Updating Anomaly Statistics—Multiple Anomaly Pieces

Abstract

Satisfactory geotechnical performance could be marred by hazards induced by unexpected presence of geologic anomalies. Engineers may need to assess the likelihood of anomaly occurrences and the size statistics, and to estimate the corresponding risk of unsatisfactory performance. Procedures are developed to infer the current statistics of anomalies with respect to encounter/no encounter of anomalies by a system of borings. Those anomaly pieces encountered may or may not be subject to further excavation, measurement, and removal. The probability of detecting a given anomaly piece embedded within the site will depend on the boring grid spacing. As larger-size anomaly is more likely to be encountered by borings, the sizes measured on the detected anomaly pieces represent at best a biased sample. Bayesian techniques are used to infer the anomaly statistics throughout a given site accounting for these relevant factors. Several example cases have been studied. Results appear to be reasonable and it is promising to apply the procedures for general application.