A risk-based diagnostic campaign optimization for electric power distribution networks

Failures of Medium Voltage Distribution Networks (MVDN) can result in power outages, penalties from authorities and large operational costs. To prevent the network failures, diagnostic campaigns are performed to estimate the degradation state of its branches. For underground MVDN, diagnostic mainly relies on Partial Discharge (PD) and Tan-delta measurements. These are costly and, thus, only a small portion of the network can be analyzed every year. Therefore, the prioritization of the PD measurements on the MVDN branches is fundamental for its efficient management. Within the Portfolio Decision Analysis, we develop a risk-based methodology for the identification of the optimal portfolios of MVDN branches undergoing diagnostic measurements. The branch failure likelihood is estimated through an expert-driven model based on the Multi Attribute Value Theory, while accommodating the incomplete information about the impact that network features have on the degradation mechanisms affecting the cable joints and insulators. The severity accounts for the number of users disconnected upon failure, which depends on the topology of the network. The optimal portfolio of inspections is obtained from an algorithm of the literature. Results of the method on a real industrial case study concerning the network of the Milan area are discussed.