Risk evaluation

The term “risk” is very general. Some specifications can be found for example on the web pages of Wikipedia. The European project INTACT (impact of extreme weather on critical infrastructures) gives a more focused and refined explication (www.intact-wiki.eu).

Eurocode 7 Geotechnical Design (EN 1997-1: 2004, EN 1997-2: 2006) is valid for European countries, and in pait 1 General rules, in chapter 2.1 Design parameters, paragraph 2.1 .(8), it states:

In order to establish minimum requirements for the extent and content of geotechni- cal investigations, calculations and construction control checks, the complexity of each geoteclmical design shall be identified together with the associated tisks.

First, it is recommended to separate out geotechnical structures that are associated with negligible risk and for which the minimum requirements will be satisfied by experience and qualitative geoteclmical investigation. Briefly stated, this is where all the know-how of our predecessors can be used and relied upon.

For all other geotechnical structures, the minimum requirements should be satisfied by a calculation process, which in the case of EC 7 is based on the limit state principle. However, for geotechnical structures associated with an abnormal risk, alternative provisions and rules should supplement the recommendations and standards specification included in EC 7.

Therefore, to establish optimal geotechnical design requirements, the EC 7 recommends the division of geoteclmical structures into three geoteclmical categories, 1, 2 and 3. From the previous statement, it is obvious that Eurocode 7 is focused on the conventional types of geotechnical structures, for which more detailed specifications are given.

Boundaries for individual geoteclmical categories are not strictly defined, although great scope is given to the individual states accepting EC 7. Therefore, the authors of the final version of EC 7 are putting forward some more detailed specifications in additional publications (Frank el al., 2007, 2011; Bond and Harris, 2008). For example, Frank el al. recommend incoiporating such geotechnical structures into the third geoteclmical categoiy, for which at least one criterion falls into the highest scale of evaluation, in other words:

• • Large and atypical structures;
• • Structures associated with abnormal tisk;
• • Structures which are interact with atypical or rare foundation conditions;
• • Structures which are loaded atypically or abnormally.

On the other hand, the EC 7 indicates the possibilities on how to determine the risk with which an individual geotechnical structure is associated, as in:

A. Direct specification ofgeotechnical structures falling into different geotechnical categories.

Here, underground structures (tunnels) can be used as an example. EC 7 specifies that tunnels in hard rock fall into the second geotechnical category (GC). Therefore, it means that all others are falling into the third GC. Nearly all large dams fall into the highest GC, as instead of the general rules expressed in EC 7, some additional conditions specified e.g. by ICOLD (International Congress on Large Dams), should be followed.

B. Risk evaluation based on the specification of:

• • Complication of ground conditions (complication of geological and geotechnical conditions);
• • Demanding nature and level of geotechnical structure;
• • Impact of failure of geotechnical structure on the environment (the so-called consequence classes).

C. Risk evaluation on the specification of uncertainties with which the main phases of the design and construction of geotechnical structures are associated.

More detailed specification of the possibilities for В and C are presented, insofar as the process of risk evaluation for both cases makes it possible.

2.1.1 Risk specification taking into account ground conditions,

Risk specification taking into account ground conditions, type of structure and impact of failure on the environment

The simplest way to specify any risk is to divide individual factors into three levels. Ground conditions (geological and geotechnical conditions) can be:

• Simple, modest or very complicated.

Geotechnical structures with respect to the level and nature of their demands can be:

• Unpretentious (simple), moderate (conventional) or very demanding.

The impact of structure failure on the environment generally is indicated by EN 1990 (EC 0), which uses the term “consequence classes” with respect to the impact on human life, society and the environment. The impact of any failure based on the first generation of EC 0 can be divided into:

• • Practically negligible, insignificant;
• • Moderate;
• • Very high.

By a mutual combination of these three factors, a differentiated evaluation is possible (see Table 2.1).

Table 2.1 Proposed levels for evaluation of structure, ground and the impact of failure to give a total risk evaluation (for classification into GC)

However, there is a general agreement that geotechnical structures falling into the first GC have, for all the aforementioned factors, the lowest level of classification and the geotechnical structures falling into the third GC have at least one factor falling into the highest level of classification.

The geotechnical structures falling into the second geotechnical category have therefore different types of combinations, which is why the EC 7 is focused in this direction. To be able to distinguish between different structures falling into the second GC, some proposals were made to further subdivide the category. For example, one proposal suggested three subdivisions for the design of a specific structure, according to the different demands on the quality of soil samples and the tests performed on them (Vanicek, 2016).

