* The Monadnock Building, nearly 200 feet high, with a pressure on the founda- tion soil of 3,750 pounds per square foot, settled very uniformly five inches.
Old Colony Building, 215 feet high, with 3,220 pounds per square foot on foun- dations, in nine months settled from 4T35 to 5£ inches. Both these buildings are con- sidered successful examples of Chicago practice. — Engineering News, October 13, 1892.
t Handb. d. Ingen. Wissensch., Bd. II, p. 196. Wochenbl. d. Ver. deutsch. In- gen., 1882, p. 53. Experiments by Brenneke and Forchheimer.
ASSOCIATION OF ENGINEERING SOCIETIES.
:'Ag:^V-.vV/^::MS « j'
AUG fc 8 1898
FOUNDATIONS FOR TALI^ttyLDINGS.
in which b = the width of base of the foundation pier at the surface. At a depth of h + h' the width of base 62 is equal to
b2 = b + 2h tan 4> -f- 2h! tan $', in which 4>' represents a change in the angle of pressure direction due to a water-soaked material, which has a much smaller angle of repose than the same sand above the water level.
If the area of the soft clay at this depth is not sufficient to give a value per unit of surface less than its safe supporting power, then the dimensions of the base of the column must be increased.
In addition, one must take into consideration the weight of the earth itself which lies between the base of the foundation and the strata of the soil considered.
Narrow limits of pressures in water-bearing ground. — The natural angle of repose of water-soaked soil becomes very much less than when dry or simply moist.
The very common condition in the foundation of a building is that the level of the ground water is within a few feet of the base ; therefore, although the angle of the pressures of the foundation may commence, when the ground is dry, to extend itself over considerable area, yet, as soon as the water level is reached, this angle becomes very much less.
For sand, the natural slope, when water-saturated, is about 24°, and this would limit the pressure directions to only 12°, as shown in Fig. 3.
In reality, the natural slope of repose of the earth under a founda- tion base is not capable of too close an analysis. It is undoubtedly affected by depth, degree of moisture and the lack of uniformity in the character of the soil. Yet an investigation can usually be carried out on the lines which have been suggested, so as to leave no doubt as to the limiting thickness which a top crust of soil ought to have, to properly distribute any weight placed upon it, to a weaker subsoil.
Earthquake effects. — Vibrations are more or less injurious to all structures, and good construction seeks to reduce them to a mini- mum. The relationship of the foundation of a building to its super- structure is of much importance in this respect, and in any country subject to earthquake shocks due regard should be paid to avoiding their effects.
Careful investigation in Japan, extending over many years' observa- tions, as well as in other earthquake countries, shows that unless the locality is situated directly over the center of the disturbance there is seldom any damage to well-constructed substructure work.
In certain portions of South America so well is this understood that the lower stories of the buildings, while of heavy masonry, or adobe, will have pliable basket-like construction in the upper stories.
The movement of an earthquake is vibratory, and in those parts of the
10 ASSOCIATION OF ENGINEERING SOCIETIES.
United State3 in which they have occurred the amplitude of the vibra- tions is comparatively small, so much so that well-constructed masonry in considerable mass in the ground is capable of taking up the oscilla- tions without damage. '
The greatest intensity of the shock and the amplitude of move- ment is at the beginning, rapidly diminishiug during its duration, much like the vibration of a tuning fork, or like sound-waves.
Therefore the oscillations of a foundation of a structure will be the same, and no more, than those of the ground in which it is built. The building above, however, may become subject to a cumulative vibration derived from the oscillating base, and this is the usual cause of disaster.
This is due to the fact that the vibrational period of the building itself, caused by the first shock, becomes a multiple of the earth vibra- tions, and the amplitude of them is thus increased, though in reality those of the earth are diminishing.
Such oscillations might throw additional strain on the foundations of a building, and from the existence of different vibrational periods in one part of a structure over another — if the building rested upon inde pendent piers — one or more of them might be called upon to carry greater pressure to the soil than others.
Steel buildings safe in earthquakes. — Before the days of recent steel skeleton structures, independent authorities in different por- tions of world, who were seeking a proper design for buildings which should be earthquake proof, recommended iron frame construction.
The idea being the building should be tied and braced together in all parts*; that they should be light in weight as well as strong ; that any vibration should be as a whole, and not greater in one part than in another.
The superstructure of the modern steel frame building complies with all these conditions, although no thought of earthquakes had any- thing to do with the general growth of the design.
Not so with the foundations however.
Movement of buildings during earthquakes. — The moment of inertia of a heavy overhanging roof, or top of a tall building, seeks to keep it at rest, and if the base is set in motion by a sudden shock, great forces occur tending to cause rupture between the foundation and the superstructure.
The building being strong enough to resist rupture at this moment of time, the great roof now moves forward, and the energy of its move- ment may be increased by coinciding with the vibration of the ground. It resists any sudden checking of its motion only by causing again strains to occur of the greatest magnitude.
How the foundations should be designed. — It appears to me that in a building of great height, in which, of course, the amplitude
FOUNDATIONS FOR TALL BUILDINGS. 11
of the vibrations at the top is extremely liable to be greater than at the base, that this base in all its integral parts should move as a whole, and no part of the foundation should be able to transmit an unequal move- ment to the superstructure.
Considerable mass and weight in the foundation will in itself take up and destroy part of the movement of the earth before transmitting it to a building upon it.
Shallow foundations not proper for earthquake vibra- tions.— The method of independent piers, as now built in common practice, makes use of small mass and weight, and would transfer any earth movement in the quickest and most direct manner to the steel frame resting upon them.
Prof. Milne found at the college at Tokio, Japan, a difference in the intensities of the earth movements during an earthquake, even over a very limited plot of ground. For this reason it is certainly better to make a single foundation for a building, or one which must move as a whole. ' It has also been determined that there is less vibration at a depth of some feet in the ground than on the immediate surface.
Of course, in buildings covering very large areas of ground this be- comes well nigh impracticable, or unduly expensive, but a continuity in the foundations should be sought to as great an extent as possible.
Foundations should be designed to act as a whole. — If due regard is paid to the principle mentioned in the first part of this paper — to avoid throwing upward reactions upon connecting walls, or of transmitting shearing and transverse strains upon insufficient masonry construction — foundations can be so designed as to act as a whole, and without causing deformations or cracks in the superstructure.
Solid masonry often cracked by earthquakes. — In the earthquake which occurred a few years since in the Vaca Valley of Cali- fornia it was observed that less damage was apparently done to some of the older and more " flimsily " constructed buildings than to those of more firm and rigid masonry. This was due to the fact that the old buildings were loose-jointed, and simply separated and pulled apart in many places without uniform vibration as a whole. On the other hand, the firm brick walls cracked throughout their entire lengths.
This is in accordance with Prof. Milne's investigations in Japan. He says: "An important point, which constructors should keep before them, is to avoid coupling together two parts of a building having different vibration periods, or else to couple them together so securely that they shall move as a whole."* The trouble with the Vaca Valley buildings was, they had sufficient strength to gather great vibrations, but not enough to resist final rupture.
* Inst, of C E., Vol. C.
12 ASSOCIATION OF ENGINEERING SOCIETIES.
Proportions to resist earthquakes. — Theoretically, the weight of a high building should decrease uniformly from the roof to the foundation. The weight of one story, as coucentrated mostly in its floor system and in its exterior walls about the floors, should not be carried upon too slender piers to the story below.
Unfortunately the modern demand for light and space makes a very undesirable condition of affairs in this respect; in the first one or two stories above the foundation very often the ground floor is given up to shops, and nearly all the space which should be in walls or heavy mas- onry piers is converted into large windows and openings. The entire building over this floor is generally carried upon iron pillars. The vibra- tion of the massive structure above them can ouly be transmitted to the foundation by means of these small columns, throwing a duty upon them which is most tremendous; and, in fact, they are unsuited to taking up these vibrations and transmitting them to the foundations and the ground.
The building doe3 not vibrate as a whole, and cannot do so with this method of construction. And particularly in the modern structures of great height should attention be paid to this principle of providing mass and weight in the base, with the least possible amount in the top of the building. Between the roof and the foundation both mass and weight should be gradually proportioned without such open construction as to permit of the independent vibrations of different parts of the building.
Steel beams in concrete. — Concrete and masonry have not, as a rule, much transverse shearing or tensional strength. When used in foundation work it should be the aim to so proportion their dimensions and positions that they will be subject to compression only. Steel beams introduced into concrete do away with the deficient strength of the con- crete alone and renders it safe for transverse strains. For this purpose this construction becomes most valuable in foundations, not the least of which is that the exact dimensions needed are susceptible of accurate calculations.
Magnitude and weight. — It is not wise, however, in many cases, especially when earthquake vibrations are liable, to make the founda- tions too shallow, simply because the steel beams in themselves may have sufficient strength to take the strains that come upon them, for the reasons which we have explained before, of providing a base of magnitude and weight to take up the vibrations transmitted through it.
In the first part of this paper I have mentioned the fact that there is no objection to foundations of masonry alone without depending upon an interior steel stiffening. What I mean by this is that such a founda- tion, proportioned mostly for compressive strains, requires considerable depth in order to secure sufficient area of base, and hence acquires large mass and weight. The considerations which have just been shown are
FOUNDATIONS FOR TALL BUILDINGS. 13
among the chief ones to prevent undue transmission of vibrations to a superstructure built upon them and to give great rigidity to the structure as a whole.
Monolithic foundations. — I find numerous examples in German cities of successful monolithic concrete foundations under heavy buildings. But in all such cases the great thickness of the concrete is evidently relied upon to give sufficient strength to the base to resist uneven reactions. The Nicolas Church in Hamburg rests on a bed of concrete 8 feet thick, while under the tower this thickness is increased to 1H feet. Aud other buildings are recorded with depths of from 5 to 6 feet.
In contrast to these the monolithic foundations of Chicago have generally failed. The City Hall settled very unequally, as much as 14 inches ; but the bed of concrete under it was about three feet thick, and this appears to have been as great a thickness as was used in any of the other buildings upon such foundations.
A monolithic concrete foundation stiffened with heavy steel beams at right angles to one another, and not too shallow, makes an ideal system of construction, provided correct dimensions are given to the base.
Foundations of the "Call" Building, San Francisco. — The building now being built by Mr. Claus Spreckels, in San Francisco, and commonly known as the "Call " Building, rests upon a bed of sand. It is almost square, being 70 feet by 75 feet in plain dimensions. The foundation consists of a layer of concrete 24 inches thick, then a tier of 15-inch steel beams spaced 18 inches centers, and then another at right angles to this of the same dimensions, all void space being filled with concrete. On top of this solid platform rest 20-inch beams grouped together under the columns as shown.
On the sides of the lot adjoining other buildings the steel beam and concrete platform is extended, so as to be used in supporting them. All beams are spliced, so as to be continuous from end to end. The outside columns are anchored to the foundation with two anchor bars of steel H inches by 8 inches, which are fastened to the under side of the lowest tier of beams and extend up and into the columns themselves, to which they are riveted. Messrs. Reid Bros, are the architects for this building.
Pile foundations. — To us on the Pacific Coast it appears strange, in investigating the foundations of large buildings in other-parts of the country, to notice the avoidance of piles as a means of founding.
Both in Chicago and New York the underlying hard pan and rock appear to be within the reach of long piles. At Chicago, firm clays appear from 45 to 50 feet from the surface, and the rock about 80 feet, while in New York the rock generally occurs in less than fifty feet.
14
ASSOCIATION OF ENGINEEKING SOCIETIES.
The shock of driving piles next to other buildings undoubtedly has been one cause which has in some cases precluded their use; but it would appear that in Chicago it has been largely due to a distrust occasioned by the failure of several buildings founded on piles a num- ber of years since. By the advice of several prominent engineers of late years a very few of the most recent buildings have been placed upon piles.
Piles make firm foundation. — Under the conditions of piles
Plan of Foundation of the Ce.ll BuitdtHg* Fig. 4.
being driven into a firm soil there is no better foundation, provided they are kept continually wet, and hence safe from decay.
When they depend entirely upon a small surface friction on their sides, as in our own shore mud and clays, we know even then that when driven to great depth they are safe for moderately heavy loads.
I desire to carefully avoid any discussion here of the very broad subject of the supporting power of piles, beyond the general statement that it is but a rare situation and condition of things when full and
Section of the "Call" Building, San Francisco. Fig. 5.
16 ASSOCIATION OF ENGINEERING SOCIETIES.
ample supporting strength cannot be secured by a properly driven pile foundation.
Should be driven deep. — If piles driven thirty or forty feet do not show sufficient resistance during driving, and the character of the soil is such that one does not care to risk its remaining with the same supporting power after a lapse of years, I can see no reason for stopping at this depth.
This is, and always will be, one of the chief causes of pile founda- tion failures, viz., the inability to recognize the proper depth to which piles should be driven. As I have said before, it is a very broad subject, and has many most interesting and abstruse matters relating to it, which would require an even longer paper than this one to make clear.
Grillages. — A grillage of timber on the heads of piles makes a most efficient base upon which to found a structure. In distinction from the very common method employed — largely in Europe — of placing concrete upon and around the heads of piles, it is the more usual American practice to build upon a timber grillage.
I think it is obvious, for the reasons stated before, that transverse breaking strains should be avoided in masonry of any kind ; also be- cause of a ready means of tying and bracing together in various direc- tions all the piles of a foundation, and thus causing them to act as a unit for stresses of all kinds that the timber grillage is much the more preferable method of construction.
Cantilever construction. — A not uncommon condition of affairs, in designing the foundations for a tall building, often occurs in which it is necessary to keep entirely within the lot line and the outer line of the walls of the building with the substructure work. Under such circumstances it is difficult to avoid throwing undue strain on the outer edge of the footings. If the soil is a compressible one it is of the utmost importance that the center of the ground areas and that of the pressures transmitted to them should be concentric. In the case of buildings founded directly upon the soil this is often very difficult to manage satisfactorily, aud in fact, under any method of founding, it is undesirable. In New York this problem has been met, in several notable instances, by the device of constructing the building upon the great steel cantilever beams, which overhang from tubular piers founded upon the solid rock, aud placed entirely within lot lines.
By this means the center of pressure from the weight of the building can be transferred to the center of the foundation areas.
Broad base most desirable. — It is always better construction in any structure to found on a broad base, and no matter how firm a foundation may be secured, a great, tall building, two hundred or more
FOUNDATIONS FOR TALL BUILDINGS. 17
feet in height, resting upon a base, the exterior edge of which encloses an area smaller than the plan of the structure itself, is not in the most desirable condition of stability.
Of course such designing is not the result of choice, but from necessity.
Foundation direct on sand. — Of late years it has become somewhat more common practice in the city of New York to found some of the great buildings directly upon the sand, which is a natural forma- tion there, while in Chicago there is an undoubtedly growing tendency to found upon piles driven into the hard clay, or upon wells excavated down to the hard strata in the soil and then filled with concrete. The practice in both places, by a process of evolution, is simply approaching correct theoretical principles.
Sand generally safe. — Sand, provided the same is not liable to future disturbance by nearby excavations, or from flooding or other causes, is, as is very well known, a particularly good foundation, and per- mits of loading with considerable pressures per square foot and with a minimum of compression.
The late distinguished engineer, Alexander Holly, boldly founded an important structure upon a quicksand, but first took the precaution to permanently enclose it.
Factor of safety for supporting power of soils. — It is considered good practice, and entirely permissible, to strain up to 16,000 pounds per square inch on the steel beams used in tall building construc- tion. This is about 50 per cent, of the elastic limit of the metal.
Why are not these conservative principles applied to the supporting power of soils in foundation work? Why is it more correct to load a soil which shows considerable compression and change in shape under a loading of, say 4,000 pounds per square foot, with a constant and un- changing pressure of from 3,000 to 3,500 pounds?
The author is satisfied that upon wet clay soils, or loam, or upon sand which is more or less impure, more moderate values of loading must be used than has been customary in the past in many of our greatest buildings.
Main principles. — The first and chief axiom in all foundation practice is to know the exact character of the soil upon which the struc- ture is to be built.
All other things become secondary to this great principle, and in fact resolve themselves into simple mechanical problems capable of definite solution.
All soils have certain safe supporting power, and are likewise sus- ceptible of a definite amount of compression.
The loading of the same must be safe within the limits of such press- ures as produce measurable settlement. 2
18
ASSOCIATION OF ENGINEERING SOCIETIES.
C
Pq
N
o
FOUNDATIONS FOR TALL BUILDINGS. 19
The investigation of the soil must be complete for considerable depths and particularly is this necessary for pile foundations *
Having acquired a true knowledge of the physical characteristics of the building site, the foundation design must be such as to recognize the various principles commented upon above.
Architecture is a science and art largely based upon historical pre- cedent, but the foundation of a modern high building is a question of engineering construction. Excepting that precedent furnishes informa- tion with regard to the strength of materials it is a dangerous rule to follow blindly.
DISCUSSION.
Mr. Wagoner. — I would like to ask Mr. Hunt if the soil under the "Call " Building is sand ?
Mr. Hunt. — I have been informed that it has a uniform sand foundation to a depth of forty feet or more.
Mr. G. W. Percy. — I have enjoyed the paper very much, and espe- cially what was said about building foundations in Chicago. Some twenty-four years ago, just after the great fire in Chicago, I was engaged on some of the heavy buildings there, and I have a special interest in Chicago foundations.
