270297.observations on the om

8/13/2019 270297.Observations on the OM

1/8


2/8

when faced with unavoidable uncertainties ofnatural ground conditions. It was used more intui-tively than formally by many engineers in the past

but was recognized as a design approach by Ter-zaghi (Terzaghi and Peck 1967) as the observa-tional procedure. The formal ingredients andname of this method were laid down by Peck in

his Rankine lecture (Peck 1969). In this approachthe construction process may start with a design based on more optimistic assessment of naturalground conditions than in the conventional ap-

proach. At the same time, carefully planned meas-ures for detecting possible differences betweenassumed and actual ground conditions are pro-vided in the design, as well as carefully planned

provisions for actions to be carried out in theevent that significant differences do occur. By thisapproach the design process is extended into theconstruction phase. It takes advantage of observa-tions and data gathered during construction toadapt the design to actual ground conditions in anorderly and planned way. It may be characterizedas a learn-as-you-go approach and it was success-fully applied in practice by many designers oftunnels, excavations, foundations, ground treat-ment works, embankments, waste disposal struc-tures etc. (for an extensive review see Nicholsonet al. 1999 and Allagnat 2005).

The success of Terzaghis learn-as-you-go approach to the observational method, and particu-larly its formalization by Peck drew much atten-tion by the geotechnical community. Many casehistories adopting this method and several impor-tant contributions to its improvement were pub-lished. However, it also raised questions about itslimitations and proper use. A whole variety ofopinions emerged, from the one that the method isan element of a quality control system, to the onethat it allows for less intensive site investigations,up to the view that it stimulates the introductionof innovations and that it is one of the most pow-

erful weapons in the civil engineering arsenal.According to Powderham (1998), starting with amore optimistic initial design may tend to createconcerns about the safety, which could inappro-

priately be associated with uncomfortably lowsafety margins. This circumstance has discour-aged a wider use of this method.

Concerns about the limitations and the properuse of the observational method deserve furtherclarification. This is attempted in the paper by in-voking the nature of uncertainties and their related

probabilities, as well as by recognizing the role ofrisk management.

Uncertainties may invoke hazards, i.e. eventswith the potential for consequences. Uncertaintiesmay also be associated with probabilities. Theyenable the use of powerful analytical tools pro-vided by the theory of probability, statistics andthe theory of reliability. The combination of thequantified measure of the consequence of an

event, with the probability of its occurrence, isknown as risk. The combination is often replaced by the product, when risk becomes the statisticalexpectation of the consequence of an event, orexpectation of the cost incurred by the conse-quence, when the later is expressed by monetarymeans.

Risks may be studied by the risk analysis andmanaged by the risk management. The risk man-agement, widely used for some time by the finan-cial community and more recently by some otherengineering fields, is now gaining more and moreacceptance by geotechnical engineers (see e.g.Clayton 2001). The risk in engineering may belimited to acceptable or agreed levels by the ap-

propriate design measures, so that the risk man-agement enters the engineering design. Recogniz-ing, evaluating and managing risk are the core ofthe observational method.

It is worth recognizing that the observationalmethod assumes that the changes of the design are

possible, that it deals with uncertainties in theground conditions and that its implementation in-curs cost in addition to the cost of the conven-tional design. When design changes are not possi-

ble or not feasible, the method is not applicable.Where there are no uncertainties in ground condi-tions, or if they are negligible, if such cases existat all, the use of this method is uneconomical.

2 DUALITY OF UNCERTAINTIES

The probability is usually used as a standard wayto quantify uncertainty. However, it is importantto recognize the dual nature of uncertainty, asemphasized and discussed in details by Beacherand Christian (2003) and Christian (2004).

Two types of uncertainties may be recognized:the aleatoric 1 uncertainty, or uncertainty based onnatural variability of a property, and epistemic 2 uncertainty, or uncertainty originating from thelack of knowledge, or from the insufficientknowledge of a property, or from the consequence

1 After the Latin word aleator - gambler2 After the Greek word episteme - knowledge


3/8

of these. The two types of uncertainties play animportant role also in geotechnical engineering.The aleatoric uncertainty mostly deals with thespace or time variability of a certain ground prop-erty in a more or less homogeneous ground layer,and is thus associated with randomness The epis-temic uncertainty deals with the lack of knowl-

edge in site characterization, the inadequacy ofthe mathematical model used in design, or the precision to which model parameters can be esti-mated. Thus, some refer to the first uncertainty asobjective, because it rests in the field, and to thesecond as subjective, because it rests in ourminds.

