Risk Assessment of Electrical Infrastructure Assets by a Multi-Objective Fine-Tuned Fuzzy Inference System

Risk assessment of aging electrical infrastructure assets aims to provide informed asset management decision support, such as asset renewal and maintenance. To this end, this letter proposes a systematical and practical method incorporating multiple inputs from utilities and regulators. Specifically, the probability of asset failure (asset health index) and the consequence of asset failure (criticality index) are combined into a Takagi-Sugeno-Kang fuzzy inference system (TSK-FIS) for risk quantification. To ensure the monotonicity of TSK-FIS, sufficient conditions are first derived and modeled as constraints. Then, the TSK-FIS parameters are optimized for three objectives: 1) the optimality objective enables the TSK-FIS to learn the expert knowledge from existing tools, ensuring alignment with established utility practices; 2) the robustness objective enhances its ability to prevent minor input changes from causing drastic output deviations; 3) to ensure such small input variations can still be reflected in the output to a meaningful extent, the differentiation objective is formulated to improve output resolution and mitigate over-concentration. Case studies on real-world power asset data of Singapore power grid have validated the proposed method.