At that time the system of building on isolated piers was just beginning to be recognized as the proper thing. It was very seldom practiced. It was the common practice there among the architects, when any science at all was used in building, to make the maximum load about one and one-half tons to the square foot. I was engaged in the office of the leading architects in Chicago. After the great fire they did a great amount of work. In a year and a half they put up a mile and three-quarters of street frontage of buildings to be used for office purposes, and these buildings ranged from four to seven stories high. This firm was then an advocate of isolated foundations, although they did not practice it, except on some of the buildings in which they were very anxious there should be no cracks or any marked settlement. In the Kendall Building, which is an office building of fire-proof con- struction, some of the boys from our office were sent there every month to take levels through the building, to see if any one part was settling faster than the rest. We figured on its settling about two inches,
* A most important bridge foundation failed in Philadelphia, because in the driving of the piles they were left with their ends just about to penetrate an un- suspected soft strata of mud. If a careful boring had been made beforehand it is need- less to say the piles would probably have been driven deeper and the accident averted.
20 ASSOCIATION OF ENGINEERING SOCIETIES. '
and that the pressure would be one and a half tons to the square foot. When it was found that any column was not settliug as fast as the rest, pig-iron was taken into the building and placed upon that column until it settled equally with the rest. A large amount of pig-iron was used to bring about this result.
Something of the same nature was practiced on the Auditorium Building, I believe. The building was of uniform height and weight up to the base of the tower where it emerged from the roof. The weight of the tower above the roof was between three and four thousand tons. The architects loaded the foundations of the tower with about three or four thousand tons of pig-iron, representing the excess of weight of tower above the roof, and kept that load on the foundations until the entire building was up to its roof-line. Then as the tower was carried on higher they would remove the pig-iron from the base, the object being that when the building was at the line of the roof the weight was equal on foundation, and equal settlements should have taken place, and they removed the pig-iron as they added brick and stonework. So when the tower was completed it had no more weight upon the foundation than when at the roof-line. That was quite suc- cessful, but not entirely so. Some considerable settling took place at the tower, which the architects explain by claiming that some five hundred or six hundred tons were added by changes and alterations made on the tower.
I think Mr. Hunt has explained the real causes why these very lofty buildings settle so much more than the one and a half inches which used to be recognized as the proper amount of settling.
It is evident that, given a thick layer of clay, such as they have in Chicago, of seven or eight or ten feet thick, and a softer layer of clay under this, the ordinary style of building of five, six, and seven stories high, and a pressure of one and a half or two tons to the square foot, as the case may be, distributed over the entire area of the building, it was not sufficient to cause the entire strata to bend or yield, and therefore that the settling of the building was just the compression of the harder clay. But with these enormously increased loads of the high buildings the whole body of this upper layer of hard clay settles, and it is the lower strata that is overloaded rather than the upper one. I think that is a proper solution of it, because it is quite certain that an ordinary building of six, seven and eight stories in height, loaded to two tons even to the square foot; does not settle more than two inches or thereabouts, while some of these heavy buildings have settled five or six inches. I think this increased load is transmitted to the softer strata below.
In regard to the question of piles in Chicago, the reason given in those days for not using piles was that there was something peculiar
FOUNDATIONS FOE TALL BUILDINGS. 21
about Chicago clay which was not adapted to their use. In most places where piles are driven in mud or clay, and even where it is quite soft, when a blow will drive the pile two or three inches, if you let it stand six months and then apply the same blow, it will not move it ; but in Chicago it is the reverse, and you strike a pile after it has been left this length of time and it will go out of sight. It is claimed that after the fire, build- ings on pile foundations settled considerably.
In this city we have a very good hard sand, and there are many build- ings where the foundations are loaded to about four tons to the square foot without any perceptible settling, not perceptible enough to make any cracks or dislocations.
I would take some exceptions to Mr. Hunt's remarks, as I gathered from what he said that he did not consider it best to load foundations so near the yielding point, and that the soil should not be loaded more than one-half of its elastic limits. I do not see that the argument applies to foundations. In the case of masonry, or most any material employed in buildings, there is a possible deterioration going on ; but in the case of foundations there can be no deterioration ; the foundations of sand under a building do not deteriorate with age, and the load may be very near to the point to which some yielding would take place. But if loaded double or treble that load, no serious consequences would result, there- fore I do not see why it is not prudent and safe to load a foundation to near its yielding limit.
I would also make one other suggestion. The members of the Tech- nical Society know I have advocated the use of twisted rods with con- crete in foundations, and that I have made some experiments in this line. I make a uniform foundation, a platform for the building to rest on. This method is a great economy in materials as compared to some others. Take such a case as this : A platform of concrete, say 4 feet thick, with a sufficient quantity of twisted rods placed both top and bottom of the concrete, and one-fourth of the amount of steel in this foundation, would be equal in strength in every particular to a foundation where steel beams are put in.
Prof. Marx. — It may be of interest to know that Mr. Sooysmith, who has probably carried out more important foundation work in this country than any other engineer, is to bring this subject out at the April meeting of the American Society of Civil Engineers. I just happened to glance through a little paper which details the point he expects to bring out. Mr. Sooysmith calls attention to the fact that pile foundations in a number of instances have not been satisfactory, owing to decay of the piles due t6 a change in the subsoil water levels. The supposition was that the subsoil water level would remain permanent, but this supposition was found to be wrong. The piles are alternately exposed to water and air, and under these conditions of course they rot away.
22 ASSOCIATION OF ENGINEERING SOCIETIES.
In the matter of sand foundations Mr. Sooysmith mentions the fact that in New York City, in the case of large buildings built upon sand, such a method of founding is rather dangerous ; that sand is a good ma- terial when confined, but that oftentimes when a neighboring building is torn down and a new building put in its place, the sand gives way to some extent, causing a subsequent settling of the building. Then the responsibility for whatever injury may occur falls upon the man who has erected the new building. He therefore suggests — and I think it is the method he has carried out — that the foundations be carried down to the solid rock by the use of pneumatic caissons.
I mention this as showing that the subject of foundations is interest- ing, and is agitating engineers at the present time.
Prof. Wing. — A question that comes up in regard to foundations composed of steel and concrete is that of the durability of steel in concrete. While it has been accepted that steel will last, that it is indestructible in concrete, yet I think there has been no definite determination of that fact, and it is not definitely known that the material will last. Observation of structures that have stood for some time I think points to the fact that it will last.
The plan of building on a continuous floating foundation under a building is one that depends upon whether the pressures of the build- ing are uniform over the whole area or not. This inequalit)7 of pressures in the case of the Chicago buildings is corrected by giving each column a bearing area under it proportioned to the load it carries. As I under- stand the matter, they carry out this design so completely that where two foundations of concrete meet they separate them by a board, so there shall be no communication between the two foundations, thus prevent- ing the concrete under two columns from acting as a beam and produc- ing eccentricity of pressure on the foundation bed.
Mr. Percy. — I want to state another interesting fact about the foundations of the Post Office in Chicago. The matter created con- siderable discussion at the time the foundations were put in. It has been referred to very frequently as a failure of continuous foundation. I had an opportunity of observing the way it was put in. If I remember, the concrete was supposed to be four feet thick over the entire area of that great building. I watched the process of putting it in through the cracks from time to time. There was a layer of six inches of broken stone laid down, and then cement grouting poured over it ; then another layer of stone six inches thick, with cement grouting, and so on. The papers spoke of it as the most perfect method of laying concrete that had been devised. The result has been that the building has settled, and there are cracks in the building in some places two inches, in other places eight inches wide, and in some places perhaps more ; there has even been a breaking of columns and beams throughout the building.
FOUNDATIONS FOE TALL BUILDINGS. 23
Prof. Soule. — I would like to have Mr. Hunt, or any other gentleman in the room who knows, state what has been the amount of settling of the Appraisers' Building here in San Francisco. You remember its foundation is a thick layer of concrete. I suppose some of you saw that construction. I believe the concrete is spread over a larger area than the building stands on. I think the concrete is about six feet thick. Of course the building is a heavy brick structure. I have been told that the settlement has been considerable, but as far as I know it has been pretty uniform over the whole area. I fancy if the weight of the building itself had been evenly distributed over that concrete slab, that with the thickness and weight of the slab itself the building would have settled uniformly. There was a discussion at the time as to whether the Appraisers' Building should rest on piles, as does the Post Office Build- ing next to it, or on a concrete slab. The latter plan was adopted, I believe, by the late Gen. Alexander and Col. Mendell.
With regard to pile foundations I think that sometimes after long periods of time unequal settlement occurs in the building, due to a different cause from those mentioned by Prof. Marx. In Venice a great many buildings stand on piles driven in the mud of the islands of the archipelago ; they were put up from the year 800 to 1000, and so on to 1200, during the period of the glory of Venice. Some of the heaviest buildings there were erected, I think, about the year 1200 or 1300. From an examination of some of those I am satisfied that while unequal settlement in some of the buildings has caused them to lean over in a rather dangerous way, in fact in some places to threaten to fall down if unsupported by their neighbors, the settlement has been caused in some instances by unequal weights on different portions of the piling, while in others I am very sure (and I draw this conclusion from personal ob- servation) that the settlement has been caused by the long continued weight upon the piles, causing an actual telescoping of the fibers one into the other ; in other words, an actual shortening of the piles through the cellular spaces being diminished. I saw in some places evidences of the piles having been compressed in the direction of their length to a con- siderable degree.
In our pile foundations, if we expect them to endure heavy loads for hundreds of years, the question of a thorough equalizing of the load coming upon the piles would, I think, be quite an item in affecting the final stability of the structure.
Mr. Percy. — I would like to say one word in answer to Prof. Soule's question. The foundation of the Appraisers' Building was put in before I came here, but when I arrived the building was not finished. I am acquainted with the superintendent of the cement work, and he told me what would indicate a dangerous condition of the foundation. There
24 ASSOCIATION OF ENGINEERING SOCIETIES.
was a great body of concrete laid over the entire area, and at the north end it was resting on rock, while the other end was resting on soft mud, which, of course, we all realize to be a very dangerous condition. I have watched the building to see if there were any cracks from settling. There are one or two small cracks on each side, showing some movement, but nothing of the extent that would be expected under such conditions as this superintendent described to me. I think the Appraisers' Build- ing has stood very well, and there has been very little movement of the foundation.
In regard to pile foundations and the capping of piles I would like to hear some further discussion. Instead of simply capping the piles with timber I am strongly in favor of digging around the piles six inches below the tops of the piles and ramming concrete all around the piles. In this way we get a full bearing upon the piles, and in addition to that we get the bearing capacity of the soil, whatever it may be ; it binds the piles together as well as any grillage could do, and it is much cheaper. On the whole, I think the advantages are in favor of capping with concrete.
Within the past six months I have put up a building in this city alongside of a building resting on piles. I went below the foundations of the building, and I was curious to see what sort of bearing the building had upon the piles. I dug about the cappings, and I found places where I could push my rule in between the caps and the tops of the piles. The building had not been loaded heavily enough to bring the capping down. The piles were not cut off on an exact level line, and therefore the grillage did not x'est on them properly. But in using concrete the way I have described, every square inch gets a bearing, and, to my mind, it is better than grillage.
Mr. Hunt. — The discussion has followed pretty near the lines I thought it would. I have been somewhat disappointed that nothing has been said in regard to the earthquake vibrations. There are only a few points that I would like to answer.
With regard to Mr. Percy's remarks about a foundation made of twisted steel rods in concrete, placed near the top and bottom, there is no criticism whatever to be made. The construction is exactly in line with the proposition I tried to bring out, namely, that when a continuous foundation is placed under a building it should act as a beam from one column to another, aud the foundation must be of sufficient strength to allow this.
When the foundation is made entirely of concrete it should be made in the very best manner and of very great thickness. Not doing this has resulted in failures and results that were not satisfactory.
Prof. Wing alluded to the principle, used in Chicago, of separating
FOUNDATIONS FOE TALL BUILDINGS. 25
the foundation. I have here a drawing of the Old Colony Building, which is twenty stories high. The foundation plan shows the method of keeping the foundations entirely separate, even when the concrete bases extend close to each other. All the concrete areas have a distinct line of separation between them. This idea, when first started in Chicago, was used to such an extent that every single column had its own independent foundation, and with great care they kept them sepa- rate. Now they have commenced to group them more or less together.
The great Manhattan Building in New York, one of the highest of buildings, is constructed on a foundation in accordance with the ideas of Mr. Sooysmith, as represented in the paper he will present to the American Society of Civil Engineers, alluded to by Prof. Marx. His argument is on the line I have introduced here with regard to vibrations, only he had no reference whatever to earthquakes. It is that all build- ings are subject to vibrations, however small, and sometimes they create considerable disturbance. Even the running of a hoisting donkey engine in the construction of a building causes vibrations that result in a settling of the building, The driving of piles will cause damage to adjoining buildings.
Prof. Wing. — I think my remarks in regard to building a con- tinuous foundation have been misunderstood. Take a building like the Auditorium Building in Chicago, having a large tower in one portion of the building, and the other portions being of less weight and of less height, provided the foundation cannot cover a greater area than the building, if the building is constructed on a continuous foundation, there will be an inequality of pressure, and provision must be made for settle- ment where conditions like those met with in Chicago exist. In some cases the cantilever construction has to be used, as illustrated in Mr. Hunt's paper. If the building is simply of square construction, of equal weight in all portions, and of uniform height, I can see no objec- tions to the plan of putting under a continuous foundation.
Mr. Hunt. — I am not an especial advocate of platform founda- tions, excepting when the conditions are favorable. I see no objections to this class of foundations in certain cases, and think they have distinct advantages. Of course, in a building of the size of the Mills Building in this city, it would be impractical to put it on a platform foundation covering the whole area. But the point I have tried to bring out in the paper is that the pressure upon the soil should be such that practically there would be no settlement. We know that all soils will support a certain amount of weight without compression, but they should not be overloaded. That is the principal point in all foundation practice ; it is the underlying principle. It makes no difference if the foundation is upon soft mud, if we only establish the point of how much load it will carry without compression. The soil should not be overloaded.
26 ASSOCIATION OF ENGINEERING SOCIETIES.
Under certain conditions, where it would be too costly to secure more land, or something of that kind, of course that changes the situa- tion. But if it is possible to avoid it, I think the soil should not be loaded so as to compress it to the poiut of perceptible settlement.
Mr. Leonard. — In regard to the settling of columns, a device has been used in Chicago, and I am told it is to be applied in New York, consisting of a space left for the hydraulic adjustment at the base of the column, so as to keep the building perfectly adjusted. As it tends to settle, the column is raised and steel wedges are inserted under it. It is taken care of in this way, and at the end of three or four years the build- ing is supposed to have settled to a permanent position and to need no further adjustment.
Mr. Curtis. — On general principles there seems to be something about a foundation which is bound to settle if it is upon the natural soil, and I think we are always likely to have a higher respect for the scrip- tural man who founded upon a rock than for the man who built his house upon the unconfined sand.
To sum up the discussion, the idea suggested might be that the foundation that would meet all the objections to piles, or to timber and grillage, would be cylinders or wells sunk to the solid substratum and filled with concrete and capped with a concrete base for the whole structure, constructed somewhat upon Mr. Percy's plan, with steel near the bottom and near the top.
Mr. Percy. — I will say, Mr. President, that the nearest approach that I know of to such a method is a church in Paris, the Church of the Sacred Heart. They commenced its construction some years ago. It is a very large, heavy church, and situated on quite a high hill. They found they had a very unsatisfactory foundation. It was a mixture of clay with other materials, and was not at all satisfactory to the engi- neers, and they sank cylinders down 80 to 90 feet under all the main piers. They went down to the solid stratum and filled those cylinders with concrete, and instead of making a continuous platform under the entire building, heavy arches were sprung from one pier to another. These large cylinders were put down only at points where there would be great bearing. This is the nearest approach to what you suggest, which, I agree with you, would be the perfect foundation, answering all requirements that have yet been suggested.
LOCOMOTIVE COUNTERBALANCING. 27
LOCOMOTIVE COUNTERBALANCING.
By G. R. Henderson, Member of the Association of Engineers of
Virginia.
[Read before the Association, June 27, 1896.*]
The subject of locomotive counterbalancing has recently been quite a favorite one, and there have been many valuable papers on this theme, but most, if not all of them, have been deficient in one par- ticular ; in that they have not clearly and simply indicated how to proceed with each part of the problem. For instance, one paper gave very carefully worked out formulae for determining the effect of recipro- cating weights, and how to correctly balance them, but the proportion of reciprocating weights to balance was passed by with a mei'e reference, as though of small consequence, when in reality it should be the funda- mental question. In the following it is not the writer's intention to advance new theorems, but to select such points from previous papers (including those by Messrs. Parke and Sanderson before the New York and the Southern and Southwestern Railroad Clubs respectively) as, with a few logical suggestions, will place the subject in the hands of every Master Mechanic.
In developing these rules, three cardinal points have been borne in mind :
(1) The amount of reciprocating weight that can be left unbal- anced may be a definite function of the total weight of the engine.
(2) The total pressure of wheel upon the rail must not exceed a certain definite amount depending upon the construction of bridges, weight of rail, etc.
(3) The vertical influence of the excess balance must never be suffi- cient to lift the wheel from the rail.
The first proposition is based on the assumption that the greater the mass, the greater may be the disturbing force without seriously affecting it, on account of its greater inertia.
The second is evidently a rational deduction, not needing any demonstration.
The third is necessary in order to avoid the wheels' jumping off the rail, thereby causing a real "hammer blow."