The recommended evaluation into five different categories is in principle comparable with the proposal for a second generation of Eurocodes, where five consequence classes (CC) are recommended - CCO to CC4. However, the Eurocodes do not cover cases with zero impact (CCO) or alternatively with extremely high impact (CC4). According to prEN 1997-l:20xx (E) Draft April 2018, consequence classes CC1-CC3 are combined with a different geotechnical complexity class (GCC), where this GCC involves the classification of a geotechnical structure based on the complexity of the ground and ground structure interaction and where three GCCs are distinguished. The result of this combination is the division of geotechnical structures into three geotechnical categories, so this means that the original classification into three GCs was retained.

The initial qualification and identification of geotechnical structures for these individual geotechnical categories should start during the earliest phases of project preparation and should become clearer over time.

Risk specification taking into account four main phases of a geotechnical structure design and execution

In this case, the risk specification is based on detailed evaluation of the uncertainties with which the main phases of the structure design and execution are connected. This specification is the more general. However, it still allows deeper insight into all aspects of the problems associated with geotechnical structure design and execution. Therefore, the individual phases will be discussed in more detail both in this chapter (from the viewpoint of risk specification) and in Chapter 3 (but from the viewpoint of risk reduction).

Here it is very useful to state that basic principles, the design of the engineering structure, and their execution are all linked. The designer has the main responsibility for a safe and optimal design. The designer should take into account all uncertainties, which will be briefly discussed in this section, and subsequently call on them when working on the geotechnical design report. But the designer is not the only person responsible for any risk of a structure failure or loss of its serviceability. Other partners share this risk, namely the person

Figure 2.2 Main phases of geotechnical structure design and execution

responsible for the ground (geotechnical) investigation report (GIR), the person responsible for structure execution (the contractor) and finally, but quite importantly, the investor (the owner). Only a very high level of cooperation among all these people can bring the expected results.

Since different people are responsible for them, the following four main phases of the whole process are mentioned (Figure 2.2) when these phases are also underlined in both first and second generations of EC 7.

Ground model

The Ground/Geotechnical model (GM) is part of the Ground (Geotechnical) investigation report (GIR) and consists of two main parts, a geological model and an overview of tests performed on the ground, either in the field or in laboratory tests.

A geological model provides in 2D or 3D a substitute visual representation of the real geological environment. A geological model specification is time-dependent process, as over time it improves from the conceptual level up to the model used by the designer. Later on, it can be clarified in more detail when real conditions are detected dining the phase of structure execution. The uncertainties linked with the geological model (or its credibility) strongly depend on:

• • Complexity of the geological environment;
• • Actual state of exploration of this geological environment;
• • Extent of the ground investigation and its quality;
• • Ability and professional skill of the persons responsible for the site investigation and interpretation.

The last point is closely related to column 4 in Figure 1.1. The geological model is not only a geometrical interpretation of the real environment. This model should also be interpreted and supplemented by the expected interaction of this geological environment with the proposed structure.

Experienced engineering geologists can specify this interaction either fr om the limit state point of view (how the structure is influenced by ground conditions) or from the environmental view (whether the proposed structure can have a negative impact on this geological

Figure 2.3 Uncertainties in a geological model caused by a great distance between investigation points

environment), as the final evaluation can be used also for the environment impact assessment (ELA) process.

An additional note focuses on the extent of a ground investigation. Close cooperation is needed between engineering geologist, designer and owner. Most geotechnical engineers will agree with the following statement. The total construction cost can be significantly reduced by spending slightly more of the budget on ground investigation and laboratoiy and field tests in order to capture the subsoil conditions more precisely. However, the question concerning an economically optimal budget is duly justified and has to be kept in mind.

The aforementioned problem can be shown for 2D structures such as earth structures of motorways and railways. A very important question relates to the distance of the investigation points. For a larger distance, the interpretation of the geological model (the geological profile in longitudinal direction) is more complicated and some irregularities can be missed, e.g. Figure 2.3. Being able to specify the optimal distance is therefore a very sensitive problem to which all partners should propose solutions.

The last note focuses on the difference between the embankment and cut from the view of the ground profile control during structure execution. The ground profile for the cut can be (and should be) controlled, either for the excavated slopes or for the bottom of the excavation, while the control of the ground profile under the embankment is not performed.

