Statistical Analysis Offers a Rational Approach for Controlling Foundation Risk in Design–Build Projects
Ross McGillivray, P.E., a Senior Consultant in Ardaman’s Tampa office, has co-authored an article in the January/February edition of Geostrata magazine, published by the Geo-Institute of the ASCE.
Mr. McGillivray and his collaborators, Dr. Gregory Beecher and Thomas M. Waits, P.E., have developed a statistical decision-making model for predicting foundation costs and quantifying uncertainty on design build projects. Design-Build Contractors often have insufficient geotechnical engineering data when preparing bids, and this leads to covering potential risk with contingency costs, or worse yet, accepting risk and exceeding foundation cost budgets. Mr. McGillivray and his co-authors apply statistical analysis to a large data set from Tampa Bay to develop a predictive model of pile depth. They apply the approach to designs for a proposed new Howard Franklin Bridge in Tampa/St. Petersburg. They suggest that “the value of applying statistical analytics to selecting test-pile locations, and ultimately to forecasting production-pile lengths, is that the risk of excessive variation in pile length can be better evaluated and planned for, leading to “more confident forecasts of cost and schedule.”