Starting with the above assumption, we arrive at the following conclusions :
A. Each wheel should be balanced for all revolving weights attached to it.
* Manuscript received July 20, 1896. — Secretary, Ass'n of Eng. Socs.
28 ASSOCIATION OF ENGINEERING SOCIETIES.
B. The connecting rod is to be considered as part revolving and part reciprocating weight ; the proportion of weight of rod which is to be considered as revolving weight varies with the length of the rod as given below :
Length of rod in) 5 6 ? & g 9 & 1Q n & 12
feet, )
Proportion as ) 57 M 53 52 51 revolving weight, j
C. The part of weight of connecting rod considered as revolving weight, should be entirely balanced in the main wheel.
D. The amount of reciprocating weight that can remain unbalanced without seriously affecting the locomotive may be found by the formula :
Wt Wr =
360 Wr = unbalanced reciprocating weight on one side (including portion of main rod).
Wt = weight of locomotive in working order.
E. The remainder of the reciprocating weights should be counter- balanced by dividing the amount equally between the driving wheels on the side, provided that the sum of the static weight on any one wheel, plus the centrifugal force of this overbalance, does not exceed the maximum pressure allowed for the particular type of engine in question at the maximum speed at which it will run. If some wheel loads are heavier than others, the lighter wheels may take a part of the overbal- ance which the heavier wheels cannot without exceeding the specified limit ; nor must the centrifugal force exceed 75 per cent, of the static load on wheel.
F. The center of gravity of counterbalance must be opposite the crank.
G. The counterbalance should be brought out from the face of the wheel as far as clearance for the rods and proper design will permit.
H. The center of gravity of counterbalance should be placed as near the rim as possible, and the weight of the counterbalance reduced by this method.
I. Make reciprocating parts as light as possible.
Section A is self-evident. B is taken from .one of the papers above referred to. C comes under the same ruling as Section A. In D the
value Wr = t , is taken as representing good practice of the present
day. It may be found that some different divisor will be more gener- ally acceptable, but it is believed that the above will give good results.
LOCOMOTIVE COUNTERBALANCING.
29
30 ASSOCIATION OF ENGINEERING SOCIETIES.
To determine the centrifugal force for Section E, the following formula is obtained from Weisbach's " Mechanics of Engineering," Vol. I, page 609 :
P == .00034 u2 Gr. where
P = Centrifugal force.
u = Revolutions per minute.
G = Weight in pounds.
r = Radius in feet. Now 'letting
S = Speed in miles per hour. D = Diameter of wheel in inches.
we have
S X 5280 X 12 S X 1056 g
u = = = 336 j.
and
3.1416 X D X 60 3.1416 X D
S2 ■ iizoyo
and substituting,
u2 =112896 L
S2 P = 38.4j-2 Gr.
As in most locomotives P = 1 , then we may put simply,
P = 38.4 J^j G.
If now we assume that the maximum speed in miles per hour of the locomotive equals the diameter of driving wheel in inches, then,
S2
jj2 = 1 and P = 38.4 G, or say
P = 40 G.
It is also necessary to observe the limits of rail pressure. This will be different on various railroads, but on the Norfolk & Western it was taken as follows :
American type of locomotives . . 28,000 pounds per wheel.
Ten-wheel " " 26,000
Consolidation " " 25,000
[These loads are per wheel and not per axle or pair of wheels.]
Referring to Section F, it is found that the displacement of the counterbalance necessary to correct the effect of the weights and balance not being in the same vertical plane is so small on outside cylinder engines, that it is accurate enough to place the balance directly opposite the crank. By bringing the counterbalance out as suggested in G, it is possible to still more lessen the irregularity explained just above.
Sections H and / need no explanation.
LOCOMOTIVE COUNTERBALANCING.
31
32 ASSOCIATION OF ENGINEERING SOCIETIES.
Having taken up these various points, the method of counterbal- ancing locomotives can now be reduced to the following :
RULE.
Divide total weight of engine by 360, this to be subtracted from reciprocating weights (including proportion of main rod) of one side of engine, and the remainder to be distributed among the driving wheels on one side.
The sum of forty times the amount of reciprocating weight allotted to any one wheel and the static load on the wheel, must not exceed the specified allowance for rail pressure, nor must forty times the amount of reciprocating weight balanced, exceed 75 per cent, of the static weight.
The weights to be put in each wheel will be inversely as the dis- tance of center of gravity of counterbalance from center of wheel is to the crank radius, and must cover all revolving weights as well as the proper proportion of reciprocating weight.
In order to obtain the best results both for the engine and track, the following points should be remembered :
1. Keep the spread of cylinders as small as possible.
2. Make pistons of malleable iron, wrought iron or steel, to reduce weight.
3. Make piston rods of steel, and hollow.
4. Make crossheads of cast steel, of light ribbed construction.
5. Make the rods of steel and of an I section.
6. Keep counterbalances near the rim of wheel.
7. Keep counterbalance as far out as possible.
No. 1 can only be done when designing the engine.
No. 2 can be accomplished in various ways ; however, the single- plate pistons have the objection that they freely transmit the heat of steam side to exhaust side of the piston, but double-plate pistons are not readily examined, as they should be, especially when very thin. Besides, a cast-iron wearing surface is desirable, while bolts and rivets are equally undesirable. A design of piston that promises very favor- able results, and will meet all the above objections, is shown in Fig. 1. The center is malleable iron, and the wearing ring cast iron, the latter fitting against a shoulder at one side, while a brass retaining ring is cast in and opened out on the other side, making practically a single piece. It also takes ordinary cylinder heads.
For No. 3, the use of nickel-steel has been suggested.
No. 4 depends entirely on the arrangement of guides, etc.
For No. 5. Fig. 2 shows the. favorite form.
No. 6 may be accomplished as shown in Figs. 3 and 4, in preference to Fig. 5.
No. 7 is limited by the clearance necessary for the rods, etc.
EIVETED JOINTS. 33
RIVETED JOINTS.
By Joseph E. Worcester, Member of the Boston Society of Civil
Engineers. [Bead before the Society, April 15, 1896.*]
In spite of the fact that the tendency of the present time is more and more towards the use of riveted work in the construction of bridges and buildings, it is somewhat surprising that we hear of no change towards improvement in the customary methods of calculating the strength of riveted joints. Perhaps it will not be accepted without proof that the tendency of the times is in the direction just noted, but a little careful consideration will show this to be the fact, at least in this country.
The earliest iron bridges in general use hereabouts were constructed with stiff compression members made up of all sorts of rolled sections, or of cast-iron columns, tied together by means of forged rods, which, later on, were superseded by eyebars, and connected by means of pins. Likewise, until quite recently, the only iron used for framing buildings was in the shape of cast-iron columns upon which the beams were sup- ported, the only connections being made by means of straps and bolts. As the science of bridge building developed, and the demands of rail- roads became more pressing, the speed of trains became greater and the loads heavier, it was found that the light pin-connected structures were gradually being rattled to pieces, the vibrations increasing to an alarm- ing extent. In the endeavor to find some form of construction which would not show these defects, it was natural that we should turn our eyes towards the practice of European engineers and see what advan- tages we could gain by adopting more of the riveted form of construc- tion, of which many fine examples were in use on the other side of the Atlantic.
It was at this period that a distinguished member of this Society, the late Edward S. Philbrick, designed the many plate girder bridges which have so well served their use in this vicinity for a generation, and which have never proved unequal to their original requirements until their metal was nearly eaten away by corrosion. One, if not more, of these stood until holes were rusted entirely through the web.
About this time some of our bridge companies began constructing riveted lattice bridges. These bridges were a distinct advance upon the earlier forms of pin-connected structures. Their trusses never became
* Manuscript received July 22, 1896. — Secretary, Assyn of Eng. Socs. 3
34 ASSOCIATION OF ENGINEERING SOCIETIES.
shaky in spite of the unscientific connections, the frightfully bad inter- sections of web members, and the faulty and incomplete systems of bracing ; but whenever they have givers out, it has been on account of the fact that the floor beams and stringers were not of sufficient strength to carry the increased loads, or the connections in the floor system were not as efficient as the trusses. Many examples of these bridges are doing good service to-day, though probably most of them have had their floor system strengthened or wholly renewed.
Within a short time one of our members has told us that compar- ing two bridges, a pin and a riveted, of equal theoretical strength, the riveted bridge was very much the stiffer, and consequently, in his opin- ion, the better bridge.
As time went on, and experience accumulated, we find railroads specifying that riveted girders and lattice trusses shall be required for longer and longer spans, until now we see riveted joints containing two hundred and fifty rivets, to be driven in the field, used in the center truss of a heavy four-track drawbridge, and we see our railroads using plate girders for highway bridges of one hundred feet span. We see all modern specifications for railroad bridges requiring riveted lateral brac- ing at the track level ; while, upon the other side of the water, we see such a bridge as the Firth of Forth riveted throughout, in spite of the fact that our eyebar practice was well understood and carefully considered in England at the time this structure was built.
In our building construction we see the same tendency. The loose strap and bolt attachments are giving place to riveted connections, and the cast iron columns are being superseded by rolled steel sections which will permit better riveted connections.
The object which we are striving for and gaining by these changes is the rigidity which seems to be inherent in riveted work.
Notwithstanding this tendency towards the increased use of riveted work, we are still using the same methods of proportioning riveted joints that have been in common practice for fifteen years, in spite of the fact that this practice involves many manifest absurdities.
The earliest authorities on the use of rivets, with more reason than we are in the habit of accrediting to them, took into account only the shearing value of the rivet, but for a long time now we have been taught to use either the shearing strength or the bearing value of the rivet against the metal opposing it, whenever the latter appears to be less than the former. These two strains are all that are considered in modern practice, though some have even gone so far as to consider the fiber strain caused by the bending moment.
In proper riveted work the writer ventures the assertion that neither one nor the other of these strains is exerted to any extent, and
EIVETED JOINTS. 35
it is with the intention of proving this assertion that the present paper is presented. It is not denied that before riveted joints will fail both bearing and shear will come into play, but when we are considering proper work we mean a class which is not on the point of failure and has not even reached the limit of elasticity, which it does as soon as a joint shows any permanent set. The rigidity, which is the essential characteristic of this form of construction, would not appear if rivets allowed a motion to take place between the thicknesses of metal con- nected. The force, therefore, which should be considered in designing the riveted joints, is that force which is exerted by the rivets to restrain the parts from all motion and to hold them in the precise position in which they are riveted.
The easiest way to demonstrate that it is neither bearing nor shearing strength which causes the rigidity of riveted work, is theoreti- cally, though we can quote also a number of practical illustrations which help to confirm this position. When a rivet is driven hot it is supposed to fill the hole, which is usually one-sixteenth to one-eighth inch larger in diameter than the rivet. This filling of the hole which specifications always stipulate, is more or less perfectly accomplished. When the holes are fairly concentric in the various thicknesses through which the rivet passes, the punch and die being not far from the same size, and the rivets are driven by a powerful machine, the metal will upset into the hole for a considerable distance until it apparently fills the void spaces ; but when the rivets are driven by hand or when the holes have large tapers, or when the rivet passes through a number of thicknesses, there are sure to be voids of considerable dimensions. Under the most favorable conditions, however, we must remember that at the time the rivet is driven it is at a much higher temperature than the surrounding metal, and as it cools it must inevitably contract, and in so doing draw away from the surface of the hole against which it was pressed. A section cut through a riveted joint in the axis of the rivet will show in a superficial examination the rivet to be in close con- tact with the surrounding metal, but with a magnify ing-glass, or with the point of a needle, it is very easy to see that the contact between the rivet and the surrounding metal is not as close as that between the head of the rivet and the surface of the plate, or between the various thicknesses of plate tied together.
It is this fact of there being a little play around the rivets which prevents their acting either by bearing or shear until there has been a little slip between the plates, that is, until the joint has passed its limit of elasticity. Up to the point when the slip occurs the force which prevents motion must be the friction caused by the pinching together of the surfaces by the rivet heads. That this frictional resistance must
36
ASSOCIATION OF ENGINEERING SOCIETIES.
be the force upou which we depend for the rigidity of our riveted structures may be practically shown in various ways. In the first place a great many experiments have been made on testing machines which invariably show that as the strain is applied the joint in the first place stretches only just as much as the metal itself stretches and in direct proportion to the amount of the strain. During this period, if the tension is relaxed the specimen returns to its unstrained position, but as soon as the friction is overcome a certain motion occurs, the
|
Strain in lbs per Rivet |
|||||||
|
0 JO - 0 H fi 5 |
0.00 |
SI |
15000 20000 25000 |
5000 10000 |
w 0 ore C )— i s a |
||
|
| |
1 1 — --j — i 1 |
||||||
|
""" |
|||||||
|
y^T |
|||||||
|
s |
1 |
||||||
|
0.05 |
/ |
||||||
|
/ ' |
1 |
||||||
|
J |
|||||||
|
/ |
|||||||
|
1 |
.- 1 1 |
||||||
|
0.10 |
4 ' |
0.000 |
|||||
|
/ |
J^J^- ~* ' |
||||||
|
/ |
A*rn\\ i i 0.001 |
||||||
|
/ |
. — ^ |
1 III |
|||||
|
/ |
r/ |
0.002 |
|||||
|
0.15 |
/ |
r |
|||||
|
/ |
0.003 |
||||||
|
/ |
|||||||
|
X / |
0.004 |
||||||
|
/ |
i |
||||||
|
0.20 |
J? |
0.005 |
|||||
|
/ |
|||||||
|
i M / |
0.006 |
||||||
|
/ |
|||||||
|
/ |
0.007 |
||||||
|
0.25 |
LL |
1 : , / |
|||||
|
/ |
0.008 |
||||||
|
— L h/ m |
|||||||
|
«z |
0.009 |
||||||
|
0.30 |
_L J |
0.010 |
|||||
|
) p =• p o p p | O | © O | ©j Cj j o |
J+ J" J" J |
||||||
|
Strain in lbs i |
»er Kivet |
Diagram Showing Slip of Eivets.
extent of which depends upon the void spaces around the rivet. After this the specimen shows a permanent set and the joint will go on stretching in an irregular fashion, as the metal around the rivet comes into full bearing. The diagram herewith presented illustrates this slipping very forcibly. It is taken from a series of experiments tried at the Watertown Arsenal in 1886, and is a fair sample of the several hundred tests made at that time.
RIVETED JOINTS. 37
In this diagram, the curve on the left shows the total elongation of the joint until nearly the time when failure takes place, the ordinates indicating the amount of strain per rivet, and the abscissae showing the stretch of the joint. The curve on the right shows the early part of the same curve on a magnified scale, the points at which the elongation is recorded being indicated by circles.
As will be observed, up to a strain of about 4,000 pounds per rivet, the elongation is approximately proportioned to the strain, but at about 4,300 pounds a sudden slip occurs. The exact point of slip is not recorded, the curve at that point being indicated by broken lines, which are produced by extending the general direction of the lines above and below the nearest observations.
Another method of showing practically that rivets do really hold by means of friction is by considering what would occur in the case of a plate girder with a large number of flange plates. In such a place as this it is impossible for any ordinary riveting machine to more than par- tially fill the holes. Any one who has had experience in cutting out very long rivets must have noted that they are easier to back out than short rivets through two or three thicknesses. The reason is that when the pressure of the riveter is applied to the end of the rivet it begins to upset at the extreme point, and as it upsets it fills first the part of the hole nearest the driving-head. As this fills out into the irregularities of the whole, it jams, so that it is impossible to force enough metal in through the hole to fill out the voids near the head end. We have, therefore, in this case even more play around the rivets than would occur where the hot rivet was completely upset. But these rivets have to transfer the strain from the flange angles to the outside plate, often a distance of several inches. If we imagine such a plate girder put together with loose-fitting pins, which indeed it would be were it not for the friction, it is evident that the girder would have to get a very con- siderable set, and the web and angles perhaps fail before the whole flange would come into play, if it could at all.
Another example of about the same action is where we see connec- tions made through loose fillers. We can often find in good practice the end-uprights of a stringer which transfer the whole shear from the stringer web to the supporting floor beam, packed out to the thickness of the flange angle by means of a bar of the same width as the upright. In this case if there were not friction exerted it is evident that the rivets would bend quite appreciably and thus allow the stringer to drop before bearing and shear came into play. Another evidence of the fact that it is the friction which is effective may be found in numerous examples of girders which were formerly constructed with very thin webs, and which with a bearing strain of the rivet against the web up to and
38 ASSOCIATION OF ENGINEERING SOCIETIES.
above the elastic limit of the material, have not shown the least sign of motion. The writer has in mind the case of a bridge on the Old Colony Railroad which was removed a few years ago, where this bearing strain caused by the every-day traffic of the railroad, amounted to 20,000 to 30,000 pounds per square inch, without allowing anything for the effect of impact.
That this fact of frictional resistance has been really recognized by engineers can be evidenced by the sensible though not very common practice of allowing a greater strain in bearing on metal enclosed by thicknesses acting in the opposite direction, than where not so enclosed. This practice appears to have originated in the thought that in a case of single shear the rivet would naturally bear upon the edge of the hole nearest the plane of shear with greater force than upon the other side of the plate, but the effect of the specification is in the line of taking account of the friction, of which we naturally have twice as much in the case of enclosed bearing as in that of not enclosed, and this fact is often advanced as an excuse for the practice.