FIELD AND LABORATORY TESTING

The main aim of this section is to specify geoteclmical data for each lithological layer of the ground or for discontinuity between these layers, also taking into account fluctuation of groundwater.

The uncertainties associated with this section depend on similar factors as for the geological model. However, a widely experienced specialist on laboratory and field tests, with a good background in soil and rock mechanics, plays an important role here. The credibility of the results also depends on an appropriate selection of the individual tests.

What is predictive ability of the test results for the control of limit states? - paiticularly for calculation model, which can be as analytical or numerical one with demands for different input data. Therefore, the designer of factual geoteclmical structure should recommend or at least oversee proper selection of test and input data. Some geoteclmical data have great value for the contractor, e.g. for an optimal selection of construction technology, which can have a close relationship to the bidding process. Classification of ground for earthworks is also very important for earth structures in transport engineering.

The results of field and laboratory testing are generally presented in the Ground (Geo- teclmical) investigation report (GIR) in tabular form, with a summary of all results for individual lithological layers or discontinuity. This form of presentation should be supplemented by an interpretation of results, namely with respect to:

• • Comparison of the obtained geotechnical data with existing data obtained for a similar type of ground in the past. Whether they fall inside the range of expected (standard table) values or are outside this range is seen.
• • Comparison of obtained data regarding groundwater with the expected fluctuation.
• • The form of specification of geotechnical parameters. These parameters can be acquired from the test results directly, mostly from lab tests. Such parameters are denoted as measured values. But the geoteclmical parameters can also be acquired from indirect tests, e.g. from the field tests as penetration and pressiometer tests. Such parameters after that are denoted as derived values. The derivation method used should be mentioned, as should the theoiy, correlation or empiricism used.

Geotechnical design model

The Geotechnical design model (GDM) is closely connected with the ground model. Both are continuously evolved as the design proceeds, and the individual steps of GDM development can be used for the individual steps of the design. The final version of the GDM specifies the characteristic values of the geoteclmical parameters for different lithological layers of the ground or for discontinuities between them. These characteristic values should be selected as a cautious estimate of the value affecting the occurrence of the limit state and are used subsequently for the calculation.

The comparison of both aforementioned models (GM and GDM) will show⁷ very clearly what values were selected for the calculation from values obtained in the phase of investigation. This comparison is very important for the control phase. On the one hand, it shows how conservative or optimistic the designer was, and on the other hand, how close the selected values are to the values obtained on samples from exposed ground during structure execution (ground excavation).

The specification of the characteristic values is the most sensitive part of the whole project. By this phase, the geoteclmical design differs from the design of other civil engineering structures. Two designers can differ significantly.

The existing version of Eurocode 7 from 2004 refers to two u'ays by which characteristic values can be selected:

• • Standard tables of characteristic values related to soil investigation parameters (standard tables of strength and deformation properties based on soil index properties and on soil classification).
• • Statistical methods with statistical evaluation of the results obtained for an individual layer from lab or field tests.

In both cases, the characteristic value shall be selected as a cautious estimate of the value affecting the occurrence of the limit state, complemented by w'ell-established experience.

The exploration of these two ways depends on:

• • The phase of ground investigation;
• • The risk with which the structure is cormected.

It is obvious that standard tables can be utilized for structures connected with low risk or at the end of the first step of ground investigation (after desk study and site reconnaissance), when preliminary soil classification is possible. Statistical methods, on the other hand, are typical for structures connected with high risk or can be applied at the end of the design investigation phase, when for each layer there is a sufficient number of measured results. From the above parameter selection alternatives, it is obvious that a combination of both ways is typical for geotechnical structures associated with medium, moderate risk or after the preliminary phase of ground investigation, when the number of results of strength and deformation properties is insufficient for statistical evaluation.

Calculation model

The calculation model is the most frequently used method of limit state verification. The calculation model will describe the assumed behaviour of the ground for the limit state under consideration. Nowadays, two basic calculation models are used:

• • Analytical model.
• • Numerical model.

The analytical calculation model divides control into two basic parts - control of Ultimate limit state (ULS), which is the limit state of failure, and control of the Serviceability limit state (SLS), which is mostly associated with deformation.

