Publisher’s version / Version de l'éditeur:
Engineering Journal, 40, 7, pp. 973-977, 990, 1957-10-01
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.
Questions? Contact the NRC Publications Archive team at
PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.
NRC Publications Archive
Archives des publications du CNRC
This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Foundation failure of the Transcona grain elevator
Baracos, A.
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
NRC Publications Record / Notice d'Archives des publications de CNRC:
https://nrc-publications.canada.ca/eng/view/object/?id=ae4b9c41-a0a5-448d-8cf9-91c2338d6079 https://publications-cnrc.canada.ca/fra/voir/objet/?id=ae4b9c41-a0a5-448d-8cf9-91c2338d6079NATIONAL RESEARCH COUNCII,
CANADA
:.;
4
:,;
8 e* ? g p T . . '<8- $
, _ i
THE FOUNDATION FAILURE
OF
THE
TRANSCONA
GRAIN ELEVATOR
A.
Baracos,
M.E.I.C.REPRINTED FROM THE ENGINEERING JOURNAL
VOL. 40, No. 7, JULY 1957
RESEARCH PAPER No. 42 OF THE DIVISION OF BUILDING RESEARCH
This publication is being clixtribtrted by the Divisiotl of Building Reseurch of the National Research Council as (I contribution towarcls better
building in Canada. It should not be reprorlrcced in zohole or in port, zcith- orit permission of the original publisher. The Division tootrld be glld to be of nssistance in obtaining such ~ ~ e r m b n ' o n .
Ptrblicc~tions of the Dioision of Bzlilcling Research may be obtained b y mcriling the appropriate remittcrnce, ( a Bank, Express, or Post Office Money Order or a cheque mcrde 71ayable nt pcrr in Ottau;c~, to the Receiver General of Canada, credit National Resecrrch Council) to t h e National Research Council, Ottazca. Stamps are not crcceptcrble.
the
Transcona Grain
A.
Baracos,
M.E.I.C.Depm+tment of Civil Engineering, University of Manitoba
THE METHODS of soil mechanics
1 ,
developinent in the past twenty- fivc: years have made possible the de- te, ,nination of the ultimate bearing c a ~ ~ a c i t y of soils. The safety factor necessary for sound engineering practice precludes correlation of the ltimate bearing capacity with the nalytical determination of the bear- ng pressures for the impending fail- ure. Thus only on rare occasions actual failure occurs is a tor- n possible. The foundation fail- 1913 of a million-bushel grain r at Transcona, a few miles from Winnipeg, Manitoba, provides such an opportunity. It is the purpose of this paper to correlate the data on the founclation failure of the Transcona grain elevator with a re- cently coinpleted field and labora- tory soil mechanics investigation using the latest analytical methods.At the Builcling Research Congress helcl in London, England, in 1951, a notable paper 011 "The Bearing
Capacity of Clays" was presented by Dr. A. W. Skcinpton'. The Transcoi~a elevator failure was used by this author as one of his examples and is incltidecl in his table of "Field Data on Ultimate Bearing Capacity of Clays". The table, however, was in- complete as presentecl, with data on actual soil properties missing for two structures, one of them the Trans- cona elevator.
Ailother speaker on the same pro- gram as Dr. Skeinpton was R. F. Legget, ~I.E.I.c., Director of the Division of Building Research of the National Research Council, Ottawa. As he was spealcing about "Special Foundation Problems in Canada", Mr. Legget also used the Transcona ele- vator as an example! I n discussion, the desirability of coinpleting the table in Dr. Skempton's paper was stressed. Mr. Legget thereupon promised at the meeting to expedite the study of this foundation failure. This paper prepared cooperatively
by the Division of Building Research ancl the University of Manitoba repre- sents the fulfilment of that promise. Owing to the necessity of com- pleting other outstanding soil and foundation studies, it was not until late in the summer of 1952 that the work herein described could be started. Only then was it found that some time previously two soil borings had been put down at the site of the elevator under the direction of Pro- fessor R. B. Peck, of the University of Illinois. Unfortunately, it was then too late to correlate the proposed Canadian investigation with Dr. Peck's work but it was decided to proceed as planned.