The most conclusive experiments on the question of friction in riveted joints that the writer is aware of, were made in France two years ago, by M. Dupuy, Inspector-General of Bridges and Highways, who was intrusted by the Ministry of Public Works with the duty of making special inquiry into the causes of deterioration of metallic struc- tures. A full account of his experiments and conclusions was presented in Les Annates des Ponts et Chaussees, January, 1895.
The experiments were conducted with the greatest care, especially to determine the strength of the rivet before reaching the limit of elas- ticity. The conclusions which he reached are based upon a very clear and convincing argument drawn from the experiments. The following points which he arrived at seem to be unassailable :
1. Rivets are stretched bars undergoing a tensile strain higher than their initial limit of elasticity. The fibers of the circumference appear to be more stretched than those of the center.
2. Rivets do not quite fill the holes, but exercise a very strong clamping effect which causes between the plates a resistance to slipping equivalent to a welding.
3. The resistance to slipping of riveted plates increases as the limit of elasticity of the metal of which the rivets are composed increases.
4. The limit of resistance to slipping is extremely variable. The causes of this variation appearing very numerous and depending (a) upon the nature of the metal of which the rivets are composed, and (&) upon the temperature at which the rivets are driven ; (c) upon the temperature at the completion of the operation of driving ; (_d) the method of rivet- ing ; (e) the manner in which the operation is conducted.
RIVETED JOINTS. 39
5. The resistances to slipping upon which we can count in practice in connections composed of three rivets or more, in pounds per square inch of rivet section to be sheared, are shown by the following table:
|
Steel Rivets. |
Iron Rivets. |
|||
|
Original Limit of Elasticity in pounds per square inch. |
29,900 |
32,700 |
25,600 |
29,900 |
|
Hand-Driven Rivets heated to a bright red heat, the dies leaving no mark on the plates, the operation being finished when the driven head has be- come black. |
6,400 |
7,110 |
5,690 |
6,680 |
|
Power Rivets heated to a white heat, driven with a pressure equal to 85,000 pounds per square inch of rivet section, the pressure being maintained until the head has become black. |
8,530 |
9,390 |
7,110 |
8,250 |
The limits of elasticity mentioned above are those which are found in testing specimens previously heated to a dull red.
The metals employed in making rivets should have an elongation of at least 12 per cent, for iron and 18 per cent, for steel.
6. If a riveted joint is subjected to a strain sufficient to cause the thicknesses to slide, even if the motion is enough to distort the rivets, the distortion will remain after the strain is removed, but it ap- pears that no further distortion can be produced without a greater strain than that which caused the original distortion.
After establishing his conclusions in regard to riveting, M. Dupuy goes on to deduce rules for designing bridges, which, for the most part, are admirable, especially those relating to the desirability of avoiding secondary strains as far as possible by making perfect intersections of diagonals with chords, etc., but he draws some conclusions which seem to run counter to the ideas which have been gaining ground in Ameri- can practice, and which require further demonstration before we can accept them. For instance, he recommends that panels shall not be made longer than 11' 6" to 13' 0". He also recommends that in double Warren systems of trussing, verticals shall be used at each panel point to connect the two systems and so equalize the strains. He further ad- vises that bridges of several spans shall be made continuous over the piers where there is no danger of a settlement, though he would not ad-
40
ASSOCIATION OF ENGINEERING SOCIETIES.
vise the continuity to extend over so great a number of spans as to in- duce large strains on account of temperature changes.
The writer has attempted, by careful examination of the reports of tests at the Watertown Arsenal, to verify the experiments of M. Dupuy, but the examination has not proved altogether satisfactory, for the rea- son that the reports do not indicate precisely the point at which the first slip takes place. The stretch of the specimen is indicated at certain intervals, and it is easy to see that at some point the slipping takes place, but the points of observation are not sufficiently close together to indi- cate the exact point where the motion begins to occur.
Another reason why the experiments are not wholly satisfactory is that the bulk of the experiments have been tried upon a single row of rivets driven as in boiler shells, the rivets in most cases being much closer together than three diameters center to center, under which cir- cumstances the frictional action does not seem to work to so great an advantage.
The following tables give the average of a large number of tests, the strain indicated being the shear per rivet at the last observation point before the slip occurs :
TABLE I.
Earlier Tests of Rivets in Single Shear.
Force required to produce a slip in pounds.
|
Thickness of Metal. |
|||||
|
i |
3 8 |
i |
1 |
||
|
• |
r 7 16 1 |
1775 3810 |
|||
|
Iron Rivets . . . - |
1 1 16" 3. 4 15 16 . 1 |
3750 |
3904 |
5200 7000 |
8625 |
|
Steel Rivets . . . - |
1 11 TS |
4000 |
4333 |
||
|
I 4 |
5000 |
RIVETED JOINTS.
4]
TABLE II.
Later Tests of Iron Rivets in Double Shear. Force required to cause a slip in pounds.
|
Diameter of Rivet. |
Thickness of Metal. |
||||
|
l 4 |
3 8 |
h |
f |
3. 4 |
|
|
9 |
4012 |
4150 |
|||
|
l l 1 6 |
4000 |
5012 |
4525 |
||
|
13 1 6 |
3833 |
5250 |
6130 |
6300 |
|
|
1 5 T6 |
4400 |
4740 |
5400 |
7200 |
6700 |
It appears from these tests that, while the thickness of the metal against which the rivet bears plays little part in the frictional strength of the joint, it seems as if it is necessary to get a tolerable length of rivet before the clamping effect can be fully developed, for we see in no case an increase of strength at all proportional to the thickness. We even find for the thickest plates less strength in some cases than with thinner metal. It does, however, appear that with the thinnest metal, that is, where the rivet is very short, we fail to get quite so high results as with a little longer rivets.
The explanation of the fact that, in these tests, the larger diameters of rivet do not seem to be as efficacious as the smaller, may be partly due to the fact that the pitch of rivets is more suitable for the small than the large sizes, and they should not be considered as discrediting M. Dupuy's results on account of this disagreement, but it is exceed- ingly desirable that further experiments may be tried, to enable us to determine whether he is correct in assuming that the frictional strength is proportional to the area of the rivet section.
It is interesting to note that the behavior of the specimens under test was exactly the same as the French experiments showed. This is clearly described in the following extract from the report of Mr. J. E. Howard, C. E., which accompanies the tests quoted above :
" The progress of the test of a joint is generally marked by three well-defined periods. In the first period greatest rigidity is found, and it is thought that the joint is now held entirely by the friction of the rivet heads, and the movement of the joint is principally that due to the elasticity of the metal.
" The second period is distinguished by a rapid increase of stretch of the joint, attributed to the overcoming of the friction under the rivet heads and closing up any clearance about the rivets, bringing them into
42 ASSOCIATION OF ENGINEERING SOCIETIES.
bearing condition against the fronts of the rivet holes. Rivets, which are said to fill the holes, can hardly do so completely, on account of the contraction of the metal of the rivet from a higher temperature than that of the plate, after the rivet is driven.
" After a brief interval, the movement of the joint is retarded, and the third period is reached. The stretch of the joint is now believed to be due to the distortion of the rivet holes and the rivets themselves.
"The movement begins slowly, and so continues till the elastic limit of the metal about the rivet holes is passed, and general flow takes place over the entire cross-section, and rupture is reached."
If we, then, assume that the experiments, both here and in France, are sufficient to establish the fact that the first action of the rivets is to hold by friction, the question arises, whether it is not possible to adopt specifications based upon this force, which will give more perfect results than those attained by using bearing and shear.
A few of the objections are :
1. Want of tightness of the rivet. Of course, if the rivet is not tight it does not hold by friction, but this is no argument against using friction in proportioning the joints, because a loose rivet is bad anyway, and one such in an otherwise perfect joint cannot be said to do any good, whether the rivets are figured by bearing or shear.
2. The rivet may be driven through thicknesses which are not in perfect contact, and from the stiffness of the plate may appear to be tight when it is not really exerting as much pressure as the metal is capable of. This defect, while it certainly may be a serious one, espe- cially when hand-driven, could not exist if the thicknesses were properly clamped together during the process of driving, and it can only be said that it is as much a source of weakness, no matter how the joint was originally proportioned.
3. Another objection which occurs to the writer, is the doubt whether a joint may be lubricated by oil or paint applied to the contact- surface to such an extent as to appreciably alter the coefficient of fric- tion. As to this point, it is exceedingly desirable that more experi- ments may be tried, and possibly some of the members present can supply information on the point which the writer has not succeeded in finding. Without mentioning the fact that it is of very doubtful utility to apply paint or oil to surfaces which are to be riveted in close contact, it is the firm conviction of the writer that the heat of the rivet and the squeezing effect produced by the process of riveting entirely dissipate any effects from this cause.
Among the many advantages may be mentioned the following :
1. Simplicity. Evidently it would be much simpler if we could
neglect the thickness of plates and consider only one value for a single
shear of a certain-sized rivet.
RIVETED JOINTS. 43
2. Comparative accuracy. While we may not be able to deter- mine exactly the frictional strength of a rivet as long as it actually holds by friction, it does not seem as if we could gain much by considering two other functions, such as bearing and shear, which do not really act at all.
3. The doubtful and uncertain questions arising in case of indirect transmission would be entirely avoided.
4. When we realize that rivets hold by friction we free our minds at once from two complications which are apt to be troublesome in designing, viz., the questions of fatigue and reversal of strains, for neither of these has any effect upon a joint holding by friction.
5. A true conception of how rivets hold would prevent some details of construction which are now quite frequent. It is often convenient to drive two knees (between which a gusset or a beam web is to be inserted later) to some supporting members, the knees being purposely left a little wide apart to allow easy assembling. It is sometimes found con- venient to drive the flange rivets of a girder before the web rivets, and the writer has even known of the flange angles of a large girder being shop-riveted to the flange plates and shipped separate from the web plates, which were afterward inserted and hand-riveted. Such practices which destroy all possible frictional action, could not be tolerated if we counted on this force.
6. It may be well to note that in adopting this method of figuring we are running no risks, because in the few cases in which this method would allow a larger strain on the rivet than the old method, even though the friction should give out and the rivet should slip so that the bearing would come into play, we have abundant evidence that no disaster can result worse than a slight set in the joint.
As to the safe values to allow for the frictional resistance, it may simplify the consideration to think of the force as depending upon the clamping power of the rivet multiplied by the coefficient of friction.
The clamping effect is produced by the contraction of the rivet in cooling. If the pressure of the machine, or of the bolts in the case of hand-riveting, is sufficient to bring the thicknesses riveted into close con- tact and the rivet is properly heated, the contraction is greater than the possible elongation of the metal of the rivet within the elastic limit. The result is that the strain in the rivet is just enough to stretch it. M. Dupuy verified this fact by some very ingenious experiments, which are fully described in his paper above quoted. This being the case, we have one of the elements upon which the friction depends, determined with considerable precision. The other element, the coefficient, must neces- sarily be determined experimentally. Judging by experiments of Rennie, it would be in the neighborhood of T47 and considering the fact
44
ASSOCIATION OF ENGINEERING SOCIETIES.
that the pressure is very great, and that there is always a certain amount of unevenness around a hole which would tend to increase the co- efficient, it seems as if we could count on about this figure with reasonable certainty. This is confirmed by the result of Water town tests of rivets driven in slotted holes. In this case it was found that § rivets in single shear required a force of about 5,000 pounds to produce a slip. This is equivalent to a strain of 15,340 pounds per square inch, or probably a co- efficient of friction of T^. The elastic limit multiplied by -£} would give for steel about 12,000 pounds per square inch of rivet section, and for iron about 10,000 pounds, but these values are above the usual allow-
73
Fte. 3
oc
Fiff. 4
|
o |
ooooooooocb o o o\ |
|
|
~o |
a VS-HitfU 4s — (MMtcl*-— / |
|
|
o |
o |
/ |
|
o |
o |
/ |
|
o |
o |
|
|
o |
o |
\ |
|
o |
o |
\ |
|
oooooooooo oo o ^ |
Fiff. 5
|
1 ° |
0 |
o o o o o ct> |
o o |
|
|
o |
— |
-3-1/2-PHcIi )k- |
--.VPiuli |
.__.] |
|
o |
j |
|||
|
o |
/ |
|||
|
0 |
||||
|
o |
1 |
|||
|
o |
y |
|||
|
: o |
o |
o o o o o o |
o o |
° ) |
Fi.sr. (>
Sketches of Details Designed with and without Regard to Friction.
ances for shear. It seems, therefore, to the writer, that if we should only figure out rivets for shear, recognizing all the while that it is the friction which is holding and assigning safe shearing units, we should be approaching much nearer a truly economical and consistent design than by considering bearing.
This may be a step backward, but it is possible that in departing from the rules of our predecessors we have not gained as much as we think.
While it is not the object of this paper to go into the question of
RIVETED JOINTS. 45
a proper specification, it may be well to state that considering the question of friction, there ought to be more distinction made than is usually done between allowable shear on hand-driven and machine- driven rivets.
A few sketches are herewith presented showing ordinary bridge connections as they would appear designed in the common way and also by means of frictional resistance.
The sketches on the left, Figs. 1, 3 and 5, represent ordinary joints calculated for bearing and shear as in ordinary specifications. Those on the right, Figs\ 2, 4 and 6, show the same joints calculated upon the basis of frictional resistance.
As will be noticed, the number of rivets required is sometimes more and sometimes less, when the friction is taken into account. While on the average the number of rivets used may be about the same, it is thought that the distribution can be improved and more strength gained by considering the friction.
DISCUSSION.
By Edward S. Shaw. — The subject of the increase of strength of riveted joints under ordinary working strains, due to the friction of the contact-surfaces of the plates or shapes riveted together, and the ques- tion of how much of the strength of the joint is due to this friction, are interesting matters, which doubtless have not heretofore had the consideration that they deserve from engineering writers and ex- perimenters.
The existence of this frictional resistance has long been known, but the older authors of engineering text-books, so far as the knowledge and memory of the writer goes, dismissed the subject in much the same way as Trautwine, who says : " The friction between the plates in a lap, or between the plates and the covers in a butt, produced by their being pressed tightly together by the contraction of the rivets in cooling, adds much to the strength of a joint while new, perhaps as much as 1.5 to 3 tons per square inch of circular section of all the rivets in a lap, or of all on one side of a single-cover butt ; or 3 to 6 tons of all on one side of a double-cover butt. In quiet structures, this friction might con- tinue to exist, either wholly or in part, for an indefinite period ; but in bridges, etc., subject to violent and incessant jarring and tremor, it is probably soon diminished or entirely dissipated. Hence good author- ities recommend not to rely on it, and it is therefore omitted in what follows."
If we admit the universal existence of this friction as a factor, even
46 ASSOCIATION OF ENGINEERING SOCIETIES.
if not the controlling one, in determining the number and strength of the rivets, it would seem also to have an appreciable effect upon the net strength of the plates connected by splices in a riveted bridge joint, for beyond the last rivet or last line of rivets in the joint, the splice plates extend, pressed to the main plates with a certain strength of grip which must be overcome before the plate can tear on the outer rivet holes.
Mr. Worcester should have the thanks of all engineers interested in bridge designing and construction, for the careful and able manner in which he has collected the somewhat meager data existing upon this subject and drawn general conclusions therefrom.
It is to be hoped that he will continue his investigations so far as to present a complete and rational method, at least for the spacing of rivets in the webs and flanges of plate girder bridges.
In regard to the general application of this method, there would seem to be more than a few difficulties, especially in connection with field-riveting.
We have learned to regard the proper upsetting of the rivet, and filling the hole as well as possible, as of utmost importance. It is, however, too much to expect of the average field-riveting gang, that they will get perfect heads on all their rivets, and make them tight and well upset in the holes at the same time ; and it has been the practice of the writer to accept a limited number of unsatisfactory rivet heads, especially in difficult positions for driving, provided that the rivets were otherwise well driven.
The substitution of the author's method would require a reversal of this rule, for if the plates are to be held together by the grip of the rivet heads, it is essential that there should be no small, flat or thin- edged or badly eccentric rivet heads, for though at first the grip may be sufficient with a flat head or small bearing surface, yet rust and other causes are very liable to diminish or destroy this surface and grip.
The writer must express views in opposition to the fourth conclusion of the author, viz. : " When we realize that rivets hold by friction, we free our minds at once from two complications which are apt to be trouble- some in designing, viz. : the question of fatigue and reversal of strains, for neither of these has any effect upon a joint holding by friction."
In contradiction to this may be placed the old theory, well ex- pressed by the statement of Trautwine quoted above in regard to the effects of impact and vibration ("jarring and tremor"), the same causes, or results of the same causes, which are supposed to produce fatigue in the main members and which are operative in necessitating a diminu- tion of the allowable strains, with increase of live load or reversal of strain. The rivets being in a high state of tension, (according to the French experimenter quoted by the author, beyond the primary elastic
RIVETED JOINTS. 47
limit of the rivet material) it would seem to the writer that they must be especially sensitive to the fatiguing effects of dynamic influences acting upon them through the media of the plates or parts connected, while the action upon the rivets caused by the movements of extension or compression of the members connected, however slight, will apparently, in time, tend to diminish their grip and the resulting friction, in about the same ratio that the microscopic molecular changes of deterioration or fatigue in the maiu members are produced.
Therefore, if it is proper to vary the unit strains in the principal members with variations in the ratio of minimum to maximum, then it would seem to be equally proper and thoroughly consistent to vary them in a proportional or similar manner in the rivets.