At this point, it is suitable to present at least a short note on the principle of the limit state design. The characteristic values of the geotechnical parameters represent a certain value selected with respect to the solved limit state. The characteristic value of deformation properties for the SLS will be close to the average value, as the result of the deformation calculation should be as close as possible to the consequently measured value. On the other hand, the characteristic value of the strength properties will be somewhat conservative and will be influenced primarily by:

• • Length of the slip plane along which a given geotechnical structure can fail. For a longer slip plane, there is a higher chance for a compensation of variation of shear strength on both sides. Tire selection of a characteristic value for the shorter slip plane will be more conservative.
• • The ability of the geotechnical structure to transfer loads from weak to strong zones in the ground.

However, for the calculation of limit state of the ULS type, the characteristic value (X_k) is mostly not used directly, but the design value of geotechnical parameter (X_d) is used, when the relation is:

X_t = XJy_u (2.1)

Where y_M is the partial safety factor for a material property.

The design is fulfilling the ULS limit state when the equilibrium on the slip plane is reached. The numerical calculation model examines the stress-strain state of the ground, which is in some interaction with the proposed geotechnical structure. For the observed element the change of stresses or deformation can be obtained for a given change of loading. Nevertheless, even then it is appropriate to solve two basic limit states (ULS and SLS) separately, when the characteristic values are applied for solving the SLS while the design values are preferred for the ULS.

The uncertainties of the analytical model are relatively high, and for each typical geotechnical structure or some part of it, there should be a separate definition. For example, for the problem of slope stability, the calculation model should take into account not only the geological model, but also:

• • Seepage and pore water distribution;
• • Short- and long-term stability;
• • Type of failure (circular or non-circular surface, toppling, flow).

The appropriate selection of the slope stability method is very important, as each has different basic assumptions (e.g. the Bishop or Janbu methods etc.). Therefore, the selected method should be well set out together with basic assumptions.

The uncertainties of the numerical method (mostly Finite Element Method (FEM)), or of the risk, are connected with the credibility of the substitution of a real zone by the geological and geotechnical design models with finite elements. Generally, this risk is associated with:

• • Coirect and precise division of the observed zone (geotechnical design model) into individual elements;
• • The function expressing the change of properties within individual elements - basic function;
• • The constitutive model, which is expressing the dependence of the deformation changes on stress changes, and finite elements on model structural components;
• • Finite elements to model interfaces;
• • Boundary conditions;
• • Ability to model technological sequences.

Therefore, the selected numerical model should always be closely specified, at least for the possibility to control the input data of geotechnical parameters and for the aforementioned points respectively, for the possibility of controlling the presented results. In general terms, the model should be validated.

More specifications concerning calculation models for earth structures can be found in Vanicek and Vanicek (2008) and will be raised also in Chapter 3, mainly concerning how the uncertainties (risk) can be reduced.

Geotechnical structure execution

Risk connected with geotechnical structure construction has two different levels:

• • The teclmical level, which will be discussed in Chapter 3.
• • The legislative, juridical level.

This second level is mostly associated with implementation of new geoteclmical structures in the vicinity of, possibly also in close interaction with, older existing ones.

All we really know is that each change in stresses initiates changes in deformations. Therefore, if stress changes in the ground induced by a new structure also affect the ground below the older structure, these stress changes must also create deformations below this older structure. However, the owner of the older structure usually agrees with the new one only under the condition that “the new structure will not have any influence on the older one.” This condition is an obvious contradiction. However, it is accepted and transformed by the designer and the contractor into a new condition, namely that “the change does not cause visible deformations - cracks.” Consequently, both the design and construction technology adopt this new condition. Special care should be devoted to neighbouring historical structures that are usually much more sensitive to the additional deformations. As a certain protection, the passportization of the old visible cracks on an existing older structure can be done beforehand.

However, there is another sensitive problem. The owner of the older structure might allow some provisional actions to guarantee the safety of both structures during the construction period. However, after that, there can be a demand to deactivate these provisional actions. For example, this could be for anchors situated below the older structure. A typical case in this direction is the collapse of the twin towers of the World Trade Center in New York. The excavation of the ruins was significantly held back while the stability of external walls was restored.

A short additional remark should be made dealing with the lowering of groundwater. This lowering can have a negative impact on neighbouring structures as it is at the same time increasing effective stresses. The increase of effective stresses below the older structure leads to additional settlement, mostly likely a differential one. The calculation has to evaluate this risk and propose some counter measures, if needed.