Quantifying the risk of geotechnical site investigations

The site investigation phase plays a vital role in any foundation design where inadequate characterisation of the subsurface conditions may lead to either a significantly over designed foundation that is not cost-effective, or an under-designed foundation, which may result in foundation failure. As such, the scope of an investigation should be dependent on the conditions at the site and the importance of the structure. However, it is common for the expense dedicated to the site investigation to be a fraction of the total cost of the project, and is typically determined by budget and time constraints, and the experience and judgement of the geotechnical engineer. However, additional site investigation expenditure or sampling is expected to reduce the financial risk of the design by reducing the uncertainties in the geotechnical system and protecting against possible foundation failures. This research has quantified the relative benefits of undertaking site investigations of increased and differing scope. This has been achieved by simulating the design process to yield a foundation design based on the results of a site investigation. Such a design has been compared to an optimal design that utilises the complete knowledge of the soil, which has only been possible due to the use of simulated soils. Comparisons between these two design types indicate the performance of the site investigation to accurately or adequately characterise the site conditions. Furthermore, the design based on the results of the site investigation have been analysed using the complete knowledge of the soil. This yields a probability of failure and, therefore, has been included in a risk analysis where the costs associated with the site investigation have been measured against the financial risk of the design. As such, potential savings in financial risk for increased site investigation expenditure have been subsequently identified. A Monte Carlo analysis has been used in this research to incorporate the uncertainties in the foundation design process. Uncertainties have been included due to soil variability; sampling errors; measurement and transformation model errors; and errors related to the use of a simplified foundation response prediction method. The Monte Carlo analysis has also provided the means to obtain results in a probabilistic framework to enable reliability and risk analyses. Computer code has been specifically developed with an aim to: generate a simulated soil that conforms to the variability of soil properties; simulate a site investigation to estimate data for a foundation design; simulate the design of a foundation and conduct a reliability and risk analysis of such a design. Results indicate that there are significant benefits to be derived from increasing the scope of a site investigation in terms of the risk and reliability of the foundation design. However, it also appears that an optimal site investigation scope or expenditure exists where additional expenditure leads to a design with a higher financial risk due to the increased cost of the site investigation. The expected savings in terms of financial risk are significant when compared to the increased investigation cost. These results will assist geotechnical engineers in planning a site investigation in a more rational manner with knowledge of the associated risks.Thesis(Ph.D.) -- School of Civil and Environmental Engineering, 200