Basic Principles for Foundation Design

Information Requirements and Foundation Design Process

The Institute of Civil Engineering (ICE 2012b) outlines the typical information requirements, the design reports and the project phases that are followed to design and construct a foundation that is presented in Figure 3.2.

Figure 3.2 Information requirements and the foundation design process (Source: modified from ICE 2012b)

The term in bold and underscored represents rhe main concern of this book - namely, evaluation of analytical methods in foundation design with a field load test database. In the early stage, the key questions are as follows:

• 1. Site geology, hydrogeology and history: What are the geological ages and depositional environments of the main deposits under the site? What subsequent changes may have occurred (e.g. landslides, influence from human activities)?
• 2. What will be builtf: What are the forms of structure, likely construction methods, main site constraints, applied loads and permissible structural movement?
• 3. Engineering knowledge: Whar existing knowledge do we have? For example, relevant technical literature, good case histories and local experience.

The collation of available data into a desk study report is an important and cost-effective means of managing ground risks in a project (Phoon et al. 2019). It is often the case that designers “inherit” ground or site investigation reports from previous or adjacent projects and studies. It is very important that the scope, adequacy and reliability of this information should be assessed in the context of current project requirements. As shown in Figure 3.2, the following steps are recommended to design a foundation (e.g. Bowles 1997; ICE 2012b; Hannigan et al. 2016; Poulos 2017):

• 1. Locate the site and the position of load.
• 2. Plan and execute site investigation to assess site stratigraphy and variability.
• 3. Perform in situ testing to assess the appropriate engineering properties of the site.
• 4. Conduct laboratory testing to supplement the in situ testing and to obtain more detailed information on the site.
• 5. Determine the engineering properties relevant to design. For variable ground conditions, different models could be used to allow proper consideration of site variability.
• 6. Assess foundation requirements based on experience and relatively simple methods of analysis. In this preliminary design, considerable simplification of both geotechnical profile and structural loading is necessary.
• 7. Refine the design based on more accurate representations of the structural layout, applied loadings and ground conditions. From this stage and beyond, close interaction with the structural engineer is an important component of successful foundation design.
• 8. Detail the design with the structural engineer towards a compatible set of loads and foundation deformations.
• 9. Verify the design by in situ foundation testing. If the behaviour deviates from that expected, the foundation design may need to be revised.
• 10. Monitor the performance of the structure during and after construction (e.g. settlement at a number of locations around the foundation).

Overall, foundation design is an iterative process. Once relevant information has been obtained for the site, an experienced engineer should be able to identify appropriate conceptual designs for the foundation. Analyses would then be carried out to check that this concept is acceptable from stability and settlement considerations. If one or more requirements are not met, the concept should be modified (e.g. pile foundations would be deepened, or shallow foundations would be widened) and re-checked. It is common to consider a few different options. Throughout this conceptual design process, the way in which uncertainties can be best managed to minimize risks needs to be considered. Given the intrinsic uncertainties in ground conditions, it is always important to verify the critical design assumptions.

Foundations of existing highways and over-river bridges may have significant functional value. Hence, re-use of foundations of existing bridges during reconstruction or major rehabilitation can result in significant savings in cost and time. Also, planning for re-use during the construction of a new bridge will meet an important sustainability criterion. Some engineers are advocating that we should consider sustainability besides safety and serviceability in design. The FHWA report (Agrawal et al. 2018) summarized numerous case examples on the re-use of bridge foundations in the United States and Canada to present a detailed process for resolving integrity, durability and capacity issues encountered during the re-use process. This document is not meant to be used as a guideline, only as a decision-making tool in addressing technical challenges and risk in re-using bridge foundations (Agrawal et al. 2018).

General Considerations

As noted by ICE (2012b), any foundation design has to include a consideration of the following:

• 1. Acceptable stability and deformation-. The foundation needs to carry the loads safely. Adequate FS against collapse is required, and excessive foundation movement must be prevented. There can be a wide range of collapse or deformation mechanisms that may need to be considered.
• 2. Risk management-. There can be a wide range of risks, including health and safety, technical issues (e.g. ground, groundwater and foundation behaviour), performance of the foundation constructor, environmental, commercial (costs and programme) and design interfaces (communication). These risks, during and after foundation construction, need to be assessed and managed.
• 3. Costs and programme for construction: The preferred foundation solution will usually be the most economical to build, provided that the long-term performance and associated risks are deemed to be acceptable. Calculation of the foundation cost may be based on assessing the quantities (e.g. volume of concrete for strip footings or number, depth of piles) for the permanent works and using unit rates for each of the main foundation elements.

Foundation Selection – the Five S’s

The overall design process needs to be well managed to ensure that there is good communication between different design teams and between design and construction. To provide a framework for selecting the most appropriate type of foundation, a useful mnemonic is the “5 S’s" (ICE 2012b): soil, structure, site, safety and sustainability. In this regard, Table 3.4 provides

a series of questions that a foundation engineer should ask him or herself before developing a conceptual design. Each term was discussed in ICE (2012b).

