GOVERNANCE-AWARE, AI-DRIVEN ANOMALY DETECTION FOR DER-RICH DISTRIBUTION NETWORKS

Frederick Madriaga

PhD Candidate, Faculty of Business and Economics, Monash University (Zema Scholar)

Supervisors: Prof Russell Smyth, Prof Le Hai Vu

Research interests: Distribution networks with high DER, spatiotemporal machine learning, causal inference and explainability, cyber-physical security and governance, market-aware resilience, utility analytics and MLOps.

Frederick researches how to keep increasingly digital, distributed electricity networks secure and dependable as Australia moves toward net zero. He develops governance-aware anomaly detection that combines time-series machine learning with causal inference to localise disturbances on feeders with high penetrations of rooftop solar, EV charging, and batteries. Beyond detection, his work focuses on explainability and policy relevance – linking technical risk signals to sector governance frameworks and market outcomes so operators can prioritise interventions that strengthen reliability and consumer trust.

Before commencing the PhD, Frederick worked as an Infrastructure & Cloud Technical Project Manager/Solution Architect on national-scale identity and security programmes in high-assurance, hybrid-cloud environments. He holds a Master of Cybersecurity from the University of Adelaide and has led cloud migrations, security architecture, and AI/ML integration initiatives.

At Monash, under the supervision of Professor Russell Smyth and Professor Le Hai Vu, Frederick is building methods and tools that translate complex data into clear operational decisions. The goal is to deliver practical insights and policy-relevant guidance that help distribution businesses operate DER-rich grids safely, efficiently and transparently.

Working in critical-infrastructure projects, I repeatedly saw how security controls, operational workflows, and compliance frameworks can drift apart – especially when new DER technologies arrive faster than policy updates. This PhD lets me bridge that gap: to pair modern ML and causal methods with governance so grid operators can act quickly, confidently, and in a way that stands up to stakeholder scrutiny.