By John C. Moses. — The so-called " factor of safety " has frequently been termed a " factor of ignorance," and the writer of the paper has clearly shown the truthfulness of the charge in the case under discus- sion. But no engineer can really be satisfied with factors of ignorance, and so we find him constantly trying to eliminate the guess work and substitute what is known to be correct. The most satisfactory method is to make experiments and reason from them as a basis. It is positively appalling to think of the amount of mental labor involved in arriving at our present theories of the action of structures — mental labor that a few actual experiments would so greatly assist. These experiments are not made because, on the one hand, no manufacturer is sufficiently dis- interested to do a work that will not bring him a direct gain unshared by others ; and, on the other hand, no purchaser feels called upon to pay for experiments that will benefit everybody as well as himself. No one questions the need, but they say it benefits everyone, and so everyone should help pay for it. Consequently it is not done. The writer speaks of the possible effects of paint as a lubricant of riveted joints — he thinks its effect is inappreciable, but perhaps ten years from now someone will spend five hundred dollars and find out. We have waited many years for the experiments on riveted joints in tension cited in the paper. The French experiments were made possible by government aid, but we cannot expect that in this country.
A recent article in the Engineering News demonstrated that the standard beam connections in common use are only worth from one- fourth to one-half the commonly assumed values " if the effect of friction be neglected." We all know this, but we also know that rivets have a value aside from their shearing and bearing values, and we guess that it is enough to counteract the effects of eccentricity. This is a very crude approximation. A hundred beams connected as in actual practice, and loaded until the elastic limits of the joints were reached, would give us a new basis for our theories, and probably enable us to save a pound or two
48 ASSOCIATION OF ENGINEERING SOCIETIES.
of metal in a joint, while at the same time making it a stronger piece of work. To accomplish a better result with less money is to be an engi- neer in the truest sense. How long would it take to pay for the experi- ments if an average of one pound a joint could be saved as the result ? Meantime the profession is daily making a ridiculous assumption, be- cause no one makes the experiments. What better work could be found for an engineering society to do ? Manufacturers can be interested if some one will take the initiative. Testing machines and men to run them can be found in Cambridge and Boston, and the cost of the mate- rials need not be great. Our last president has suggested dividing our Society into groups, each one in turn furnishing the paper of an even- ing. The report of the Committee on Standard Connections for Beams would make an interesting subject for some of us, and if properly done it would be a distinct addition to engineering knowledge and a new demonstration of the usefulness of this Society. There are, of course, many other questions awaiting similar solutions ; this one is merely sug- gested as an example, the solution of which is indicated in the evening's paper. One cannot help thinking, as he reads that paper, how much more satisfactory it would be if the author had been able to base it on more extensive experimental data. If he could have positively told us that it was correct to use the shearing values of rivets when spacing them in flanges of girders, the information would save several per cent, of the cost of the work done in the establishment with which the writer is connected. As it is we will feel safe to use a somewhat higher bear- ing value for rivets than has been our custom heretofore.
By James E. Howard. — That riveted plates are held to- gether by a very substantial gripping pressure exerted by the rivets in cooling, there can be little doubt. A consideration of the coefficient of expansion by heat shows that rivets need cool only over a limited range of temperature in order to be strained to their elastic limit. This zone of temperature is of such limited range that the known changes in the modulus of elasticity and elastic limit under higher temperatures has but slight influence in the results, and it is believed that under favorable conditions rivets in their final state may be left gripping the plates of a joint with a force nearly or quite coincident with their elastic limits.
In hydraulic riveting the conditions seem favorable for reaching good results. Rivets in their upset state may, however, have a lower elastic limit in consequence of the upsetting than possessed by the metal in a rolled bar.
That rivets are strained nearly to their elastic limit the tests of the joints seem to afford some proof.
In the early stages of a test it is frequently observed that the scale starts off the rivet heads, and when this occurs under a comparatively
RIVETED JOINTS. 49
low stress on the joint it is taken to signify that the rivets were nearly ready to scale when the joint was at rest.
A critical comparison of the strains developed in a joint and those which should be developed in the solid plate, adopting a modulus of elasticity of 30,000,000 pounds per square inch, shows that joints very frequently elongate more than a solid plate should elongate, and that permanent sets appear early in the joint.
So far as these minute changes in shape are at present explainable it would appear that while in the main frictional resistance prevents general slipping of the plates, yet slight distortions are permitted to occur.
It is difficult to ascribe an adequate cause for this behavior. Per- haps it signifies the release of internal strains, where conflicting strains existed due to the contraction of the parts while cooling after riveting.
Possibly the most advantageous case for maximum frictional re- sistance is found in a joint containing a single row of rivets, running at right augles to the line of applied stresses.
Where several rows of rivets are used, the outer rows necessarily are obliged to permit some movement of the joint in order to bring the in- side rows into action. Just why Mr. Worcester considers experiments made in a single row of rivets unsatisfactory (see page 40) does not appear to be explained, nor why the spacing should exceed three diam- eters for the frictional resistance to work to advantage. Data seems needed to show what part of the maximum frictional resistance existing or supposed to exist in a joint is available for use in a structure. Per- haps an experimental inquiry into the behavior of joints exposed to al- ternate and variable stresses and also vibratory influences would aid in answering this question.
By Gaetano Lanza. — (1) The chief objection to Mr. Worcester's theory is, of course, the fact which Mr. Howard has already mentioned, that the stretch due to the application of the load on a riveted joint is always greater than that due to the stretch of the metal, and hence that there is no load which does not cause slipping.
(2) Moreover, if, as I suppose Mr. Worcester must have tried to do, we seek for a load at which the rate of slipping increases very decidedly, we shall find, in many cases, that the increase in the rate of slipping is gradual.
(3) Were the action such as Mr. Worcester describes, it would be a dangerous proceeding to design joints on the basis of the friction alone, without considering their strength, as it would be possible, on that theory, to make a joint having more frictional resistance than strength, or, at least, where the frictional resistance formed a large proportion of the strength and the factor of safety was very small.
4
50 ASSOCIATION OF ENGINEERING SOCIETIES.
(4) The experiments of Dupuy seem to me to be too few in number, and also, for the most part, to have been made on joints with too few rivets to warrant the conclusions drawn. The determinations of stretch were much coarser than those of the tests at Watertown arsenal, and the fourth group of joints were tested on a machine in which piston friction was allowed to vitiate the results.
Of the other three groups, the first was the only one where there were as many as four rivets on each side of the joint, and these were hand-riveted; the number of tests being four. In the second group of seven joints there was only one rivet on each side of the joint, and in the third group of twelve joints there were only two rivets on each side of the joint.
By J. P. Snow. — It seems to me that the web rivets in plate girder flanges get the advantage of frictional resistance more than those in almost any other situation. If the web tends to fail by bearing, that is if it starts to buckle around the rivet hole, the angles held by the flange plates prevent it and this action tends to increase the friction rather than to diminish it, as would be the case in a simple connection where the rivets act in single shear. Oftentimes, when designing under the usual specifications, the thickness of webs is governed by the rivet bear- ing near the ends. I think that the usual unit strains might be safely exceeded in these cases. So far as I have been able to judge from observing old girders, the thickness of webs is a part that may be allowed to vary more widely from established units than other members of the structure, and it is hardly conceivable that one could be so badly designed as to fail from the insufficiency of flange rivets. It must be the friction that helps in these cases, and I believe that it is legitimate to take advantage of what we know must exist in proportioning new work, although I acknowledge I have never had the courage to raise these units appreciably.
It is troublesome to arrange a general specification so that it will cover extreme cases without running into absurd results ; this makes it difficult to provide for all the varying cases where it seems advisable to depend wholly or partly on the friction. The cautionary legend that Mr. Cooper puts at the head of his Standard Specification is well placed ; much depends on the judgment of the designer. It is, however, feasible to provide for a different unit for hand-driven and machine-driven rivets and for rivets in single and enclosed bearing. The practice on the Bos- ton and Maine Railroad is to allow 25 percent, more on machine-driven than on hand-driven rivets aud 25 per cent, more on rivets enclosed between two thicknesses than on those acting in single shear.
I heartily agree with the paper in the matter of making these allowances, and in crediting the friction with a generous ratio of its
EIVETED JOINTS. 51
ultimate value in the case of web rivets in plate girders. In cases of connections where the rivets act in single shear, however, especially when they are hand-driven, it will hardly do to allow more on a rivet than the bearing area or shear could properly carry, because these surely are the ultimate dependence if the rivets from any cause become loose. The examples shown by the author call for more rivets in this class of connections than the usual rule. How this would work out for different thicknesses cannot be told without a definite specification. It is prob- able that in thin metal less rivets would be called for by the new rule than the old. The usual number should be reduced with caution, for although it may be that rivets do not get loose much worse in thin metal than in thick, yet after a rivet is loose the thin metal goes to destruc- tion much the faster, and it is hardly to be expected that any set of rules will prevent rivets getting loose occasionally. This is off the ques- tion somewhat perhaps because the paper considers " proper rivets only " aud those which become loose cannot be called proper; but we must deal with actual conditions and try to design work to meet them.
In the structures under consideration we must unfortunately de- pend on hand-driven rivets and generally on those acting in single shear to perform the most important function of all, that is, to connect the various members together ; while the rivets of less importance, that is, those holding together the integral parts of a member, can be machine- driven. This condition operates in several ways to the disadvantage of riveted trusses. Being so vitally important, and at the same time of so low efficiency, we must have plenty of them. They are what hold our structures together and we must not depend too much on so uncertain an element as friction.
As to the superior rigidity of riveted trusses alluded to in the paper, it is probable that the form of section has as much to do with it as the style of connection. The flanged sections in pin trusses vibrate but little, while flat bars in riveted ones are but little more rigid than if connected by pins. I think if satisfactory joints could be arranged in pin-connected trusses when using flanged sections like angles, zees and channels throughout, the resulting structure would be as satisfactory in the matter of rigidity as it would be if riveted.
It is very interesting to compare European practice with ours, as is done by the author. On the whole it seems to me that the Americans have been the most ready to drop what is bad in their past practice, and to adopt the good from foreign design. In our solid-floor bridges, we have certainly improved on the English designs, and in our long-panel trusses we have dropped most of the objectionable features of early American flimsiness. The Europeans seem to cling to their clumsy short panels. Some American designers claim that long panels only are right, that
52 ASSOCIATION OF ENGINEERING SOCIETIES.
panels less than 20 feet should be frowned upon. It does not seem to me that the length of panel jier se affects the efficiency of the bridge at all. This element as well as the style of the bridge should be adjusted to meet the conditions. Within twelve months we have designed bridges for the Boston and Maine Railroad with panels varying from 2 feet to 23 feet, and so far as I know they are equally strong and service- able. It was surely our endeavor to make them exactly equivalent. Our practice is equally varied in the matter of pin and riveted work. I believe that each has its place. He who will use both and can keep each in its proper field is certainly employing the art of bridge building to the best advantage.
By Joseph R. Worcester. — The interesting questions which have been raised by the discussions of the paper have shed light on a number of points which have not been sufficiently elaborated in the paper itself, and the author is very glad to acknowledge that many of the criticisms are well founded. There have, however, been a few points raised which perhaps can be explained.
With regard to Mr. Shaw's points about field rivets it appears to the author that the danger in this class of work is not so much from imperfect heads as from the fact that the holes are not completely filled. If such be the case, unless we are willing to allow considerable slip when the load is applied, it is essential that we should provide rivets enough to hold by friction. The head must be very bad if it is not sufficiently enlarged to grip the plates together when the rivet cools. In the author's opinion it is better to use bolts where good rivets cannot be driven, for a bolt well drawn up is much better than a loose rivet: one loose rivet in a joint is absolutely useless until the others have yielded.
So far as the action of repeated or reversed strains affect the question of friction, if we have no motion between the surfaces clamped together, we have a constant straiu in the rivet. It is not apparent, then, how the rivet can be weakened by any number of variations of stress.
Of course, if there is any motion the strength becomes less with each slip and we soon have a loose joint. It is therefore all the more important to provide rivets enough at the start to prevent motion.
Mr. Howard's careful consideration of the effect of the friction is very instructive, and his explanation of the variation of the elongation during the earlier stages of the tests from that which would occur in solid metal appears quite reasonable. The author's only reason for suggesting that a less spacing than three diameters may not give as good results as a greater is from observation of the large number of Water- town tests made with all kinds of spacing. It certainly seems necessary that more tests should be made to definitely determine this point.
RIVETED JOINTS. 53
While the author agrees with Prof. Lanza that it is somewhat dif- ficult to determine from the Watertown tests the point where a decided slip takes place, in most cases it is very evident that there is such a point, as may be seen by carefully plotting the extensions. The author cannot agree as to there being any danger in considering the friction in the method suggested in the paper, and it is not easy to see what is meant by a joint having more frictional resistance than strength. It is not, of course, proposed to govern the thickness of the plates or parts connected by the frictional resistance of the joint, as this element is in- variably settled by other considerations.
With regard to the experiments of M. Dupuy, while the author does not quote them as being conclusive they seem to be particularly instructive on account of the thoroughness with which the minute elon- gations were observed under light strains and also on account of the care exercised in determining the physical conditions of the rivets.
The remarks of Mr. Moses with regard to the difficulty of securing an adequate number of tests must be apparent to all. It is very much to be hoped that in the future this subject will not be so neglected as it has been in the past.
The author is pleased to note the extent to which Mr. Snow has been willing to consider friction in his own practice.
In conclusion the author cannot but repeat what he has stated in the paper that while in good practice, that is, in work that is not strained above one-half its elastic limit, practically all riveted joints are held by the friction alone, it certainly seems wrong to consider as the basis of strength, elements, such as bearing, which cannot possibly come into play until the joint has yielded.
54 ASSOCIATION OF ENGINEERING SOCIETIES.
A LOW CRIB DAM ACROSS ROCK RIVER.
By J. W. Woermann, Member of the Engineers' Club of St. Loris.
[Read before the Club, May 6, 1896.*]
The dam which is herein described constitutes one of the structures of the Illinois and Mississippi Canal, better known as the Hennepin Canal, and was built for the purpose of furnishing slack water naviga- tion in Rock River above the Lower Rapids, which are located at Milan, 111. The location and general engineering features of this canal were briefly described in a paper devoted to the concrete construction on the same, read before the Club about two years ago, and published in the journal of the Association for November, 1894. Since that time the four and one-half miles of caual around the Lower Rapids, together with about eight miles of slack water above the dams, or about thirteen miles in all, have been opened to navigation, and some coal traffic is being developed. During the same time eight miles have been com- pleted on the Eastern Section, immediately adjoining the Illinois River, including the masonry for seven locks, one aqueduct and a number of minor structures. The concrete abutments for this dam or dams, as there are two sections of it, one or either side of Carr's Island, were described in the former paper, so that only the cribwork or dam proper will be considered at the present time.
LOCATION.
The determination of the best location for the dams involved a com- 'plete survey of this vicinity, including not only ordinary soundings over a considerable stretch of the river, and the taking of topography along the shores, but also the preparation of profiles, showing the elevation of bed-rock at the most feasible sites. The topography along the shores was necessary, as some of the locations required a greater or less extent of levee to protect the adjoining lands from overflow.
The location finally selected requires a dam across the south channel of Rock River, at the head of Carr's Island, 764.2 feet in length, and another dam across the north channel 598.3 feet in length, about 800 feet below the head of the island, giving a combined length of 1,362.5 feet of crest. Connecting the two dams is a levee about 1,000 feet long to protect the island from overflow at high-water.
* Manuscript received July 20, 1896.— Secretary, Ass'n of Eng. Socs.
A LOW CRIB DAM ACROSS ROCK RIVER. 55
DESIGN.
The general design of the dam was prepared by Major W. L. Marshall, Corps of Engineers, U. S. A., the general plan of which can be most readily seen by a glance at Plate VI accompanying this paper. It was designed for a rock foundation to withstand a maxi- mum head of four and a half feet, and may be described as a rock-filled crib, the woodwork consisting principally of six-inch by eight-inch pine timbers laid flatwise. The main part of the dam is thirteen and a half feet in width and the apron six and one-half feet, making the total width of the base twenty feet. Immediately adjoining the cribwork above is a filling of clay and quarry refuse of about the same width as the crib- work, and rising in height to the top of the sheet piling.
The main dam and the apron are both covered with four-inch oak plank, and the upstream face of the dam with two rows of two-inch pine sheet piling. The oak plank on the main dam are closely fitted together, making it practically water-tight, so that the vertical pressure of the water above the coping is added to the weight of the material in the dam in giving increased stability to the structure.
The coping of the main part of the dam is built ou a slope, rising two feet from the sheet piling to the crest, with a view of preventing projecting limbs and other irregular objects from getting caught on the up-stream face and pounding more or less upon the coping, as is usually the case where the slope is in the opposite direction.
From the crest of the dam to the apron is a fall of three feet. The top of the apron is about six inches above extreme low-water, but at the stage at which the ice usually goes out it is covered with water, more or less, forming a cushion which prevents the ice from cutting the apron as it otherwise would. In the spring of 1895 the ice went out at an unusually low stage, with a thickness of six to twelve inches, but did no damage. The six-inch by eight-inch transverse pine timbers, projecting two inches above the level of the apron, were cut down to the level of the oak plank to some extent, but this was anticipated and was not con- sidered important.
NORTH COFFERDAM.
