Quantifying intra-organisational risks : an analysis of practice-theory tensions in probability elicitation to improve technical risk management in an energy utility

摘要

Assessing the likelihood of future events is core to technical risk management at Scottish Power Generation (SPG). Events can include failures resulting in unavailability of key assets, or incidents impacting staff safety or the environment. Eliciting probabilities from engineers to quantify the likelihood of future, uncertain events is challenging given the diversity of assets across the multiple, heterogeneous power plants operated by SPG. Such probability assessments inform investment decisions intended to manage technical risks and support regulatory compliance. Through interviews with engineers we reveal the opportunities for heuristics and bias that explain some of the historical disparities in assessments intuitively evident to risk managers. We propose better ways of obtaining judgemental probability assessments based on a study involving engineers and a control group of post-experience students. We find that the choice of scale descriptors impacts the probability values of defined events as judged by engineers. Consequently we suggest changes to the risk management system, including new design features to better frame and capture probability assessments. As a consequence of our study, the technical risk management process is being enhanced in a number of ways including the creation of a single organisational-wide framework, clearer guidelines, and better knowledge management.