Dr. Peck has now published his results" accoinpanied by an eye-
Although the foundati~n failure of the Transcona grain elevatcr oc- curred as long ago as 1913, the c o n d i t i o ~ s involvccl have since fre- quently been discussed in con- nection with soil mechanics prob- lems. The author clescribes a recent investigation made jointly by the Division of Building Research, N.R.C., and the University of IManitoba.
witness account of the failure by Mr. L. Scott White". The investigation nom7 to be described was rather Inore extensive than the American study, the two being generally complemen- tnry, agreement between the various test results being reasonably close, even though carried out quite inde- pendently except for a check by t h e author on a few of the soil samples obtainecl by Dr. Peck.
General Description of the Structure
Development of Canada's vast wheat lands in the earlv , art of this ccntury resulted in serious congestion of the Winnipeg railroad yards dur- ing peak periods of grain movement. Construction was therefore started in 1911 on the Transcona elevator in
conjunction wit11 oile of the worlcl's largest railroacl yards to facilitate rapid grain movement and to give relief to the Winnipeg railroad
The plan of the elevator is in Fig. 1. I t consists of a dryer 18 by 30 by 60 feet high, l~ouse 70 by 96 by 180 feet '111d the bin house 77 by 1 9 5 by feet high, all constructecl mainly of reinforced concrete. The foundation failure occurred under the bin house which was designed for storing one inillion bushels of grain. It consists of 65 circular bins arranged 13 in each of 5 rows running north a n d south. The 48 interstices between bins are also used for storing grain. A raft foundation, of reinforced concrete 2
feet thick, supports the bins ancl the conveyor tunnels under the bins. The depth to t h e bottom of the footings was 1 2 feet below t h e ground sur- face. The design bearing pressure was 6,600 Ib. per square foot based on load bearing tests for which the data are no longer available. Dead weight of t h e bin house was very nearly 20,000 tons.
General Description of Soils
Greater Winnipeg lies in t h e basin of the glacial Lake Agassiz which ex- isted cluring the recession of the Wis consin ice sheet. Generally, in thi area, the soils may b e conveliiently grouped as follows. The t o p 10 feet or less consist of relatively recent de- posits of organic soils, flood-deposited silts and silty clays ancl outwash from higher ground, and modified lacus- trine deposits. Under these a r e found 40 feet or less of glacial lake de- posits forming hvo distinct layers of approxin~ately equal thickness. The top layer is a brown clay a n d is dis- tinctly varved with many fractional inch-thick layers of silt spaced be- tween layers of clay % inch or more thick. The bottom layer is a grey- coloured clay, softer than t h e over- lying material and having numerous
calcareous silt pockets a n d containing limestone gravel and stones at the greater depths. Beneath the clays are found glacial deposits of rock flour, silts, sands and gravel. The upper portions were deposited as the gla- cier receded and are underlain by subglacial drift which has been acted on by the full weight of the ice sheet. The subglacial drift is highly consoli- dated and supports many of the heavier stluctures in the \Vinnipeg area. The total thickness of the drift is about 10 feet but varies consider- ably from this value. The entire area is underlain by Ordovician limestone.
Description of Failure
The storage of grain in the bin house was begun in September, 1913, with considerable care taken to clis- tribute the gmin uniformly. On Octo-
ber 18, when 875,000 bushels of wheat were stored, a vertical settle- ment of a foot was noted within a n hour after movement had been de- tected. The structure then began to tilt to the west and within 24 hours was resting at an angle of 26" 53' from the vertical and the west side was 24 feet below its original posi- tion. The east side had risen 5 feet above original elevation. Eye witness accounts"~-' stated that the structure acted monolithically with only a few superficial cracks appearing. Its com- ing to rest, approximately 24 hours after the n~ovement began, corres- ponded with the cupola falling off
the top of the structure.