The construction of the dams was in charge of Mr. L. L. Wheeler, M. Am. Soc. C. E., as Resident Engineer, with the writer as principal assistant, and Mr. Geo. T. McGee as instrumentman. The best form of cofferdam to use in shutting off the river was a matter of considerable investigation, and the contingencies and probable cost were estimated for several different styles. Inasmuch as Rock River is usually at a low stage during the summer, and as the spring of 1894 had been unusually dry, it was decided to build a simple earth embankment across each
56 ASSOCIATION OF ENGINEERING SOCIETIES.
channel with riprapping on the up-stream side to protect them from wave- wash. This plan received greater favor also from the fact that the cofferdam around the guard lock had heen successfully built in this manner, and because the bed of the river was of such a nature that teams could be driven over it wherever the depth was less than three feet. As the north channel contains the deepest water, the greatest depth at that stage being about five feet, it was decided to build that dam first, and leave half of it incomplete to serve as a temporary sluice- way during the construction of the South Dam.
An area below the north abutment of this dam was stripped in April to furnish a site for a quarry, and early in June the construction of the cofferdam was commenced. The stripping from the quarry, together with the quarry refuse, formed the body of the cofferdam, while a ridge of riprap was kept in advance on the lower side to prevent the current from washing away the loose earth. While the embankment was being started from the north shore of the river, five cribs, sixteen feet square, were settled in line adjacent to Carr's Island. The cribs were built in the shallow water near the shore, by simply boring a hole in each end of each timber and dropping them over the loug bolts which held the timbers together at each corner of the crib. The cribs were placed fourteen feet apart, the top covered with four-inch oak plank, and weighted down with rock and bags of sand. Six by eight-inch timbers, the ends of which were supported by the cribs, were then shoved down into the water, furnishing a length of about 130 feet to sustain the rip- rap from being carried away by the current. The riprap and earth cofferdam was then extended to the island, above this protection, and the flow completely shut off. The weight was then taken off the cribs and the lumber used in the construction of the permanent dam. Sub- sequently the riprap was removed and used for filling in the permanent structure, — this recovery of the riprap being the principal argument in favor of placing the stone on the down-stream side of this cofferdam.
The end of the cofferdam was kept about two feet above the surface of the water, the wagons being dumped while they stood on the steep slope at the end. A small amount of riprap was placed on the up-stream side, above the water-line, to protect the embankment from wave- wash. The teams returned to the shore by driving through the water on the upper side of the cofferdam. On account of the south channel remaining entirely open, the construction of this cofferdam only raised the water surface about four inches. The foot of the island was far enough down-stream to keep the backwater from coming up and inter- fering with the work. Low secondary cofferdams,' a few inches in height, were then built below the main cofferdam, to exclude the seepage that came from the latter. The areas enclosed were from fifty to two hun-
A LOW CRIB DAM ACROSS ROCK RIVER. 57
dred feet in length, according to the irregularity of the bottom, and were kept dry by means of hand-pumps. The entire amount of material in this cofferdam, including what was in the secondary dams, was about 300 cubic yards of riprap and 800 cubic yards of quarry stripping.
FOUNDATION.
All sand and gravel, together with as much of the bed-rock as could be readily raised with a pick, were then cleared away and the construction of the cribwork commenced. Where the rock was com- paratively smooth and solid, iron anchor bolts were set in cement, in holes drilled for the purpose, to which the foundation timbers were bolted so as to insure a greater factor of safety against sliding. The bolts were usually one and one-eighth inches in diameter and twenty- four inches long, but where the bottom was less firm longer bolts were used. Two bolts were generally placed in each panel, the spacing depending on whether the first course consisted of longitudinal or trans- verse timbers. Where the rock was loose enough to permit a trench to be picked out, six inches or more in depth, for the base of the dam, the anchor bolts were considered unnecessary. The largest pocket of clay was ten feet across, and was excavated to a depth of six feet below the river bottom before starting the cribwork. Over the three largest pockets the apron was made fourteen feet in length instead of the regular seven-foot length.
CRIBWORK.
All the timbers in the dam were six by eight inches except the top timber on the up-stream face and the top timber under the crest, which were eight inches by eight inches, and eight inches by ten inches, respectively. All of the longitudinal timbers were sixteen feet in length ' and were arranged so as to break joint regularly and to bring the joints within two feet of the middle of the panels. The bottom course con- sisted of six rows of timbers, so as to furnish more area for the support of the rock filling, and each of the succeeding courses only five, up to the level of the apron.
In laying the bottom timbers readings were taken on them at fre- quent intervals with a wye level so as to insure their being started at the proper grade. Where the bottom was comparatively regular the carpenters extended the work several panels at a time with common spirit levels. Throughout the work the bottom timbers were adzed to fit the irregularities in the rock so as to insure greater safety against sliding. Readings were again taken on the cross- timbers at the level of the apron, and on the eight-foot blocks near the top, and wher- ever the elevations exceeded the proper grades by more than a half inch
58 ASSOCIATION OF ENGINEERING SOCIETIES.
the course next following was notched down accordingly. The differ- ences in elevation were caused maiuly by the variation in thickness of the timbers. The lumber being only regular commercial stock, the thickness varied all the way from five and a half to six inches. On this account also the bottom timbers were started about two-tenths of a foot above the elevation shown on the plan so that allowance could be made for this deficiency. An absolutely level crest at an exact grade, how- ever, was not considered of sufficient importance to warrant much addi- tional expense in notching down timbers, so that no great refinement was attempted in this direction. The highest and lowest points on the crest of the North Dam were 130.04 and 129.91 respectively, and for the South Dam 130.59 and 130.46 respectively. The maximum dif- ference in each case is thirteen-hundredths of a foot, and the mean ele- vation determined from readings taken every sixteen feet are 129.992 and 130.530, Hennepin Datum.
From the level of. the apron upward the transverse timbers are of different lengths in order to properly support and reinforce the purlins, as shown on Plates V and VI. The purlins were five in number, spaced about three feet and three inches between centers. The use of the tem- plates in marking the gains to receive the purlins is shown graphically on Plate VI and requires no further explanation.
The transverse timbers are all spaced eight feet apart except that in the course above the bottom longitudinals an extra timber fourteen feet long is placed in the middle of each eight foot panel, so as to assist in receiving the weight of the rock filling. On the down -stream face of the dam, under the apron as well as under the coping, a two-foot block is placed under each joint, to which the longitudinals are thoroughly bolted. The intention was to increase the tensile strength on that side, so that in case any part of the dam should ever be called upon to act as a beam, it would have proportionately greater transverse strength.
DRIFT BOLTS AND SPIKES.
The method of drift bolting can be seen most readily by glancing again at Plate VI. The size of all the drift bolts was three-quarter inch by sixteen inch except that ten-inch bolts were used at the bottom wherever the first course happened to consist of longitudinals, and eighteen-inch in putting on the eight by eight-inch and eight by ten- inch longitudinals. One bolt was driven at each intersection, the bolt being always started through a cross-timber. Holes were bored to the full depth of the bolts one-sixteenth inch smaller in diameter.
The four-inch oak planking on the coping and apron was fastened down with two seven-sixteenths inch by eight-inch boat spikes at each purlin, for which seven-sixteenths inch holes were bored. The first row
A LOW CRIB DAM ACROSS ROCK RIVER. 59
of two-inch sheet piling was held temporarily with 20d wire nails until the second row was put on, both rows being carried along together, after which they were fastened permanently with four three- eighths inch by seven-inch boat spikes per running foot, driven without boring.
ROCK FILLING.
The filling of the dam was carried on simultaneously with the crib- work, and the stone packed in between the timbers so as to obtain as much weight as possible. The filling immediately adjoining the crib- work on the up-stream side was carried along in a narrow embankment as fast as the sheet piling was put on. This permitted the rock teams to be driven up close to the sheet piling, and made it possible to throw the stone directly into the cribwork from the wagons. When the dam was nearly in the condition in which it was to be left during the con- struction of the South Dam, this embankment was widened by casting over material from the cofferdam, until the latter was finally allowed to break through.
TEMPORARY SLUICEWAY.
The temporary sluiceway in the North Dam, previously referred to, kept the water above the dams about two feet lower than if the dam had all been completed at one time, — and allowed the South Cofferdam to be made lower by the same amount. The saving by this plan must not be measured directly by the difference in the volume of the South Coffer- dam under the two conditions, but by the fact that the average depth of water in which it was necessary to work would have been increased from about two feet to four feet, in case the sluiceway had not been used.
A temporary bracket consisting of a vertical post and two braces was set up at each panel point, as shown on Plate V. On the up- stream side of the posts, near the top, was supported a line of four- inch by twelve-inch pine wales. This brought the wales on line with the permanent sheet piling, which formed the bottom of the sluiceway. In shutting off the sluiceway subsequently, in order to complete this portion of the dam, it was necessary simply to shove the sheet piling down in front of the wale, and allow them to catch on the top of the permanent sheet piling. The vertical posts and the long braces were used in the completion of that half of the dam, and the balance of the lumber on other parts of the canal, so that there was no waste of lumber. Dur- ing the construction of the South Dam this sluiceway carried the whole discharge of the river, amounting to about 2,500 cubic feet a second.
SOUTH DAM.
The construction of the South Cofferdam was commenced on July 24th, and completed on August 4th. In this case the ridge of riprap was
60 ASSOCIATION OF ENGINEERING SOCIETIES.
extended ahead of the earthen portion on the up-stream side, so as to per- mit the teams to return to the shore on the lower side, as the increased depth, caused by shutting off the water completely, would not permit the same plan to be used as at the North Dam. On account of the greater amount of water with which it was necessary to contend in making the final closure, ten cribs were erected adjacent to the Island instead of five. These were settled on the up-stream edge of the coffer- dam. When the end of the cofferdam had been extended so as to come under the protection of the first crib, wales were placed across the spaces between the cribs and driving sheet piling was commenced from both ends of this three-hundred foot space covered by the cribs. Under this protection work on the rock and earth cofferdam was also carried on from both ends. As the ends of the cofferdam were extended part of the sheet piling was taken up and used a second time, and only in mak- ing the final closure for the last hundred feet was it necessary to double the piling. The method of dumping the wagons and the final closing behind the cribs are shown on Plate I.
When the comparatively tight embankment had been completed the weight was taken off the cribs and the lumber used in the perma- nent structure. Most of the clay and riprap for this cofferdam were taken from a waste pile of material that had been excavated from the lock pit at the south end of the dam.
As indicated by the borings, it was found that the foundation of this dam was not as good as the other, inasmuch as a stretch of hard clay was encountered 50 feet in length, and a bed of compact sand and gravel 120 feet in length. This was excavated to a depth of about three feet below the bed of the river, aud the material used as filling above the dam. As the cribwork progressed, the V-shaped trench that remained below the apron was filled with heavy stone, as shown on Plate V.
The carpenter work on this dam was commenced on August 7th and completed August 22-1, sixteen days from the time the first timber was laid. The Federal labor law, passed by Congress in 1892, prohibited us from working more than eight hours a day, without putting on a second shift, which was considered impracticable, but it did not prevent us from working on Sundays.
The construction of this cofferdam raised the water surface about two feet higher than it was during the building of the North Dam, and in order to provide a sluiceway for carrying the seepage from the coffer- dam through the permanent dam, the sheet piling was omitted from one panel until the filling above the dam was all completed. The coffer- dam was allowed to break through on August 24th, exactly one month from the time the cofferdam was commenced.
The rock filling for this dam was obtained considerably cheaper than
EtRING SOC
|
i ; i |
||
|
■71: _ |
||
|
1 1- |
3 □
Elevation of D
JOURNAL OF THE ASSOCIATION OF ENGINEERING SOCIETIES.
J. W. WOERWIANN-CRIB DAM.
i:U-v;ili(»ii of Kowii -St renin Side.
Aiu-lior Boll,
.' -Iir. ( ClLliiK. Double
|
— -- -rx — - — t|TT "f i! 1 1*1 |
|
|
III | Hi ( L.| || I1 |
|
|
P«l 1 ii !l lip llljl lljl |
|
|
-in iiii in i| i jji ij i ' |
|
A LOW CRIB DAM ACEOSS ROCK RIVER. 61
for the other from the fact that on this side about 75 per cent, of the rock was readily quarried from the bed of the river without explosives.
FORCE EMPLOYED.
During the construction of the South Cofferdam the force consisted of about fourteen teams and fifty laborers. For a few days while the preparation of the foundation was being rushed the number of laborers was increased to 130.
During the erection of the cribwork the force consisted of sixteen carpenters, and about fifty laborers, about one-third of the latter assist- ing the carpenters in carrying timbers, boring, and driving bolts and spikes. The number of teams remained the same throughout the work. The appearance of this dam during construction is shown on Plate II, while the view presented with the water running over it, after both dams were completed, is shown on Plate IV.
COMPLETING NORTH DAM.
After the completion of the South Dam the temporary sluiceway in the North Dam was closed as previously described, and the upper part of the dam completed readily under the protection of the sheet piling. As soon as the sheet piling was in place the lower braces were knocked out, so that there was nothing to interfere with putting on the purlins. As the coping was gradually extended from the abutment; braces were put in from the waling-piece back to the top of the oak plank. This permitted the post and the remaining brace to be knocked out, as the spikes in the sheet piling were sufficient to support the weight of the waling-piece. . The stone filling was thrown into the crib from a barge which was towed alongside the sheet piling. The completion of this part of the dam occupied four and a half days, including the erection and removal of the sheet piling. Plate III was taken during the com- pletion of this portion of the dam.
FISHWAYS.
At the time the dams were built a fish way was constructed at the south end of each dam. It was found, however, that they were unsatis- factory for several reasons, and during the following summer they were modified according to the plan shown on Plate V. By increasing the number of wings from five to nine the velocity of the water was checked so that fish can readily ascend from step to step. The upper end is arranged so that the fish go out into comparatively quiet water, instead of having to jump over the crest, while at the same time the amount of water entering the fishway can be regulated to suit the stage of the river. They also comply with the theory that a fishway, in order to be
62 ASSOCIATION OF ENGINEERING SOCIETIES.
found readily by the fish, should uot extend down-stream any farther than the apron of the dam. The cribs above the fish ways, together with their protected position adjacent to the south abutments, are de- signed to protect them from ice. When the fish are running up-stream the larger ones can frequently be seen entering and leaving the fish- ways. By shutting off the water at the upper end, as many as sixty fish of various species have been found in it at one time, some of which have been from two to three feet in length.
TOTAL AMOUNT OF MATERIAL IN DAMS.
North Dam. South Dam.
Ft. B. 31. Ft. B. 31.
Longitudinal timbers 47,230 73,550
Transverse " 28,350 46,950
Sheet piling 7,950 14,610
Total pine lumber 83,530 135,110
Oak plank in coping 33,540 42,840
" " " apron 15,870 19,300
Total oak lumber 49,410 62,140
Total oak and pine lumber ........ 132,940 197,250
The total amount of lumber in both dams is 330,190 feet, B. M. The cost of the labor expended in putting this in the dams amounted to $1,914, or $5.80 per M.
The total amount of rock filling in the North Dam is 1,240 cubic yards, and in the South Dam 2,350 cubic yards, making the total for both dams 3,590 cubic yards.
AMOUNT OF IRON IN DAMS.
North Dam. South Dam.
Anchor bolts 1,010 pounds. 320 pounds.
Drift " 6,050 " . 9,610 "
Boat spikes 4,750 " 6.050 "
Wire nails 300 " 400
Total 12,110 pounds. 16,380 pounds.
COST OF LABOR ON EACH DAM.
N. Dam. S. Dam.
Hualing material $283 97
Building cofferdam $729 55 1,055 34
Preparing foundation 493 30 818 04
Carpenter work on dams 948 92 964 86
Quarrying rock, filling cribs, and grading above dams . . . 1,965 54 1,970 56
Engineering, watching, and miscellaneous 362 25 402 47
Total $4,499 56 $5,495 24
A LOW CEIB DAM ACROSS ROCK RIVER. 63
making the total cost of the labor on both dams practically ten thousand dollars.
TOTAL COST OF THE. TWO DAMS.
The total cost of the two dams, including labor and material, is as
follows :
Rent of land '. $ 217 40
Labor 9,994 80
Oak lumber 2,919 00
Pine lumber 3,086 60
Explosives 151 19
Drift bolts, spikes, etc 804 98
Total $17,173 97
The total length of the two dams being 1,3621 feet, makes the cost per lineal foot $12.60.
Mradley $ Poates, Lngr's, N.Y.
Editors reprinting articles ironi this journal are requested to credit both the Joubnal and the Society before which such articles were read.
Association
of
Engineering Societies.
Organized 1881.
Vol. XVII. AUGUST, 1896. No. 2.
This Association is not responsible for the subject-matter contributed by any Society or for the state- ments or opinions of members of the Societies.
EXPERIMENTS CW VITRIFIED PAVING BRICK.
By F. F. Harrington, Member or the Engineers' Club of St. Louis.
[Kead before the Club, June 17, 1896.*]
About this time last year Prof. H. A. Wheeler delivered before this Club an interesting lecture on " Vitrified Paving Brick," in which the processes of manufacture, the methods of testing, and the uses of paving brick were very thoroughly described. The speaker reviewed the methods of testing and the results of many experimenters, and explained many ways of combining the results of the various tests in order to show the relative merits of different kinds of paving material. Although the tests given were made by reliable men, the chief objection to them as a whole is that the methods employed were so different that the results, from a scientific standpoint, are inconclusive and not comparable. Only a short time ago, also, Prof. J. B. Johnson addressed the Club on " The Resistance to Crushing of Brittle Materials," including vitrified brick, and in this address the results of tests by Bauschinger, at Munich, were exhibited and many interesting relations shown.
