Statistical PERT™ Analysis Applied to Contingency and Comparative Results with Data for Mining and Energy Projects

Abstract

Historically the construction industry has experienced a poor record of delivering projects within their sanctioned budgets. Such capital cost estimate (CAPEX) overruns are well documented in media and industry publications [1 ,2 ,3]. Large projects often run 15% to 20% over budget, risking approximately $1.5 billion US of capital each year through 2030 [4, p4]. To improve CAPEX ranging analysis, the authors propose creating a model for Statistical PERT® analysis to be applied during prefeasibility studies. The historical data used to support the research is sourced from published feasibility studies and publicly released capital costs for completed projects. The projects of interest were completed between 1997 – 2024, where project elements were analyzed to determine the distribution representative for mining and energy projects. A statistically significant sample of 50 projects were selected to test the Statistical PERT® analysis by comparing the model simulation to final outcomes in two case studies.

The research allowed development of realistic risk profiles compared to assuming a normal distribution of risk (i.e. traditional Monte Carlo analysis), to improve insight of elements with dominant interdependencies. The proposed approach is a cost-effective early assessment tool that can be used prior to investing in a qualitative risk analysis (QRA). The results also highlight areas of influence to be considered in conjunction with early execution planning to provide transparency on potential causes and measures to mitigate CAPEX overruns.