Bayesian method for characterization of geological anomaly

A geologic anomaly is formed by any foreign material not included in the design soil profile. The performance of a geotechnical system could be significantly affected if the anomaly has undesirable properties. However, the presence of such an anomaly and its properties can never be fully predicted. Methods for updating the anomaly size and location distributions based on the engineer's experience and judgement, supplemented with the encounter history from exploratory borings performed at the site, are presented. The probability of anomaly presence can be updated independently of the updating of its location and size by utilizing the mixed model. Sometimes an anomaly may have been encountered by a boring, but it failed to be detected by the boring. The updating methods accounting for this imperfect detectability due to testing errors, indirect tests and intermittent samplings are developed. The effects of imperfect detectability are shown to reduce the effectiveness of a boring program in limiting the likelihood of anomaly presence, and, if present, in limiting the anomaly size and location.<<ETX>>