About a year ago the Water, Sewer and Street Commissioners of thi3 city established a testing laboratory, where all the material used by the city could be tested, instead of each department having its own testing done, as was formerly the custom. The use of vitrified brick had already been resorted to in the construction of alleys in the city, and its extensive use for street paving was about to follow. The ques- tion of determining the quality of the material therefore became an
* Manuscript received July 8, 1896. — Secretary Ass'n of Eng. Socs. 5
66 ASSOCIATION OF ENGINEERING SOCIETIES.
important one, and the laboratory was equipped with suitable apparatus for some of the essential tests.
A review of the past will show that vitrified paving brick has been used for several years in many cities and towns in the United States, and in most cases no laboratory tests were made, while in other cases reports of tests show them to be extremely crude and variable. Take for example the abrasion test. The material of construction of the' rat- tler may be cast or wrought iron, steel, wood, or some combination of these materials; the size ranges from 12 inches to 60 inches in diameter ; the speed from 200 to 10 revolutions per minute ; the charge of bricks for the test varies from 5 to 40, usually composed of samples of different manufactures ; sometimes whole bricks are used, and at other times cubical specimens are taken from them for test ; the size of the abrading material ranges from that of shot to scrap iron pieces weighing over five pounds each ; sometimes billets or blocks of wood are used as cushions for the bricks, and frequently granite, trap rock and other kinds of stone are introduced to furnish a comparison between brick and other paving material. The duration of the test varies from twenty minutes to three hours, and often an intermediate weighing is made, the loss during the first period of the rattling process, when the edges are defaced, being neglected, and that during the second period being taken for the loss due to abrasion. Specifications in different localities are therefore variable in their requirements and show a lack of information on the subject. Thus, in some cases, a few special tests must be employed, and in others, different ones, but it is very seldom that well-defined requirements are stated covering the essential tests. The paving-brick industry is now, however, developing rapidly, and the material is being extensively used in the residence districts of many of the largest cities, so the importance of an investigation of this subject is therefore apparent.
In 1895, at its annual convention, the National Brick Manufactur- ers' Association appointed a committee for the purpose of recommend- ing for adoption a standard series of tests on vitrified paving brick and the methods of making them. This Commission is now actively en- gaged in the investigation of the subject. The first meeting was held August 1, 1895, at which time a preliminary standard method of mak- ing the various tests was agreed upon as a basis of conducting the neces- sary experiments. The experimental work was also apportioned be- tween the members, in order to make the greatest haste in the prepara- tion of the final report. The Commission is composed of engineers and brick manufacturers, and its recommendations, it is expected, will there- fore be generally accepted.
It is intended to examine vitrified paving brick in this laboratory by means of the following tests: The rattler, absorption, cross-break-
EXPERIMENTS ON VITRIFIED PAVING BRICK. 67
ing, crushing, freezing, specific gravity and hardness. New machinery is being obtained for making all these tests. The Water Commissioner, Mr. M. L. Holman, has designed and built a new hydraulic testing machine for cross-breaking brick, having a capacity of 20,000 pounds, and has just completed the design of an hydraulic machine for crushing brick and other materials of construction, having a capacity of 1,500,000 pounds. This will be the. most powerful crushing machine in the West. The following are the results of the rattler and absorption tests made up to this time:
IMPACT AND ABRASION TEST.
The rattler test is considered by most engineers to be the most important test for vitrified brick. It is a measure of the toughness of the material, when laid in the pavement, to withstand the blows of horses' hoofs and the wearing action of the wheels of vehicles.
The tumbling barrel used for making the following tests is made of cast iron. It is polygonal in form, having 15 staves, and its dimensions are approximately 24 inches in diameter and 42 inches long. It re- volves on trunnions. Within the barrel is a cast-iron partition at right angles to the axis by means of which the length can be varied. The barrel is operated by a constant speed electric motor, through a main shaft, counter-shaft and gear wheels.
The vitrified paving brick used for the tests were made of pure shale,5 worked by the stiff mud process, and burned in down-draught kilns. Five hundred well-burned samples were selected from one kiln, with the view of obtaining the most uniform specimens possible. They are re- pressed brick and have rounded edges. The dimensions are 81 x 4 x 2£ inches. The volume of one brick is therefore 82J cubic inches, and the weight is slightly less than seven pounds.
The results of the tests are shown in Figs. 1, 2, 3 and 4.
Fig. 1. — An arbitrary length of barrel of 30 inches and speed of thirty revolutions per minute were chosen to begin the work. Five per cent, of the volume was then filled, requiring eight bricks, and the per- centage of loss calculated after forty and eighty minutes' tumbling. In the same manner the barrel was filled to 10, 15, 20 and 25 per cent, of its volume, using 16, 24, 32 and 40 bricks respectively, and the percent- ages of loss calculated as before, after forty and eighty minutes' turn- . bling. The maximum percentage of loss was thus found to be when the barrel was filled to 15 per cent, of its volume.
Fig. 2. — For the experiments shown in Fig. 2, the cast-iron partition in the barrel was moved so as to make the length successively 12, 21, 30 and 39 inches, and for each test 15 per cent, of the respective volumes were filled, requiring 10, 17, 24 and 31 bricks for the charges.
68 ASSOCIATION OF ENGINEERING SOCIETIES.
The percentage of loss was calculated in each case after tumbling forty and eighty minutes. It will be seen that the percentage of loss is prac- tically independent of the length of the barrel, when 15 per cent, of its volume is filled.
Fig. 3. — For these tests a length of barrel of 30 inches was chosen, as in the tests in Fig. 1, and 15 per cent, of the volume of barrel, or twenty-four bricks, were tumbled in each experiment. The barrel was run at speeds of 20, 25, 30, 35 and 40 revolutions per minute by changing the pulley on the main shaft, and the percentage of loss was calculated after 10, 20, 30, 40, 60 and 80 minutes' tumbling for each test. It will be seen that the percentage of loss continues to increase with the increase of speed and at a more rapid rate. This evidently shows that the construction of the barrel is such that a speed of forty revolutions per minute is not sufficient to cause the bricks to revolve around the axis of the barrel, nor is it sufficient to carry the bricks up to a height that will produce the greatest loss from the impact of the bricks upon one another, due to their fall. The curves indicated by 20, 30 and 40 minutes' tumbling are quite regular, while those of 10, 60 and 80 minutes' tumbling are irregular. The irregularity in the case of 10 minutes' tumbling may be accounted for from the fact that the edges of the specimens are being knocked off during this time, while that of 60 and 80 minutes may be caused by the disintegration of the bricks from the length of time that they have been subjected to the test.
Fig. 4. — In Fig. 4 the same tests are recorded as those shown in Fig. 3. The abscissa is here changed, however, to time of tumbling, while the speeds in revolutions per minute are written on the curves. These curves may be called " characteristic rattler curves " for the material at different speeds. They show the greatest loss to be during the beginning of the tests, when the edges are being defaced, and in general the percentage of loss is less with equal successive intervals of time during the continuance of the tests.
In making the preceding experiments, the bricks for each charge were weighed in bulk on a scale reading from £ pound to 250 pounds. The time of tumbling for all the tests was observed to the second. The speed of the barrel was so controlled that it did not fluctuate from that recorded a single revolution during the progress of the work. The most striking feature noticed was the uniformity of the brick tested. Thus the total weight of any 24 of the 500 bricks weighed, taken at random, did not vary more than a half pound. It was also found that the weight of any ten bricks did not vary more than one pound from that of any other ten bricks, when taken from the rattler at the completion of any particular test. This not only shows the brick to be very uniform, but also that the rattler is constructed so as to give results that are strictly uniform and comparable.
EXPERIMENTS ON VITKIFIED PAVING BRICK. 69
The results are :
(1) The maximum percentage of loss is obtained when the barrel is filled to 15 per cent, of its volume with brick.
(2) The percentage of loss is independent of the length of the bar- rel, when the foregoing condition is fulfilled.
(3) The percentage of loss increases at a more rapid rate than the speed from twenty to forty revolutions per minute.
(4) The percentage of loss decreases with equal successive intervals of time up to eighty minutes' tumbling.
A study of the results leads to the following recommendations for a standard rattler test.
Figs. 3 and 4 show that it is necessary first to choose a definite speed for running the barrel. A suitable speed would be thirty revolu- tions per minute, since this gives about the average circumferential speed of rattlers in general use. Then obtain characteristic rattler curves as shown in Fig. 4 for brick of each manufacture at the above men- tioned speed. When these characteristic curves are drawn, it will only be necessary to find the percentage of loss after tumbling samples a given time for the test, or in other words to determine a single point on the curves. A suitable length of time for continuing this test would be forty minutes, since from Fig. 4 it appears that the edges of the brick suffer the greater loss from impact during the first twenty minutes, while during the next twenty minutes the exposed surfaces are abraded.
Having decided upon a definite speed and a definite time of tum- bling, any one of the following three methods may be chosen for the standard test :
(1) Fill the barrel to 15 per cent, of its volume with the brick to be tested, adjust the partition in accordance with the number of brick on hand and tumble at the adopted speed for the chosen length of time. For this method it would be advisable to always test about fifteen bricks and move the partition according to their size. Not less than ten bricks should be tested.
(2) Clamp the partition at a definite position in the barrel, fill to 15 per cent, of its volume with the brick to be tested, and tumble at the required speed for the chosen time. For this method a good posi- tion for the partition would be in the middle of the barrel, making two chambers 21 inches in length, so that two tests could be made simul- taneously. To make the test in this way, it would be necessary to use about seventeen bricks of standard size (81 x 4x 2* inches) or about thirteen of block size (9 x 4 x 3 inches).
(3) With partition clamped as in (2) put in the rattler the five or ten sample bricks to be tested and fill to 15 per cent, of volume of chamber with a selected uniform standard brick, kept in the testing
70 ASSOCIATION OF ENGINEERING SOCIETIES.
laboratory for that purpose, and tumble at the required speed and length of time.
RATTLER EXPERIMENTS WITH CAST-IRON BLOCKS.
A method of making the rattler test prevalent in some places has been to tumble five sample bricks with ten cast-iron blocks, each weigh- ing about six pounds, for thirty minutes and determine the percentage of loss. In order to examine the reliability of this test, fifteen " unit bricks" as described above, were selected, five of which were tumbled with ten cast-iron blocks three successive times under the above condi- tions. The length of the barrel was 21 inches and the speed thirty revolutions per minute. Fig. 5 shows the results, the numbers on the curves giving the order of the tests. It will be seen that this method gives no characteristic curve such as we obtained when only bricks were rattled, and it is therefore apparent that the method represented in Fig. 5 is not a good one.
THE ABSORPTION TEST.
The absorption test is considered a very important one for paving brick for the reasons :
(1 ) For any particular brick, the percentage of absorption is an index to the degree of its vitrifaction.
(2) From a sanitary standpoint, it indicates the relative avidity for the retention of refuse matter, the evaporation of which pollutes the atmosphere with noxious gases.
(3) It furnishes a means of determining the possibility of the dis- integrating action of frost.
The oven for drying the brick was designed for the purpose. It is made of galvanized iron lined throughout with asbestos. It is divided by a partition in two apartments, each 15 inches wide, 30 inches high, and 26 inches deep, and alike in every respect, so that only one side of oven need be described. There are four sliding grates, each holding fifteen standard size bricks, or sixty in all, and the full capacity of oven is 120. The heat is supplied by a Bunsen burner placed in the center of the bottom, the mixture of air and gas passing through numerous small holes of a special cap on the burner, and the flame impinges on an iron plate below the grates, thus heating the oven uniformly. The temperature is regulated by a damper in the flue, and read on a ther- mometer in the top of the oven.
The results of the drying tests are shown in Figs. 6 and 7. The temperature of the oven varied from 220° to 240° F. The makers of the bricks tested are designated by letters on curves, as follows :
(a) Alton Paving Brick Co., Alton, 111.
(6) St. Louis Pressed Brick Co., Glen Carbon, 111.
EXPERIMENTS ON VITKIFIED PAVING BRICK. 71
(e) Standard Paving Brick Co., St. Louis, Mo.
(d) Purington Paving Brick Co., Galesburg, HI.
(e) Barr Clay Co., Streator, 111.
(/) Townsend Paving Brick Co., Zanesville, O.
(</) Moberly Brick, Tiling and Earthenware Co., Moberly, Mo.
(h) Galesburg Paving Brick Co., Galesburg, 111.
1 7c) Galesburg Brick and Terra Cotta Co., Galesburg, 111.
(I) Royal Paving Brick Co., Canton 0.
The average of two bricks of each kind was obtained for the curves: For Fig. 6, the bricks, as received from the makers, were dried in the oven for one week. For Fig. 7, these same bricks, when taken from the oven, were immersed in water twenty-four hours, and the drying process repeated for the same length of time. These two series of tests, it is thought, cover the conditions of the state of moisture of brick likely to be received at the laboratory for tests.
Figs. 8, 9, 10, 11, and 12 show results of the absorption tests. In Fig. 9, the absorption curves of Fig. 8, for the first three days in water, are shown on a larger scale. The figures on the curves of the plates represent bricks from the following manufacturers :
(1) Alton Paving Brick Co., Alton, 111.
(2) St. Louis Pressed Brick Co., Glen Carbon, 111.
(3) Standard Paving Brick Co., St. Louis, Mo.
(4) Purington Paving Brick Co., Galesburg, 111.
(5) Barr Clay Co., Streator, 111.
(6) Wabash Clay Co. (Poston Block), Veedersburg, Ind.
(7) Des Moines Paving Brick Co., Des Moines, Iowa.
(8) Townsend Paving "Brick Co., Zanesville, O.
(9) Mack Paving Brick Co., Pittsburgh, Pa.
(10) Moberly Brick, Tiling and Earthenware Co., Moberly, Mo.
(11) Imperial Paving Brick Co., Canton, O.
Three uniform samples of each of the above kinds of brick were selected. They were dried in the oven for forty-eight hours. Two bricks of each kind were immersed whole. The results are shown in Figs. 8 and 9. Both ends of the third brick of each kind were re- moved, leaving about half bricks with two surfaces from interior ex- posed to absorb water. These results are shown in Fig. 10. Also, a small piece from the interior of the third brick of each kind, weighing about 25 grammes, was tested, the results being shown in Fig. 11. The temperature of the water in which the bricks were immersed aver- aged about 60° F. The water on the surface of each specimen was removed with a dry cloth before weighing. The whole and half bricks were weighed on a balance to the nearest gramme. The small pieces were weighed on a chemical balance.
72 ASSOCIATION OF ENGINEERING SOCIETIES.
RESULTS OF. ABSORPTION TESTS.
Figs. 6 and 7 show that it requires four days to thoroughly dry vitrified brick when subjected to a temperature ranging from 220° to 240° F. under all usual degrees of moisture, and that in forty-eight hours they are practically dry. Thus, 94.1 per cent, of the whole amount from tho, bricks in normal state of moisture was evaporated in two days, and 95 7 per cent, of the whole amount contained in the samples previously immersed for twenty-four hours was driven off in two days.
Figs. 8 and 10 show that both whole and half bricks continue to absorb water up to twenty-four weeks, at which time the experiments were discontinued. For practical purposes eight weeks would be required to soak them. The percentage of absorption of the half bricks ordinarily exceeds that of whole bricks of the same manufacture. Thus at twenty-four weeks' immersion the percentage is greater for half bricks than for whole ones, except in numbers 3, 6 and 10. The average increase of half over whole bricks after soaking twenty-four weeks is 16.5 per cent.
Fig. 11 shows that the small pieces continue to absorb up to eight weeks, and also that the percentage of absorption of small pieces in eight weeks exceeds that of half bricks of the same manufacture except in the case of number 8, and that of whole ones of Fig. 8 without any exception. The average increase of small pieces over half bricks in eight weeks is 47.3 per cent , and over whole ones in the same time 66.1 per cent.
Fig. 12 shows the average absorption curves of whole bricks, half and small pieces found from Figs. 8, 10 and 11.
A study of the figures suggests the following method for a standard absorption test. Let about ten samples of the bricks to be tested be dried in the oven for at least two days. Then immerse in clear water and -obtain characteristic absorption curves for each manufacture as in Fig. 8 for a length of time of eight weeks. Rattled bricks should be used for these tests, since the preceding experiments show that a higher percentage of absorption is obtained when surfaces from the interior are exposed. Furthermore, this method conforms better with the conditions of actual service, since it is only a short time after the bricks are laid in the pavement before some of the exposed surfaces are worn away.
When the characteristic absorption curves of various kinds of brick have been obtained, it will only be necessary to immerse the samples, the characteristic absorption curve of which is known, for any con- venient length of time, and obtain the percentage of absorption for that time.
JOURNAL OFTH
N OF ENGINEERING SOCIETII
"~ -^^— ^—
15 percent, of volume "f barrel tilled with unit brick.
|
*s |
,-. |
||||||||||||||
|
" |
; r |
^ |
2- |
||||||||||||
|
• |
/ |
||||||||||||||
|
1 =L |
. |
-„ |
, |
, |
I |
v |
15 per cent, of volume of barrel filled with unit briek.
Fig. 3. Diameter of barrel, 24 incites. Length of barrel, 30 indies. 15 per cent, of volume of barrel filled iriih unit briek.