I t \vas reported that during the failure, the soil around the structure rose to a height of 5 feet above t h e ground surface a r o l ~ ~ l d the entire ]>in
Above: West side of elevator, showing tilt and soil upheaval. Below: East side following foundation failure; early stages of righting operations are shown under way. (Photos: Foundation Company of Canada Limited.)
house. Pliotographs taken after t h e failure show that the greatest up- heaval occurred on the west side and was considerably more than 5 feet. Calculations based on the dead weight of the bin house, 20,000 tons, and 875,000 bushels of wheat at 60 lb. per bushel, give a unit uniformly distributed pressure of 6,200 Tb. per square foot on the clay when failure took place.
The operations to right the struc- ture have been reported in detail by Allaire? T h e structure has been in successful use since its positioil was restored.
Field and Laboratory Investigation
Figure 1 shows the location of seben test holes used to obtain sam- ples for t h e laboratory tests. Holes 4 and 7 were sufficiently removed from the structure to avoid distur- bances caused b!~ the failure and the righting operations.
The remaining holes were located nearer the structure, some 60 feet from the bin house, in a n effort to ascertain the effects of failure. It was realized, however, that these holes would show the effects of almost 40 years of continued pumping that has t,tken place since the bin house was righted. Pumping has been necessary to keep the bottom of t h e bin house clry. After righting, the bin lhouse was ,tpproximately 34 feet below the prairie grade. It was not considered practical to pldce the test holes nny closer to the structure as the entire area nearer the building was dis- turbed by t~~nnelling, excavation. ctc., during the righting operations. To keep the bins dry, a 12-foot-deep trench had. in addition. been ex- cavated around the bin house on all but the south side, further cliscour- aging test holes any closer to the structure than those indicated.
The holes were bored to refusal at a c l e ~ ~ t h between 40 to 50 feet where the dense a n d coarse glacial deposits were encounterecl. A dia- mond clrill adapted for taking thin wall Shelby tubes 2 inches in diam- eter was used for boring and sam- pling. Sainples approximately g1/2 feet
long were taken at 5-foot intervals or less where changes in soil \Yere evident.
All sa~llples were examined in the laboratory ancl notes inade on colour, stratification, etc. On each sample, moisture contents, density, degree of saturation, and unconfined compres- sion strengths were determined. On representative samples, grain size, Atterberg limits, undrainecl quick tri-
axials under constant load incre- ments, specific gravity, a n d consoli- dation tests were performed. T h e un- confined compression a n d t h e un- chained triaxial tests were perfoinled on undisturbed samples trimmed to 1.5 inches diameter and approximate- ly 3.0 inches long. Both t h e field and laboratory testing were con- ducted during t h e autumn of 1952 a n d winter of 1952-1953.
Test Results
Typical results of the tests a r e shown in the Log of test hole 4, Fig. 3. N o tests were performed on the material above 1 0 feet in holes 1. 2, 3 , 5, 6, and 7 where fill placed during t h e righting of the elevator was encountered. Hole 4 showed t h e silts a n d silty clays as they probably were over the entire area prior t o the excavation for t h e foundations.
