Australia's leadership in rooftop solar photovoltaic (PV) systems has propelled the nation to the forefront of renewable energy adoption, with over 4 million installations contributing more than 26 GW of capacity as of early 2026. Yet, this rapid growth brings challenges, particularly silent underperformance that erodes efficiency and profitability. Researchers from the University of Technology Sydney (UTS) and the University of New South Wales (UNSW) have addressed this head-on with a groundbreaking rule-based method for PV system underperformance detection, published in Solar Energy in April 2026. This innovation uses only readily available inverter data to pinpoint issues, offering a scalable solution for Australia's vast distributed PV fleet.

The method stands out for its practicality—no expensive DC-side sensors or complex machine learning models required. Validated on data from 1,089 real-world systems across Australia, it achieves up to 92% accuracy in classifying faults like clipping, tripping, and yield anomalies, potentially saving billions in lost revenue.

🌞 Australia's Solar Boom and Hidden Performance Gaps

Australia boasts the world's highest per capita rooftop PV penetration, with small-scale solar generating 14.2% of electricity in late 2025. The Australian Energy Market Operator (AEMO) projects up to 74 GW of PV capacity by 2035 in its Integrated System Plan. However, underperformance plagues many systems. Globally, asset issues cost the solar sector US$4.6 billion in 2023; domestically, a mere 10% loss could exceed $1 billion annually by 2035.

Common culprits include inverter clipping (when panels produce more power than the inverter handles), grid-induced export limits, shading from trees or buildings, soiling, wiring faults, and degradation. In commercial setups, a 'maintenance gap' leads to $340 million in yearly losses. Traditional monitoring often misses subtle, recurring issues that don't trigger alarms, leaving operators blind to efficiency drains.

The UTS-UNSW Collaboration: Pioneering Practical PV Monitoring

This research stems from the NSW Smart Sensing Network's Grand Challenge project, 'Smart Energy Asset Management Intelligence.' Led by Dr. Fiacre Rougieux at UNSW's School of Photovoltaic & Renewable Energy Engineering and Dr. Ibrahim Ibrahim at UTS's Institute for Sustainable Futures, the team partnered with Global Sustainable Energy Solutions and local councils.

Key contributors include Bernardo Mendonça Severiano, Jonathan Rispler, and Jaysson Guerrero Orbe from UTS, alongside UNSW's Dr. Baran Yildiz, Dr. Yinyan Liu, Professor Anna Bruce, and Dr. Rougieux. Dr. Yildiz, a Senior Lecturer at UNSW SPREE, specializes in consumer energy resources integration and leads projects like CICCADA on PV curtailment. Their work exemplifies Australian universities' role in translating academic research into industry tools.

How the Rule-Based Method Works: A Step-by-Step Breakdown

The innovation is a suite of algorithms analyzing AC-side inverter data—power output, maximum power point (MPP), and timestamps—at five-minute intervals. Unlike DC-heavy methods, it's cost-effective and brand-agnostic across eight major inverters.

• Step 1: Anomaly Screening – Flags zero generation, sudden drops, or sustained low output compared to expected performance ratios (PR).
• Step 2: Event Isolation – Isolates incidents like inverter tripping (complete shutdowns) or clipping (plateaued output).
• Step 3: Pattern Recognition – Identifies recurring anomalies or seasonal trends using statistical thresholds.
• Step 4: Classification – Categorizes as clipping, export limits, degradation, or grid events via rule sets (e.g., MPP voltage saturation for clipping).
• Step 5: Alert Prioritization – Scores severity for operator action, integrating historical data for trends.

This deterministic approach outperforms probabilistic ML in explainability and low-data scenarios.Read the full paper.

Validation: Real-World Results from 1,000+ Australian Systems

Tested on 1,089 systems (2,213 inverters) nationwide, against 807 industry-labelled faults. Accuracies: 92% for major events (e.g., tripping), 88% overall, 56% for subtle clipping—highlighting refinement needs. It caught issues missed by standard O&M, like a council system's five-month underperformance.

Compared to prior 2025 review on data-driven fault detection, this AC-only method scales better for distributed PV.

Addressing the 'Long Tail' Degradation Challenge

Complementing this, UNSW's January 2026 study revealed a 'long tail' where 20% of panels degrade 1.5x faster than the 0.9% annual average, shortening life from 25 to 11 years due to backsheet failures and microcracks. The rule-based tool detects early signs, mitigating financial risks in solar farms.

Economic and Environmental Impacts on Australia's Energy Transition

With rooftop solar hitting record 4.4 GW output in Q4 2025, underperformance threatens net-zero goals. This method enables proactive fixes, boosting yield by 5-10%, cutting $340M commercial losses, and supporting AEMO's 2035 vision. Fleet operators gain dashboards for remote triage, slashing site visits.

For universities, it underscores PV research hubs like UNSW SPREE, fostering PhD opportunities and industry ties.

Career Opportunities in PV Research and Higher Ed

Australia's solar surge demands experts. Roles in research jobs, fault analysis, and energy modeling abound at UTS and UNSW. Aspiring researchers can excel via programs like How to Excel as a Research Assistant in Australia. Check Rate My Professor for insights on supervisors like Dr. Yildiz.

• Research Assistant in PV monitoring
• Postdoc in renewable fault detection
• Lecturer in solar engineering

Future Outlook: Scaling and Enhancing PV Intelligence

Next steps include ML hybrids for clipping and integration with batteries/grid forecasts. As TOPCon cells face UV vulnerabilities per UNSW findings, enhanced protocols pair with this tool. Australian unis lead globally, positioning graduates for green jobs.

Conclusion: A Brighter Future for Australian Solar Research

The UTS-UNSW rule-based method transforms PV system underperformance detection from reactive to predictive, safeguarding Australia's renewable edge. Explore higher ed jobs, university jobs, and career advice to join this field. For faculty ratings, visit Rate My Professor; post openings at Post a Job.

Photo by Noble Mitchell on Unsplash