While the epistemic or knowledge based uncer-tainty may be reduced by a more extensiveground investigation program, such a programwould not do much for the aleatoric or natural

based uncertainty. On the other hand, investigat-ing the natural variability of a property, for exam-

ple the shear strength, in an assumed homogene-ous soil layer may not be of much help if we failto recognize, for instance, the existence of a thinweak soil layer, which might govern the failuremode of the geotechnical structure.

Both types of uncertainties may be associatedwith probabilities. While the aleatoric probabilitymay be evaluated from the frequencies of field orlaboratory investigation results, the epistemic

probability is usually evaluated by subjective judgments based on experience.

Since both types of uncertainties are quantified by probabilities, it is important to know the typeof uncertainty the probability represents if a

proper interpretation of the analysis is sought.Furthermore, it should also be emphasized thatthe division or even the distinction between thetwo types of uncertainties may be somewhat

blurred and may depend on the conceptual modelwhich represents the natural ground conditions.

In this paper, the attention is drawn to the dual

character of uncertainties related to natural groundconditions. Recognizing this duality may have animportant influence on clarifying some ambigui-ties about the use of the observational method.The major contributions to the development ofthis method are, thus, revisited in the following

paragraphs and discussed in the light of this dual-ity.

3 MAJOR CONTRIBUTIONS TO THEDEVELOPMENT OF THE OBSERVATIONALMETHOD

3.1 Pecks ingredients of the observational method

As already stated, Peck (1969) coined the nameand formalized the observational method by iden-tifying its following eight ingredients: ( a ) Explo-ration sufficient to establish at least the generalnature, pattern and properties of the deposits, butnot necessarily in detail; ( b) Assessment of themost probable conditions and the most unfavour-able conceivable deviations from these condi-tions; ( c) Establishment of the design based on aworking hypothesis of behaviour anticipated un-der the most probable conditions; ( d ) Selection ofquantities to be observed as construction proceedsand calculation of their anticipated values on the

basis of the working hypothesis; ( e) Calculationof the values of same quantities under the mostunfavourable conditions compatible with theavailable data concerning the subsurface condi-tions; ( f ) Selection in advance of a course of ac-tion or modification of design for every foresee-able significant deviation of the observationalfindings from those predicted on the basis of theworking hypothesis; ( g) Measurement of quanti-ties to be observed and evaluated on actual condi-tions; and ( h) Modification of design to suit actual

conditions.At the same time, Peck discussed the advan-

tages and limitations of the observational method by presenting instructive case histories. Theirclose inspection reveals that the uncertainties hedealt with were mostly knowledge based or epis-temic. He also stressed that the full value of themethod cannot be realized unless the engineer isthoroughly familiar with his problem and has theauthority to act quickly upon his decisions andconclusions.

Out of the eight ingredients of the observationalmethod, the establishment of the initial design based on the most probable ground conditionsdrew most concerns. These concerns arise fromthe circumstance that when unfavourable condi-tions are encountered during construction, andthey have a rather high probability of occurrenceof 0.5 (because the initial design is based on most

probable ground conditions, which have an occur-rence probability of 0.5), contingency measureshave to be introduced to rectify the design. This

leads to cost and time overruns, while in themeantime the safety margin may decrease and the


4/8

risk increase. This situation would certainly not be favoured by many of the parties involved in the project.

3.2 Powderham on initial design and progressivemodification

Powderham (1994, 1998, 2002) recognized thatthe safety issues are essential and that a high de-gree of certainty in project performance andschedule is generally required in modern con-struction practice, which is led by teams ratherthan individuals. He further developed the obser-vational method by advocating a more conserva-tive initial design (with a probability of being ade-quate larger than 0.5), which is then progressivelymodified in small and well controllable steps asmore data and their trends become available by

observations, most probably in the direction ofsaving costs rather than introducing contingencymeasures. This situation would certainly be moreattractive to most parties involved. By this ap-

proach the level of risk may well be maintained oreven decreased as construction proceeds.

In summary, Powderhams approach is to ( a )commence construction with a design providingan acceptable level of risk to all parties, ( b) main-tain or decrease this level of risk, ( c) progressconstruction in clearly defined phases, and ( d )implement appropriate changes progressively anddemonstrate acceptable performance trough ob-servational feedback.

