This session will address the limitations, field realities, and operational integration of AI in solar and storage applications.
Key points include:
•Combining physical models and machine learning
•Detection of underperformance, mismatch, soiling, degradation, and failure signatures
•Distinguishing signal from noise in high-volume operational datasets
•Using AI outputs for O&M prioritisation and value recovery
•Limitations of AI when data is incomplete, biased, weakly labelled, or not representative
•Challenges in newer technologies where field evidence is still limited
•Cybersecurity considerations and trust boundaries in digital operating environments