Experiment* with oast imn id.., U in rattier. Brick and 10 cast iron blocks.
Revolutions per minute, :;
|
S |
Percentage of Water Evaporatec 8 s s |
|||||||||||||||||
|
- |
\NN |
V-. |
\ |
|||||||||||||||
|
-3 |
V |
W |
\ |
|||||||||||||||
|
H h |
\ |
; |
\ |
\ |
N |
|||||||||||||
|
1 |
: |
1 |
\ |
\ |
\ |
|||||||||||||
|
fs |
\* |
\ |
||||||||||||||||
|
I |
ll> |
, |
\ |
l |
\ |
|||||||||||||
|
I. J |
li\ |
» |
1 |
\ |
||||||||||||||
|
i < |
v |
» |
• |
\ |
n |
|||||||||||||
|
B1 g |
t |
1 |
| |
|||||||||||||||
|
^ |
1 |
1 |
||||||||||||||||
|
|s S j |
••, |
> |
||||||||||||||||
|
IE re , |
% |
|||||||||||||||||
|
li^ |
j |
.i |
||||||||||||||||
|
Is ?* |
||||||||||||||||||
|
°g: > ^ |
||||||||||||||||||
|
| |
||||||||||||||||||
|
i ! |
||||||||||||||||||
|
a |
||||||||||||||||||
|
I ~ |
||||||||||||||||||
|
£ |
||||||||||||||||||
|
jj |
It |
1 |
||||||||||||||||
|
3 § |
||||||||||||||||||
|
" |
j. |
|||||||||||||||||
Charge for each lest, 5 unit.
: ASSOCIATION OF
|
1 1 |
|||||||||||||||||||||
|
/ |
|||||||||||||||||||||
|
/ |
|||||||||||||||||||||
|
J iy |
|||||||||||||||||||||
|
/ |
|||||||||||||||||||||
|
i |
/ |
||||||||||||||||||||
|
T |
|||||||||||||||||||||
|
7^ |
-- |
^.Aeeta^-- |
|||||||||||||||||||
|
'/f |
e. . |
||||||||||||||||||||
|
i |
|||||||||||||||||||||
|
h |
|||||||||||||||||||||
|
t |
i |
i |
|||||||||||||||||||
|
I, |
y |
||||||||||||||||||||
|
t |
|||||||||||||||||||||
|
L |
i — |
inilllllllllllllllllllll
p==HIHH!H==HHHH
7/ me in Water- Weeks
Fro. 8. Win-It- hrk'k* ic-slt-ii. Kiii'li .nrvf i- the :iver.i^f ui" two [..ricks Kcfcienic uiiiiiI.lts nuirkt'il u
Drying Tesls. Immerse-I in wal
Time m Oven- Hours
Fig. 7. r hour*. Each curve is the average of mo bricks. Reference letlet Temperature of oven, 220° to 240° F.
JOURNAL OF THE ASSOCIATION OF ENGI N EERI NG SOCIETIES.
|
/ |
||||||||||||
|
/ |
||||||||||||
|
/ |
||||||||||||
|
/ |
||||||||||||
|
/ |
||||||||||||
|
1 |
||||||||||||
|
1 |
/ |
|||||||||||
|
* |
||||||||||||
|
i |
/ |
s |
||||||||||
|
°' |
1 |
f |
||||||||||
|
•s |
/ |
/ |
||||||||||
|
!„ |
||||||||||||
|
-- |
- :; |
|||||||||||
|
/ |
...41 |
sa! |
"' |
|||||||||
|
/ |
||||||||||||
|
' |
||||||||||||
|
t/, |
||||||||||||
|
'■ |
&= |
|||||||||||
Time in Woler - Hours Fro. 9.
Absorption Tests. Whole bricks tested. Each curve is the a
|
' |
||||||||||
|
/ |
||||||||||
|
a, |
||||||||||
|
^ |
||||||||||
|
V~" |
||||||||||
|
t |
||||||||||
|
:§ |
||||||||||
|
o< |
||||||||||
|
o |
||||||||||
|
■J |
/ |
|||||||||
|
&° |
f |
■■'' |
'' |
|||||||
|
<fc |
r |
'-' |
||||||||
|
V |
||||||||||
|
I |
||||||||||
|
H-A- |
||||||||||
|
r |
||||||||||
|
f |
Time in Water -
Fig. 11. option Tests. Small pieces fro
|
rs- |
=_ |
|||||||||||||||||||||||
|
■7" |
||||||||||||||||||||||||
|
-4 |
^ |
|||||||||||||||||||||||
|
P° |
||||||||||||||||||||||||
|
{/'" |
||||||||||||||||||||||||
|
t^ |
||||||||||||||||||||||||
|
1- |
||||||||||||||||||||||||
|
V |
- |
|||||||||||||||||||||||
|
___ |
||||||||||||||||||||||||
|
V |
||||||||||||||||||||||||
|
; |
■ |
. |
'■ 1 |
AWurpiion Tests, Halfb
Time in Water - Weeks,
Fig. 10.
.■f fit. -I i l.rk'k rt ve.l, lenvini; tivn Mir '"in c-s from :
Reference number
|
f |
||||||||||||||||||||||
|
. |
||||||||||||||||||||||
|
mi |
.— |
|||||||||||||||||||||
|
(it. |
i£S- |
|||||||||||||||||||||
|
y" |
||||||||||||||||||||||
|
'(-*- |
||||||||||||||||||||||
|
r |
||||||||||||||||||||||
|
( |
1 — 7 |
j-,, — |
Time in Water - Weeks-
Fia. 12.
Absorption Teste. Resulting average curves of whole bricks, half bricks and small pie
PAKTICLES SETTLING THROUGH LIQUIDS. 73
THE CONDITIONS NECESSARY FOR EQUALITY OF VELOCITY IN PARTICLES SETTLING THROUGH LIQUIDS.
By Ltjther Wagoner, Member of the Technical Society op the
Pacific Coast.
[Read before the Society, August 7, 1896.*]
Under the title " On the maximum velocity acquired by small bodies falling in water and glycerine," the writer published a paper in Proceedings Tech. Soc. Pac. Coast, March, 1888, wherein certain con- clusions and empirical formulas were presented differing from the views previously held, and as the question is one of practical importance, especially in the dressing of ores and the separation of bodies by air or water, the writer has been induced to take up the subject again.
The published results of Prof. Richards, A. I. M. Engrs., Vol. xxiv, furnish data much superior to anything previously had, and, as his paper may easily be found, only a short abstract of his methods will be given. Thirteen kinds of minerals were experimented with. They were first assorted upon sieves into different sizes ranging from 10-12 to 120.140 mesh. The diameter of each sieve aperture was care- fully measured and the diameter of the ore grain is taken as a mean between the sieves passing and those rejecting the grain. Fifty grains or particles of sized mineral were next dropped into a vertical glass tube, and the time required for 90 per cent, of the grains to pass two wires eight feet apart gives data for finding the mean velocity ; the experiment was repeated ten to twenty times for each size, and the mean for all was adopted. We thus have data connecting the diameter or mean sieve opening and the maximum velocity of fall in the water. To have been complete the data should have given the average weight of a grain of each mineral. The immediate object of this discussion is to examine the facts about grains under one millimeter size, and the data of Prof. Richards has all been reduced from inches to millimeters, the m.m. being taken as unit for diameter x. and velocity equal v.
METHOD OF DISCUSSION.
Referring to Fig. 1, where the diameters x of the grains are shown as abscissae and the velocities v as ordinates, it is required to find an equation connecting v with x and which will be reasonably correct for diameters
* Manuscript received August 13, 1896. — Secretary, Ass'n of Eng. Socs.
74
ASSOCIATION OF ENGINEERING SOCIETIES.
smaller than the lowest values of x = 0.1171 mm. or from x = 0 to x = .1171. The impelling force is gravity and the weight of the body is a function of its diameter. The retarding forces are the cross-sectional area and perhaps the surface, but both are functions of the square of the diameter. The equation for uniform motion may be written
where v = velocity in mm., x = diameter in mm. of the opening in the
|
300 iii.n.. |
y^ Ciulcnn. |
V=S8S X Vl.O?T XL'+.S33 |
|
|
200m.ni. /t .-' |
/ |
^ — ; |
|
|
100 lit. in. / |
Quartz. 0 in. in. 1 |
V = 100.4 S.V2 |
|
|
^i*-"'"*' 1 |
r> iii.iu. 1 |
V/.90S X2+.5195 5 iii.iu. |
Diameter of Grain = a%
Fig. 1.
sieve, k is a general coefficient and /. x is considered the unknown quantity and is found
/ * =
k2
(Eq. 2)
v and x being known, k has* been found as follows: Assume the various values of k to be proportional to the area of the curve shown in Fig. 1, or, what is the same,
k sum of v
k/ sum of vt A reduction of the experiments of Pernolet (vide Annales de mine, 1853, p. 144,) on coal, quartz, and galena 3 mm. to 30 mm. diameter combined with some experiments of the writer, gives the value of & for galena where x is the sieve aperture, equal k = 283, from which
PARTICLES SETTLING THROUGH LIQUIDS. 75
all the other values of k for the different minerals become known. Substituting the proper value of k and solving the equation (2) for its 14 diameters, there result 14 values of/, x. Several formulas were tried to find an equation for the denominator, and the simplest one/, x = (a x2 + 6) was adopted, a and b were found as follows: let s and s, be the sums of the first and second sets of seven of x2, and I and Ft the corresponding sums of/, x, then
a s + lb = F , __ F—F,
a st + lb = F/ S — S,
This method of treatment gives equal weight to each of the observa- tions. The general result of the investigation points to an increase in
value of b for diameters below 0.2 mm., probably of the form n * f
But as the data relating to form, surface and weight, in terms of di- ameter, are lacking, it is useless to attempt more approximate formula.
FINAL EQUATION. 3
— z. &
V~k i/o X* + b
making x large, b can be omitted, and
v = kj \/x~
which is the ordinary equation as given in text-books ; making x. small, the value a x2 may be omitted, and then
3 '
v = ktl x2,
a result which accords with the facts as well as with the theory, because it is clear that very small bodies must remain suspended in the fluid (v = 0), hence the exponent must be greater than one. Were the old formula correct, a body whose diameter was dx would have a finite velocity.
The above equation appears to hold for diameters as small as x = 0.0001 mm. Dr. Barus (U. S. Geolog. Survey Bulletin 39) assumes Sp. Gr. quartz, clay, etc., 2.50, and from rate of observed subsidence computes for
x = 0.0001 mm. t. 15° C, v = 0.0000278 mm. t. 100° C, v = 0.000556 "
The above formula does not consider temperature, and gives v = 0.000139 m m., a result fairly in accord with that of Dr. Barus.
The following table shows the value obtained from a discussion of eight of the thirteen minerals given in the table quoted :
76
ASSOCIATION OF ENGINEERING SOCIETIES.
Minerals.
Anthracite
Quartz . .
Pyrrhotite
Chalcocite
Antimony
Wolframite
Galena
Copper
Sp. Gr.
1.473 2.640 4.508 5.334 6.706 6.937 7.586 8.479
25.36 100.4 140.1 140.5 191.4 205.5 283 187.6
.7815 .9030 .5248 .6396 .8799 .9034 1.0770 1.0630
.6267 .5195 .9159 .7902 .5485 .4887 .3220 .3510
(a. + b.)
1.4082 1.4125 1.4407 1.4298 1.4284 1.3921 1.3990 1.4140
The mean value of (a -f b) is 1.4156, which is nearly the same as \/2, = 1.4142. This close coincidence of value as well as the more im- portant fact of (a -f- b) = constant, has a significance that the writer is unable to grasp, and should lead to renewed experiments upon spheres whereby the weight would be known and the influence of form would be constant.
The relation — is not a constant for any two minerals, unless a
and b are the same, for instance making x large and small in the case of galena and quartz.
x large, -J- ^™^ = 2.581 lowest value of ratio.
v galena
3.58 highest value of ratio.
x small,
vt quartz
Having shown that the law of velocities is greatly changed for small values of x, it is suggested that it is probable that a similar modification will be found of the law governing the settlement of fine particles upon inclined planes (Vanners, canvas, etc.), and as perhaps more than 80 per cent, of all the ore stamped is under I mm. diameter, it seems to the writer that there is an excellent field here for original investigation by the various mining schools, of the laws governing the separation of small bodies under \ mm.
Editors reprinting articles from this journal are requested to credit both the Journal and the Society before which such articles were read.
Association /;'uoii^V Engineering Societ?
Organized 1881.
Vol. XVII. SEPTEMBER, 1896. No. 3.
This Association is not responsible for the subject-matter contributed by any Society or for the state- ments or opinions of members of the Societies.
WATER SUPPLY AND SEWERAGE AS AFFECTED BY THE LOWER VEGETABLE ORGANISMS.
By the Late Clarence O. Arey, C.E., M.D.
[Read before the Civil Engineers' Club of Cleveland, June 9, 1896.*] In taking up this subject, regarding the effect of lower vegetation upon our water and sewerage, it will first be necessary to study the nature and life-work of these minute organisms which are found every- where, and to establish their place in the circle of the varied forms of life. Man lives either upon other animals or upon vegetables. These other animals that furnish food for man live either upon vegetables or upon herbiverous animals dependent upon vegetable life. All animal life is therefore dependent upon vegetation for its existence. Upon what, then, does the vegetable life with which we see ourselves sur- rounded depend? It depends upon the gases in the air and in the soil in which it is developed. Water is part of the food of all life and it is not necessary to consider it in differentiating the various forms.
What furnishes the constant supply of the elementary gases upon which the higher forms of vegetable life depend ? The life-work of the lowest forms of vegetation is to supply these gases. The bacteria, the yeasts, and the moulds do this work. They take dead organic matter as their food and reduce it to its original elements, which are mostly gases. Without them, all dead matter, unless destroyed by fire or cauterizing chemicals, would remain forever in the exact condition that it was in when death took place. When the Egyptians mummified their dead, they simply destroyed all of these organisms of decomposition. We are
* Manuscript received September 3, 1896. — Secretary, Ass'n of Eng. Socs. 6
78 ASSOCIATION OF ENGINEERING SOCIETIES.
all familiar with the fact that plants will not grow on fresh manure. This is simply because the bacteria have not yet reduced it to the ele- ments necessary to feed the plant. It is also probable that the heat, produced by the chemical changes instigated by the bacteria, acts dele- teriously upon the plant. The pea and the bean contain a considerable percentage of nitrogen, and upon investigating the roots of these plants we find that they are covered with a species of bacteria whose function it is to produce nitrogen. Clover is similar, and farmers have planted their fields with clover in order to render the soil more rich, that is, to replace the nitrogen that had been exhausted from the ground. Lately the experiment of inoculating the soil with nitrogen-producing bacteria has been made.
Now, as to the structure of the bacteria, yeasts, and moulds. The bacteria are the lowest forms of vegetable life that we have. They consist of single cells, and their function as a class is to reduce dead matter to its original elements. They are not all engaged in this work, however. Some are parasitic and live upon other forms of life. We all of us have our skins covered with a variety called staphylo- coccus epidermatis albus, and discovered by Dr. Robb, formerly of Johns Hopkins University, but now of Cleveland. Of those that are parasitic in their nature, certain ones eliminate a poison which is deadly to the host, that is, to the person or animal upon which they happen to find an abiding-place. Right here comes the all-important point regarding bacteria ; namely, how they produce disease. All life requires food ; all life gives off excretions. All bacteria absorb food ; all bacteria excrete other matter. The comparatively few disease-pro- ducing bacteria excrete poisons more deadly quantitatively than any known chemical poisons. These poisons separated from the bacteria will produce the same disease as the bacteria themselves, but do so more quickly because the living bacteria require time to multiply until they are numerous enough to produce a poisonous quantity of their excretions before the symptoms of disease show themselves. These same poisons diluted sufficiently, as in drinking water, may after a time render the person drinking the water incapable of taking the disease they produce when given in poisonous quantities. The poison- producing bacteria are the ones that we wish to keep out of our water supply, out of our houses, and out of our sewers.
The yeasts are slightly larger organisms and contain a nucleus. They are generally gas producers.
The moulds are slightly higher up in the vegetable scale ; they branch and have fruit.
The yeasts and moulds are perhaps antagonistic to the bacteria.
The greatest enemy of bacteria is sunlight. If we take two sterile
WATER SUPPLY AND SEWERAGE. 79
gelatine plates and inoculate both with the same species of virulent bacteria, and expose one for half an hour to direct sunlight, and do not expose the other, the result will be that the exposed plate will contain no growth whatever, while the one not exposed will have a luxurious growth of the inoculated bacteria upon its surface.
In taking up the subject of water supply, let us first consider a river town. Suppose that a town is located on and takes its water supply from a river and that ten miles up the river is a small town which dis- charges its sewerage iuto the river, the question which at once arises is, will the health of the lower town be good ? The answer to this question will depend entirely upon the amount of sewerage discharged by the upper town in proportion to the distance between the towns and the size of the river. Let us leave out the question of chemical waste and consider only the effect of the disease-producing bacteria that are carried in sewerage.
The sewage, when small in quantity, is discharged into the river and is immediately diluted with the river Avater. It is tumbled over and exposed to the sun, and at every tumble thousands of bacteria are destroyed. The bacteria are filtered through the green slime growing in the rivers, going to meet their death in the filtration, till, at the end of three or four miles, the water, upon examination, is found to be pure enough for drinking purposes. But if the sewage is once allowed