Permissible Foundation Movement

The magnitude of the permissible foundation movement is a key factor in designing the type, size and cost of foundations. This section provides a brief introduction of guidance on limiting settlement. Firstly, it is important to distinguish between the following (ICE 2012b):

• 1. Total settlement (s_t) - this may cause damage to services connecting into a structure but will not lead to damage to the structure itself.
• 2. Differential settlement (s_d) - because of rigid body rotation or tilt, which may be noticeable in high buildings and affect lifts, escalators, etc.
• 3. Differential settlement - because of relative displacements within the structure. It can lead to structural damage. For example, there may be differential settlement between adjacent building columns or bridge piers.

Differential settlement could induce the damage of buildings (e.g. cracks through bricks and mortar). Burland et al. (1977) noted,

Compared with the literature on the prediction of foundation movements, the influence of such movements on the function and serviceability of structures and buildings has received little attention. Yet major and costly decisions are frequently taken on the design of foundations purely on the basis of rather arbitrary limits on total and differential settlements.

Ideally, the foundation engineer would be able to predict the amount of differential settlement that a structure can tolerate and then predict the differential settlement that will actually occur because of the structural loads and the soil response below the foundation. For most practical situations, it is impossible to make this prediction accurately. The accuracy of settlement calculations is less than that for capacity calculations. Reasons could include estimation of stiffness is less accurate than strength because it is strain dependent, and settlement is more sensitive to spatial variability because it may only involve local shearing. Capacity mobilizes a large volume of soil mass, thus enjoying variance reduction because of the spatial averaging effects. Some discussions on the difficulties in estimating stiffness can be found in Whitman (2000) and Vardanega and Bolton (2016). Figure 3.3 provides a comparison between calculated and measured settlements for spread footings in cohesionless and cohesive soils. Each bar represents the

Figure 3.3 Accuracy of settlement calculations (Source: data from Burland and Burbidge 1985)

90% confidence interval (i.e. 90% of calculated settlements will be within this range). The line in each bar represents the average prediction, and the number to the right indicates the number of data points used for evaluation. As shown, the calculated value can be quite different than the observation. This is basically consistent with the observation of Phoon and Tang (2019a) who conducted a comprehensive review of model statistics, indicating a high dispersion in settlement calculation.

Also, Burland et al. (1977) cautioned, “If an attempt is made to model analytically the structure and calculate the effect of differential settlements, one obtains ridiculously low allowable differential settlements because of the unrealistically large bending moment that will be calculated.” This is because of the difficulty in simulating real soil-foundation-superstructure interaction behaviour that can be affected by the following factors (Boone

• 1996):
• 1. Variations in soil properties,
• 2. Variations in the construction sequence of the foundations and superstructure,
• 3. Uncertainties regarding the rigidity of connections across the foundation and within the superstructure,
• 4. Flexural and shear stiffness of the superstructure,

• 5. Changes in stiffness of the superstructure during construction and how load redistribution to the foundations will be affected,
• 6. Degree of movement or slip between the foundation and the ground,
• 7. Overall shape of the ground-movement profile and the location of the structure within it,
• 8. Building shape (both in plan and height),
• 9. Uncertainties in the way that the loads will redistribute in the long term as the structure settles differentially; and
• 10. Variable influence of time, both in the rate of the settlement of the ground and how creep and yield within the structure will develop.

For probabilistic assessment of differential settlements, the horizontal scale of fluctuation is important, but it is difficult to estimate because bor- ings/field tests are generally conducted far apart from each other. Tables 2.6 and 2.7 provide some guidance on the selection of this important parameter. The aforementioned complexities necessitate simple guidelines for practical applications that are summarized in the next section.

Guidelines on Limiting Settlement

The development of criteria for routine limits on permissible settlements has been established empirically on the basis of observations of settlement and damage in actual structures (e.g. Skempton and MacDonald 1956; Polshin and Tokar 1957; Grant et al. 1974; Burland et al. 1977; Wahls 1981; Moulton et al. 1985; Boscardin and Cording 1989; Boone 1996; Zhang and Ng 2005).