Although being more conservative than that proposed by Peck, the initial design proposed byPowderham may still be less conservative than theone in the conventional design approach, or, interms of probabilities, the initial design may be

based on more probable ground conditions thanthose for the conventional design approach. Pow-derham (1994) used the term more probable inthe context of the observational method to denoteconditions that are moderately conservative, with-out assigning them any specific probability value.As stressed by Powderham and Nicholson (1996),it often may be appropriate to start with a basi-cally conventional design, and then to proceedapplying the principles of the observationalmethod through progressive modification.

Case histories presented by Powderham in his papers are very illustrative. He described the useof very elaborate decision making trees, withmechanisms for triggering design changes, re-

sembling a traffic light system. He stresses the

great importance of a very strong connection be-tween design and construction teams includingadequate book keeping, good communicationamong parties involved, and adequate contractual

provisions between the client, the designer (orconsultant) and the contractor. The ground andconstruction conditions related uncertainties he

addressed were mostly knowledge based or epis-temic, as indicated by the title of one of his papers(Powderham 2002).

3.3 Muir Wood on tunnels and initial design

In his critical paper on the New Austrian Tunnel-ling Method, Muir Wood (1987) proposed a sim-

plified version of the observational method foruse in tunnelling: ( a ) Devise a conceptual model,(b) Predict expected features for observations, ( c)

Observe and compare against ( b), (d ) Are differ-ences between ( b) and ( c) explained by values of parameters, inadequacy of ( a ) or inappropriate-ness of ( a )?, ( e) Devise a revised conceptualmodel, ( f ) Repeat ( b), (c), ( d ) and ( f ) as appropri-ate.

Muir Wood stressed that the adoption of themethod in tunnelling must presuppose the abilityto supplement the tunnel support, while the ob-servational process is in progress, without the riskof collapse. It also must presuppose a general de-gree of confidence in the approach, which is ex-

pected to be based on comparable experience andon the analysis of the particular circumstance. Itcan again be argued that the above rules deal pri-marily with the knowledge based or epistemic un-certainty rather than with a simple statistical vari-ability of natural ground conditions, or thealeatoric uncertainty.

In his book on tunnelling, Muir Wood (2000),warns about possible problems if the most prob-able conditions, set up by Peck, are adopted inchoosing the basis for the initial tunnel design. By

performing a risk analysis under very simplifiedassumptions, he showed that the minimal ex-

pected construction costs for the initial design andthe ground conditions to be encountered duringexcavation, are obtained with the minimum prob-ability of being safe p, given by

11

2 p

k = - , (1)

where k is the ratio between the unit cost of sup- plementary support, to be added when worse

ground conditions are encountered, and the unit


5/8

cost of the same support if deployed as planned by the initial design. Because the unit cost of in-stalments of supports, provided for in the initialdesign, and being part of the construction routine,are lower than the unit cost for the same work,when performed out of sequence as a contingency

action, k in equation (1) will always be larger thanone and, therefore, p will always be larger than0.5, the value of probability for the most probableground conditions. Equation (1) is graphically de-

picted in Figure 1. The probability p = 0.5, for therelative unit cost k = 1, relates to the most prob-able ground conditions quoted by Peck in one ofhis eight ingredients of the observational method.

The preceding analysis is in principle equallyapplicable to geotechnical design of all geotech-nical structures, not only to tunnels. It shows that,

apart from concerns for safety, there exist eco-nomical needs for a more conservative initial de-sign than the one based on the most probableground conditions. The problem with this, as wellas with other similar, but less simplified riskanalyses, is to assign probabilities to particularground conditions to be encountered during con-struction with any degree of precision, in the casewhen the related uncertainties are knowledge

based or epistemic. On the other hand, this analy-sis shows that the Powderham proposal for a moreconservative initial design is substantiated by the

risk analysis under the requirement of minimizingconstruction costs.

3.4 Eurocode 7 requirements

Eurocode 7 (e.g. BSI 2004), the new Europeangeotechnical design code, addresses the observa-

tional method only briefly. But never the less itsdescription is interesting and therefore cited here(Clause 2.7):

(1) When prediction of geotechnical behaviour is diffi-cult, it can be appropriate to apply the approach known asthe observational method, in which the design is re-viewed during construction.

