Autonomous PV Asset-Health and Energy-Yield Optimization System

Research context & significance

Research questions & objectives

• Build an autonomous controller that jointly optimizes panel position, PV health, tracker health, weather risk and maintenance priority to improve net energy and reliability.

Research approach

1. A · Input / Context

   Weather/irradiance, electrical telemetry, inverter data, RGB/IR imagery, and tracker sensors: position, motor current, torque, vibration and temperature.

2. B · Method / Process

   AI forecasting and fault/health models feed a health-aware tracker controller. The controller trades energy gain against actuator energy, wear and failure risk.

3. C · Evaluation / Validation

   Compare fixed PV, a conventional tracker and AI health-aware control. Measure net energy, tracking error, fault metrics, maintenance lead time, loading, availability and LCOE impact.

Expected outcome

A prototype decision framework/digital twin that co-optimizes energy yield, asset health and maintenance actions using feedback from actual field behaviour.

Progress since last update

• Integrated research scope defined across energy optimization, tracker health, PV fault detection and predictive maintenance.
• Literature maturity/gap mapped: diagnosis is mature, true RUL prediction is less mature, and integrated optimization remains emerging.
• Validation comparison and KPI set specified for fixed PV vs conventional tracking vs AI health-aware control.

Tangible outputs & evidence

• Paper 1 systematic-review scope, research gap and evaluation framework defined

Key result

What should we notice?

• The research gap is not another tracker or fault detector; it is joint decision-making across energy and asset health.
• Tracker movement is worthwhile only when expected energy gain exceeds actuation, wear and failure-risk costs.
• The target is higher net yield and reliability, not maximum irradiance capture alone.

Next steps

1. 1

   Finalize SLR protocol, databases, search strings and inclusion/exclusion criteria.

   05 Oct 2026

2. 2

   Run literature search, de-duplicate results and begin screening across the review themes.

   15 Oct 2026

3. 3

   Draft evidence synthesis, research-gap statement and joint-optimization conceptual framework.

   25 Oct 2026

Success by next forum: SLR evidence map + conceptual framework ready

Risks & blockers

• Scope may become too broad across tracking, weather, PV faults, tracker mechanics and maintenance.
• Synchronized electrical, weather, imagery and tracker-condition data may be difficult to access.
• Phase the work: SLR → subsystem models → integrated controller; validate with benchmark/simulated data before field deployment.

Support needed

• Supervisor feedback on scope boundaries and SLR protocol; guidance on representative PV/tracker datasets or field access.

Possible Research Grants (Proposed)

• NRF Competitive Research Programme (CRP)
• MOE Academic Research Fund Tier 2
• AI Singapore — Research / Technology programmes