Below t h e 10-foot level t o a d e l ~ t h Fig. 1. Plan of the Transcona elevator. of 20.5 feet in hole 4, and from 25
to 2 6 feet in the other holes, a brown depth of 4 0 to 4 5 feet from t h e sur- countered in holes 1, 3 , a n d 7 and highly stratified or varved silty clay face, a highly plastic grey silty clay the transition layer were a s follo\vs: was found. ~h~ or was found with numerous tan-col-
varves more or less ~lorizolltal o w e d calcareous silt pockets a n d Unconfined compressive strcngth (lb./sq.ft.) . . . . . lG4l and consisted of layers of silt of frat- limestone pebbles. This material h a d Liquid limit . . . . . . . . . . . 75.9 tional inch in thickness between "bout lnoisture 'Ontent as Plastic; limit . . . . . . . . 22.8 closely spaced layers of clay appi.oxi- 'lay and h,Ioisture content ( % ) . . . . . . . . 49.9 mately ?h inch thick. Average test re- a lower unconfined ~ 1 . 1 . ~ . grain size grouping (70)
sults for this material were as follqws: In 3, and clay 38.7; silt 44.5; l ~ o t t o m few feet of t h e grey silty clay 13,0; 3.8
Unconfined compressive
strength (lb./sq.ft.) . . . 2160 \\(ere found to be very moist a n d soft. Unit weight of soil (lb./cu. ft.) 110 ~ i ~ ~ i d limit . . , , , , . . . . . , 85.3 Holes 1 and 7 showed no distinct
1'13stic limit . . . . . . . . 29.3 boundary between t h e grey silt a n d Because of the wide variation in Moisture contcnt ( % ) . . . 52.4 the underlying glacial drift. About 3 the bottom grey silty clay and the L,I.I.T. grain size grouping ( % ) feet of a mixture of both materials transition . layer,
110 average values clay 49.4; silt 42.8;
sand 7.4; gravel 0.4 formed a transition layer. Average are given. Unit weight of soil lb./cu. f t . ) 107 test results for t h e grey silty clay
'Under the brown silty clay to a excluding the very moist material en- All t h e salnples below t h e 10-foot depth showed complete or near com-
- ..- . . . .- ... . . .. -. . - plete saturation. Twelve undrained
triaxial tests on samples from hole 2 confirmed a negligible angle of in- ternal friction for this type of load- ing. T h e consolidation test results (for samples from hole 4) indicate a de- crease in compressibility with in- creased d e p t h . Swelling pressures determinecl by permitting undis- turbecl samples to swell under a small load ancl cletermining t h e pres- sure required to return t h e sample to its original volume, range from 560 to 2050 lb. p e r square foot and are typical of t h e Greater Winnipeg clays which contain about 3 0 per cent of the more active clay minerals
GREY SILTY CLAY GREY SILTY CLAY (montmorillonite). Preconsolidation
WITH SILT P O C K E T S WITH SLT POcKETS pressures a r e not accurately deter-
, mined on these clays b u t indicate SECTION THROUGH NORTH END OF BINHOUSE
BEFORE AND AFTER FAILURE
A F T E R ALLAIRE. A.S.C.E T R A N S A C T I O N S F E B 2 1916
FIGURE 2
that they a r e somewhat in excess of overburden pressures probably due to desiccatioli. The void ratio pres- sure cllrves are shown in F i g . 5.
indicated by a decrease in moisture d = depth of cover on
content approaching or below the footing
plastic limit. Up to 4 feet of the less p'or long continuous footings, the dense drift were found. Boring re- quantities N,, N,, and N , are pure fusal was encountered in the sub- llumbers depending on the angle of glacial drift corresponding to the internal friction,
+.
Their values are depth to which the north end of the given in most modern soil mechanics bin house settled following the fail- or foundation texts.Lire. Numerous stones prevented F,, the special case of + = 0, N,, strength and consolidation tests from beconles and N , = 0. TIle eclua- being performed on the glacial drift. tion thus becomes:
The following data, however, were q l L = NLc
+
yd (2)obtained:
Natural moisture Prandtl, in an early form of equa-
content range ( 7 0 ) 10.0- 13.4 tion (2) evaluated N,, as 5.14 ancl
Moist densitv (lb./cu.ft.)
.