Skempton and MacDonald (1956) reported observations of settlement and damage to 98 buildings. Polshin and Tokar (1957) presented tolerable settlement criteria on the basis of twenty-five years of Soviet experience. Grant et al. (1974) reviewed settlement and damage data with an additional ninety-five buildings. Wahls (1981) summarized tolerable displacements for various types of structures. Moulton et al. (1985) investigated the performance of 439 bridge abutments that had experienced some type of displacement. Boscardin and Cording (1989) used analytical models and field data to develop procedures to evaluate building responses to excavation-induced settlement. They emphasized the importance of horizontal strain in initiating damage. The larger the horizontal strain, the less the tolerable angular distortion before some form of damage occurs, as illustrated in Figure 3.4. The solid circles represent the real case histories for the damage of various buildings (e.g. brick-bearing wall, wood and steel frame and masonry-barrel vault) in Boscardin and Cording (1989). Four levels of damage are distinguished (circled numbers in Figure 3.4; i.e. negligible, very slight to slight, moderate to severe and severe to very severe). This could be of particular importance when assessing potential damage arising from tunnelling. Similarly, for bridges, Barker et al. (1991)

Figure 3.4 Relationship of damage to angular distortion and horizontal extension strain (Source: data from Boscardin and Cording 1989), where I = negligible damage, 2 = very slight to slight damage, 3 = moderate to severe damage and 4 = severe to very severe damage

observed that settlements are more damaging when accompanied by horizontal movements. Zhang and Ng (2005) studied the performance of 171 bridges and 95 buildings that have experienced certain settlement (vertical displacement) and 204 bridges and 205 buildings that have experienced certain angular distortions (or relative rotation). The histograms of settlements and angular distortions for bridge and building foundations are presented in Figures 3.5-3.8. Based on the observed displacements that are considered tolerable and intolerable, the probabilistic distributions of the limiting tolerable settlement and angular distortion for each type of structure were established.

Published guidance on routine limits of permissible settlement considers sand separately from clay. This is because structures are more likely to suffer damage if the settlement occurs rapidly than if it develops slowly over many years. Hence the settlement limits for sand are lower than those for clay. Terzaghi and Peck (1948) suggested that for footings on sand, the differential settlement (s_d) is unlikely to exceed 75% of the maximum settlement (s_max). As most ordinary structures can withstand s_d = 20 mm between adjacent columns, a limiting value of s_max = 25 mm was recommended. For raft foundations, the limiting value of s_max increased to 50 mm. Skempton and MacDonald (1956) correlated angular distortion (6) with total and differential settlement for 11 buildings on sand and concluded that for a safe limit of 8 = 1/500, s_d = 25 mm and s_t = 40 mm for isolated foundations and 40-60 mm for raft foundations. Of thirty-seven settlement results reported by Bjerrum (1963) only one exceeded 75 mm and the majority were less than 40 mm.

Figure 3.5 Histograms of settlement of bridge foundations (Source: data from Zhang and Ng 2005)

Figure 3.6 Histograms of angular distortions of bridge foundations (Source: data from Zhang and Ng 2005)

Figure 3.7 Histograms of settlement of building foundations (Source: data from Zhang and Ng 2005)

Figure 3.8 Histograms of angular distortions of building foundations (Source: data from Zhang and Ng 2005)

Skempton and MacDonald (1956) suggested s_d = 40 mm for foundations on clay and s_t = 65 mm for isolated foundations and 65-100 mm for raft foundations. With the data from Skempton and MacDonald (1956) and Grant et al. (1974), Burland et al. (1977) plotted s_d against s_max for framed buildings on isolated foundations and buildings with raft foundations. A distinction has been drawn between buildings founded directly on clayey soils and those on a stiff layer overlying the clay stratum. Framed buildings on raft foundations were distinguished from buildings on isolated foundations. Based on the results of these studies, routine limits of s_max and s_d for buildings are summarized in Table 3.5. Data in Figures 3.5-3.8 indicate that recommendations in Table 3.5 will usually be conservative, particularly for raft foundations. Routine limits of permissible movements for a wide range of structures and associated infrastructure are given in Table 3.6.

Site-Specific Assessment

It should be emphasized that the typical limits in Table 3.5 should only be used for low-risk site conditions and for simple structures. For more complex situations (e.g. heterogeneous ground conditions, variations in loading, construction duration, site history and site topography, changes in superstructure stiffness across the foundation and potential for brittle behaviour), a more detailed consideration of differential ground movements and permissible limits for foundation design may be necessary. Details can be found in ICE (2012b).

Table 3.5 Routine limits of s_max for buildings (Source: data from ICE 2012b)

Notes:

• (1) Pad and strip footings are normally considered to be “flexible” when assessing the overall interaction of the foundations and superstructure. As individual elements, they may be rigid, depending on their thicknesses compared with their lengths or widths.
• (2) These limits apply when the clay is immediately below the foundation.When a stiff layer is between the foundation and the clay, then s_d can be substantially reduced and typically s_d<0.5s_max.This stiff layer could be natural (say because of a dense sand or gravel stratum) or be constructed as part of the foundation design.
• (3) Assume linear interpolation to assess s_d, when s_max = 50 - 200 mm.
• (4) This is usually conservative, as the relative raft-bending stiffness increases, s_d can be substantially reduced (max).
• (5) s_d for sands can exceed 0.75s_max.
• (6) These routine limits are only intended for low-risk situations and simple structures.

Table 3.6 Routine limits for permissible movements (Source: data from ICE 2012b)

(Continued)

Table 3.6 (continued)