(2)P The following requirements shall be met before con-struction is started:- acceptable limits of behaviour shall be established;- the range of possible behaviour shall be assessed and itshall be shown that there is an acceptable probability thatthe actual behaviour will be within the acceptable limits;- a plan of monitoring shall be devised, which will revealwhether the actual behaviour lies within the acceptablelimits. The monitoring shall make this clear at a suffi-ciently early stage, and with sufficiently short intervals toallow contingency actions to be undertaken successfully;- the response time of the instruments and the proceduresfor analysing the results shall be sufficiently rapid in rela-tion to the possible evolution of the system;- a plan of contingency actions shall be devised, whichmay be adapted if the monitoring reveals behaviour out-side acceptable limits.

(3)P During construction, the monitoring shall be carriedout as planned.

(4)P The results of the monitoring shall be assessed atappropriate stages and the planned contingency actionsshall be put into operation if the limits of behaviour areexceeded.

(5)P Monitoring equipment shall either be replaced or ex-tended if it fails to supply reliable data of appropriatetype or in sufficient quantity.

The first paragraph of Eurocodes 7 clause 2.7explicitly states that the observational method is aviable alternative to conventional design when thegeotechnical engineer is confronted with knowl-edge based uncertainties of ground behaviour.However, as stated by Frank et al. (2004), itleaves open the manner in which safety is intro-duced in the supporting calculations. Same au-thors advise not to use the method when collapsecan occur without prior warning, as may be thecase for brittle ground or ground-structure sys-tems.

Figure 1 Required minimum probability ( p), that an ini-

tial tunnel design is safe for ground conditions encoun-tered during construction, as a function of the relativeunit cost ( k ), if minimum construction costs are to beachieved (Equation 1; after Muir Wood 2000).

1 2 3k

0.0

0.5

1.0

p


6/8

4 DISSCUSSION

As stated before, apart from the great success ingeotechnical practice, the application of the ob-servational method also raised concerns. Some ofthe more common concerns are as follows:

Starting construction with the design basedon the most probable ground conditions mayin due course endanger the safety;

What are the safety margins to be used in thedesign based on the observational method,with reference to the serviceability and ulti-mate limit states, as well as with temporaryworks?

Brittle ground or structural behaviour mayexclude the use of the observational method;

Is the observational method a substitute forthorough ground investigation works?

Legal matters, contractual and quality assur-ance restraints might hinder the applicationof this method.

In an attempt to resolve these concerns, Figures2, 3 and 4 present three different simplified hypo-thetical cases, A, B and C, which show the changeof anticipated values of a ground parameter,which governs the design, as the constructionworks proceed, and new observational data be-come available. Since the value of the parameter 3 (or its effect on the structure) is uncertain, its an-ticipated upper and lower bounds are followed

from the start to the end of the construction proc-ess. Their values at any moment are anticipated

by using all available data and information up tothat moment, including ground investigation andobservations, and by using the chosen conceptualmodel, intuitive or statistical, which defines theupper, the lower and the most probable parametervalues.

The range between the upper and the lower bounds represents the uncertainty; it covers theknowledge based component as well as the com-

ponent related to the natural ground variability. Itis assumed that the uncertainty decreases, due tothe decrease of its knowledge based component,as the construction proceeds, and new observa-tional data became available. The most probablevalue of the parameter falls somewhere betweenits upper and its lower bound.

The three cases depicted in Figures 2, 3 and 4refer to three possible situations, depending on the

3 In this discussion the value of a governing parameter

stays for its design value rather than for a value deriveddirectly from specific laboratory or field tests.

gathered and interpreted information from obser-vational data: Case A, where the lower bound and

A n

t i c i p a

t e d v a

l u e o

f a g r o u n d

p a r a m e

t e r

g o v e r n

i n g

d e s i g n

Upper boundMost probable (Peck)Lower bound (more probable, designby progressive modification, Powderham)Initial lower bound (conventional design)

Construction progress

Uncertainties

conventional design is conservative

Figure 2 Case A: Both, the most probable and the lower

bound anticipated ground parameter values increase withthe construction progress

A n

t i c i p a t e

d v a

l u e o

f a g r o u n

d p a r a m e

t e r

g o v e r n

i n g

d e s i g n

Upper boundMost probable (Peck)

Lower bound (more probable, design byprogressive modification; Powderham)Initial lower bound (conventional design)

Construction progress

Uncertainties

conventional design is conservative

Figure 3 Case B: The lower bound of the anticipatedground parameter values increase, while the anticipatedmost probable values decrease with the construction pro-gress

A n

t i c i p a

t e d v a

l u e o

f a g r o u n

d p a r a m e

t e r

g o v e r n

i n g

d e s i g n

Upper boundMost probable (Peck)Lower bound (more probable, design byprogressive modification; Powderham)Initial lower bound (conventional design)

Construction progress (time)

Uncertainties

conventional design is unsafe!

