Discovering the Outage Attribution per Energy Source Using Fuzzy Inference and K-Means Clustering for Philippine Electricity Market

Promoting a diverse energy mix improves reserve planning and anticipating forced outages, as each energy source has different characteristics that affect its reliability and the likelihood of outages for policymakers and energy planners to enhance market stability. This paper uses Fuzzy Inference and K-means clustering to discover outage attribution per energy source in the Philippine electricity market. K-means clustering is utilized in pattern discovery in the outage mix and fuzzy inference to assess the outage risk level. In addition, the paper performed a sensitivity analysis using silhouette score and elbow method with inertia to find the optimal <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$k=3$</tex>. The paper outlines three outage risk clusters: (1) High-risk outages due to aging infrastructure and fossil fuel reliance, causing frequent disruptions. (2) Moderate-risk outages from planned maintenance and renewable energy variability, indicating integration challenges. (3) Low-risk outages linked to resilient infrastructure and backup systems, ensuring stable performance and minimal disruptions. The findings indicate that low-risk outages are associated with resilient infrastructure and effective backup systems, while high-risk outages are linked to aging infrastructure and a heavy reliance on fossil fuels. Further exploration of each cluster's operational characteristics could inform energy policy and infrastructure improvements in the Philippines.