. 157. -143 TelzaglliG gives 5.7 for general shear u uquid limit, average . . . . 91.0 failure and 3.8 arbitrarily for local
-
asticlimit, average . . . 11.9.I.T. grain size grouping ( % ) shear failure. T h e general shear fail-
clay (rockflour) 6.0; silt 33.6; ure applies when the stress-strain
sand 33.3; gravel 38.1 curve (from laboratory tests) is of
T h e test holes were not estencled the tvDe s h o ~ v n in Fig. 4a, or is , &
-
to the underlying limestone. Eight approached when negligible variation test holes bored by the owners of esists in both loading and soil con- the building have shown, however, ditions.that the limestone bedrock was at a depth of approxinlately 50 feet. Theoretical Bearing Capacity 1
The relatively rapid loading of the elevator on saturated clay corres- ponds to the laboratory undrained quick triaxial test for which the un- confinecl compression test is a special case. For such conditions it is recog- nized that the angle of internal fric- tion is negligible and thus the cohe- sion is equal to half the unconfined compressive strength.
In general the ultimate unit bear- ing capacity of a soil inay be ex- pressed by:
B
q,, = N,.c
+
N,,ycl+
N,y - " (1)For rectangular footings t h e value of Nc has been shown by analytical iuethods, ~llodel studies and a stucly of actual failurcs to be a function of
L d
- and -, where L = length of
B B
footing. Recently, Skempton' has given the following formula:
N,. = 5 ( l
+
B/5L).
(1+
d/5B) (3) The theory for equation (1) as- sumes that the soil fails along a com- posite curve as shown in Fig. 4B. Although the theory i s beyond the scope of this report, it may be notecl that when @ = o, the coinposite curve extends to a depth below the bottom of the footing equal to approximate- ly onc-half the footing width. As fail-MOISTURE CONTENT PARTICLE S I Z E UNCONFltQO
PERCENT DISTRIBUTIOII STRENGTH P S F
ole log. (Test hole 4.)
4
arge upheaval on t h e the building tilts. Stability Analysis
A general examination of f d u r e a n d test data show failure was consistent with ing capacity theory. T h e u quick triaxial test confirined gible angle of internal frictio composite curve along whic failed would have theoret tended t o a depth equal
one-half the foundation width or 38Y~ feet below the bottom of the foun- dation. Since the dense glacial till occurrecl a t approximately the same clepth, it did not prevent the full de- velopment of this curve.
It may also be noted t h a t the soil upheaval all around the foundatioils due to "eclge effect" at the start of failure actually occurred. Allaire" re- ports a n upheaval of 5 feet. Photo- graphs confirm that further large up- heaval consistent with theory oc- curred on the side to which the struc- ture tilted. The actual direction of tilting is not important as even a very minor eccentricity in loading or varintiori in soil condition could cause a fdilure to either side.
The nealest test holes t o the struc- tule on the side of tilting were 63 feet distant and from t h e examina- tion and testing of undisturbed sam- ples, the soil appeared t o be un- affected by the failure. Although the failure occurred nearly 40 years ago, it is not believed that t h e loss in strength of the soil resulting from the failure has been regained. Tests on similar Lake Agassiz deposits7 do not indicate any extensive thisotropic strength regain for this material. Al- though no remoulded strength tests were performed, it has been general- ly found that remoulding results in a loss of one-half of the strength of the Winnipeg clays.
I t is also reasonable to assume that because of the nature of the labora- tory stress-strain curves a n d the pre- cautions taken to assure uniform load- ing of the elevator, that t h e Terzaghi general shear conditions were satis- fiecl. It is questionable, however, whether the assumption of local shear value (N, = 3.8) would have been applicable had t h e stress-strain curves been different.
Thc undrained quick triasial test confirmed that the angle of internal friction was negligible a n d that equa-
Figure 4.
Moisture contents and dens-
Holes Holes
51.9 48.9
107.8 112.2 The average unconfined compres- sive strength values of 1933 Ib. per square foot for holes 1, 2 , 3, 5 , and 6 , and 1850 lb, per square foot for
depth from holes 1 and 3, and hole
7 respectively. The difference in the PRESSUE LB PER cu FT.
values of cohesion, density, and nloisture content mentioned, how- ever, are small and could simply re-
flect statistical accuracy. Fig. 5. Consolidatioll test results, holc 4.