Figure 4 Case C: Both, the most probable and the lower

bound anticipated ground parameter value decrease withthe construction progress


7/8

the most probable values increase; case B, wherethe lower bound increases, but the most probablevalue decreases; and case C, where both valuesdecrease. Which of the cases will occur is notknown prior to the start of construction works.

By Pecks approach to the observationalmethod, the design follows the most probable

value, whereas by Powderhams approach, the de-sign follows the lower bound (or a value close tothe lower bound). The conventional design doesnot rely on observational data and remains un-changed during construction. Therefore, it followsthe initial value of the anticipated lower bound ofthe governing ground parameter.

The conventional design is safe, but uneconom-ical, for cases A and B. Case C is an engineersnightmare since it may prove to be unsafe whenthe safety margin is small. All three cases may

prove to be unsafe for Pecks approach when thesafety margin is small, whereas Powderhams ap-

proach is always safe for cases A and B. It is alsosafe for case C, provided there is sufficient timeto install the contingency measures. Therefore,Powderhams approach is better suited with re-spect to safety. Pecks approach would require alarger safety margin than Powderhams.

Regarding the safety margin, which should beused in the design based on Powderhams ap-

proach to the observational method, it may be ar-gued that no stage in the construction process can

be singled out. The safety margin should onlytake account of the current anticipated lower

bound of the governing ground parameter, and thetype of works, temporary or permanent, as in thecase of the conventional design, where the precisevalue of the governing ground parameter isknown in advance.

Regarding the differences between serviceabil-ity and ultimate (or failure) limit states, the an-swer to the question of suitability of the observa-tional method lays in the types of data that can be

observed and measured in situ. While we canmeasure displacements, rotations, strains, stressesand pore water pressures, and therefore controlthe serviceability limit states, the strength andstrength related parameters (bearing capacity, pullout capacity etc.) can not be observed without in-duced failures. Therefore, it would generally bedifficult, if not impossible, to modify an initiallyanticipated lower bound strength parameter (ob-tained from prior ground investigation works) asmore observations became available during con-

struction. If this is the case with a strength relatedground parameter governing design, the observa-

tional method would be reduced to the conven-tional design. This is particularly the case for the

brittle material behaviour. Following these argu-ments it can be concluded that the observationalmethod is best suited to control the serviceabilitylimit states when they govern design. It is appli-cable, but less suited, to control the ultimate (fail-

ure) limit states for ductile materials (where largedeformations or the creep type behaviour wouldindicate the approach to failure), but not so for

brittle materials (which might induce progressivefailure and leave no time to apply contingencymeasures), because again, when properly used for

brittle materials, it would be reduced to the con-ventional design approach. The latter case would

be represented by the lower bound curves in Fig-ures 2, 3 and 4, having a sudden vertical decrease.

The design in geotechnical engineering heavilyrelies on the proper and thorough ground investi-gation works. The observational method is no ex-ception, because it induces large additional costsin comparison to the conventional design. A thor-ough analysis of cost, reliability and induced riskswould theoretically give a balanced proportion

between efforts to be undertaken for ground in-vestigation works, and efforts for the implementa-tion of the observational method, but its reliabilityseems dubious at present.Legal matters, contractual and quality assuranceconstraints might indeed hinder the use of the ob-servational method. However, if proper attentionis given to those, and mutual interest exists be-tween parties involved, there are available solu-tions, as described by Muir Wood (1990) andPowderham (1996).

