Application of Bayesian Networks for maintenance and risk modelling

In an ageing electricity distribution infrastructure as found in many countries, measuring the impact of maintenance is becoming increasingly important. A realistic maintenance impact model will permit to improve existing maintenance strategies. This paper addresses this issue from a risk management perspective. The first part gives a short review of some of the existing mathematical models that are developed to assess the effect of the maintenance on component reliability. The second part presents the possibility of applying Bayesian Networks to model maintenance strategies' impact on utility risk. A practical example dealing with power transformer tap-changers is presented. Experience shows that the approach have the capacity of providing decision makers with a holistic overview of the problem they are facing as well as with quantitative model results.