Results of substitution in equation
(5) are shown in Table I. T h e unit weight, Y, of the soil covering the footings was taken as 107 Ib. per cubic foot and the cohesion as half the unconfined compressive strength.
Discussion
Basecl on results for:
The ultimate theoretical bearing capacity of 6420 lb. per square foot
using the most justifiable value of un- Brown silty clay - all holes
confined compressive strength, 1850 Grey silty clay - all holes . . lb. per square foot is remarkably Brown and grey silty clay
close to the actual bearing capacity all holcs . . . . .
. . . .
at failure of 6200 lb. per square foot. holes 1, 2, 3, 5, GThe correlation is even better than holes 4, 7 . . . . . .
statistical considerations of the data Note: actual ultimate bcaring capacity = 6200 Ib/sq. ft.
can substantiate. Reasonable correla- ultimate bearing capacity from a soil anics laboratory of the University of tion, however, is gained using the study are obvious. With the informa- Manitoba and to him the author re- other average cohesion values as tion now available and the additional cords his thanks as also to Dean A. E. shown in Table I. This is in spite studies being made on settlements, Macdonald, M.E.I.C. and to Mr. R. of such factors as pumping that may foundations on clay may be designed F. Legget, M.E.I.c., Director of the have caused soil changes since the with reasonable knowledge of the Division of Building Research, Na-
failure. safety factors involved and the fu- tional Research Council, with whose
Difficult to explain is the length ture behaviour of the structure. approval this paper is published as a
of time, 24 hours, which elapsed joint contribution from the University
from the time motion began until the Ack"owledgements of Manitoba and the Division.
building came to rest. The plastic The interest in this investiga- References
natc1.e of the soils and the gradual tion of the Canadian Pacific Railway, 1. Skempton, A. W. T h e Bearing Capaclty transfer of load from the upper stiffer through Mr. R. A. Emerson, chief ~ ~ $ ' , " ~ . I , ~ g \ ~ $ g l ~ ~ ~ a r c h Congress. clays to the softer underlying material engineer, and their permission for the 2. Pack. R. B. and F. G. Bryant. The
may b e responsible. The slow fail- sinking of the tesL holes on their
~ ~ , " ~ i n ~ l , " , " ~ r ~ t y $ d ~ ~ ~ ~ , " , " " ~ ~ ~ ~ :
ure and the varves in the brown property is gratefully acknowledged. 208. 1953.clay do not appear to have invali- The test borings were put down by 3' Scott White' L.
Failure: Eye-witness Account. Cbo-
dated the theoretical formula. the special soil boring crew under 'echn'que, 111.209-214. 1953. T O the engineer, it is most reassur- Mr. E. Hargest of the Manitoba De- 4' Winnipeg Free Press. October 20-
5. Allaire. A. The Failure a n d Righting ing that the study of the Trailscona partment of P u b l ~ c Works. The author of a Mill~on-bushel G r a i n Elevator. elevator failure and similar studies is indebted to the ~ o ~ l l l d a t i o n Corn- Trans. Amer. 80. Engineers. Vol-
December 1916.
reported for the foundation failures pany of Canada Limited for the pho- 6. Terzaghi, K. Theoretical Soil Mecha- on clays in widely separated areas, tographs which serve as illustrations.
Tg::
Art. 47. W1ley and Sons' In'.verify the present theories. The ad- The soil testing was carried out by 7.Berger, L. and J. Gnaedinger. Thixo- tropic Strength of Clays. A.S.T.M. Bul- vantages of being able to pledict the Mr. M. B O Z O Z U ~ in the soil mech- letln, september 1949.
A list of all publications of the Division of Builcl~ng Research is auail- able and may be obtained from the Publications Section, Diuision of Bttilcling Research, National Research C o ~ m c i l , Ottawa, Canadu.