5 CONCLUSIONS

The major contributions to the development of theobservational design method in geotechnical en-gineering were reviewed and discussed. Dividinguncertainties into knowledge based (epistemic)and those related to the natural ground variability(aleatoric), it is possible to anticipate cases forwhich epistemic uncertainties might decrease asmore observational data become available in dueconstruction course. These are the cases suitablefor the application of the observational method.Three such cases of the development of the upperand the lower bounds, and the most probable an-ticipated value of the ground parameter governingdesign, were illustrated. The analysis of these

cases helped to clarify some common concerns re-


8/8

lated to the proper use of the observationalmethod. The main conclusions from this analysisare:

Pecks approach to the observational method,with the initial design based on the most

probable ground conditions, is less safe (andwould therefore require a greater safety mar-

gin) than Powderhams approach, by whichthe design progressively follows the anticipated lower bound values of the governingground parameter, as more observational data

become available; Pecks approach may alsorequire more use of contingency measures;

With Powderhams approach the requiredminimum safety margin may follow the cur-rent lower bound of the anticipated govern-ing ground parameter;

Powderhams approach would require the in-troduction of contingency measures only forcases where the conventional design ap-

proach would prove to be unsafe;The observational method is best suited for de-

signs that are governed by the serviceability limitstates. It is applicable, but less suited, for designsgoverned by the ultimate (failure) limit states withductile behaviour, and it is unsuitable for the ul-timate limit states accompanied by brittle behav-iour, when it is reduced to the conventional designmethod.

ACKNOWLEDGEMENTS

Many thanks go to A. J. Powderham for numer-ous enlightening discussions that ultimately re-sulted in the appearance of this paper. However,the opinions expressed are authors own. The au-thor would also like to thank M. Jamiolkowski,who acquainted him with the work of A. MuirWood, F. Schlosser, who helped in a crucial mo-ment, all members of ISSMGE TC 37, and W.Van Impe who initiated it all.

REFERENCES

Allagnat, D. (2005). La mthode observationnelle pour ledimensionnement interactif des ouvrages . Presses del'cole nationale des ponts et chausses, Paris,

Beacher, G. B., Christian, J. T. (2003). Reliability andStatistics in Geotechnical Engineering . John Wiley &Sons Ltd.

BSI (2004). Eurocode 7: Geotechnical design Part 1:General rules. BS EN 1997-1:2004. British StandardsInstitution, London.

Christian, J. T. (2004). Geotechnical Engineering Reli-ability: How Well Do We Know What We Are Do-

ing? 39 th Terzaghi Lecture. Journal of Geotechnicaland Environmental Engineering . ASCE, 130 (10),985-1003.

Clayton, C. R. I. (2001). Managing geotechnical risk: Improving productivity in UK building and construc-tion . Thomas Telford, London.

Frank, R., Bauduin, C., Driscoll, R., Kavvadas, M.,Krebs Ovesen, N., Orr, T., Schuppener, B. (2004).

Designers guide to EN 1997-1, Eurocode 7: Geo-technical design General rules . Thomas Telford,London.

Muir Wood, A. (1987). To NATM or not to NATM,Felsbau , 5 (1), 26-30.

Muir Wood, A. (1990). The observational method revis-ited. In: Proceedings of the 10 th Southeast Asian Geo-technical Conference, Taipei, 2, 37-42.

Muir Wood, A. (2000). Tunnelling: Management by de-sign . E & FN Spon, London.

Nicholson, D., Tse C.-M., Penny, C. (1999). The Obser-vational Method in ground engineering: principlesand applications . CIRIA, London, Report 185.

Peck, R. B. (1969). Advantages and limitations of the ob-servational method in applied soil mechanics.Gotechnique , 19 (2), 171-187.

Powderham, A. J. (1994). An overview of the observa-tional method: development in cut and cover boredtunnelling projects. Gotechnique , 44 (4), 619-636.

Powderham, A. J. (1998). The observational method application through progressive modification. Civil

Engineering Practice: Journal of the Boston Societyof Civil Engineers Section/ASCE , 13 (2), 87-110.

Powderham, A. J. (2002). The observational method learning from projects. Proceedings of the Institutionof Civil Engineers: Geotechnical Engineering , 155

(1), 59-69.Powderham, A. J., Nicholson, D. P. (1996). The wayforward. In: The observational method in geotechni-cal engineering (D. P. Nicholson & A. J. Powderhamed.). Thomas Telford, London.

Terzaghi, K., Peck, R. B. (1967). Soil mechanics in Engi-neering Practice . John Wiley, New York.

270297.observations on the om

Documents