About This Lecture

Site-specificity is a unique feature in geotechnical engineering. Site investigation data obtained from one site cannot be directly used for another site. The same principle of site-specificity applies to load-test and monitoring data. However, it is not uncommon that non-site-specific data are used to support site-specific decision-making. For instance, engineers routinely adopt transformation models to estimate design soil/rock parameters, and most transformation models are calibrated by non-site-specific data. It is quite extreme and unrealistic to ban such models. In contrast, the success of such transformation models indicates that non-site-specific data may have certain values for site-specific decision-making. As we enter the era of BIG DATA, it is timely for geotechnical engineering people to ponder the value of non-site-specific databases. Computer science people have been very successful in exploiting the value in non-person-specific or non-case-specific databases (BIG DATA). It is natural for geotechnical engineering people to ask whether we can also exploit some value from non-site-specific databases. This is the main focus of this talk. The answer is YES. The talk will first introduce some exiting BIG DATA in geotechnical engineering. Some are soil/rock property databases, and some are load-test databases. Then, the talk will introduce some advanced methods developed by the author that can extract useful knowledge from BIG DATA to facilitate site-specific decision-making. Without BIG DATA and the advanced methods, such site-specific decision-making was very challenging or even infeasible, but sensible decision-making is now possible with the aid of BIG DATA and the advanced methods. Two types of decision-making will be illustrated: the determination of design soil/rock parameters for foundation design and the prediction of geotechnical structure performances.