Comparative analysis of steady-state and dynamic models for grid-connected photovoltaic farm performance monitoring: A case study of Huawei SUN2000-60KTL-M0 inverter

Abstract

This study evaluates the solar-to-electrical energy conversion performance of a 63.24 kWp grid-connected photovoltaic (PV) farm equipped with a Huawei SUN2000-60KTL-M0 (60 kW) inverter by comparing a steady-state diagnostic model against a dynamic observability assessment. A single-instant snapshot of 12 PV strings on 26 June 2026 identified three non-producing strings (PV6, PV7, PV8), reducing string availability to 75.0% and causing an instantaneous DC power shortfall of approximately 6.0 kW (23.6% of currently achievable capacity). Inverter DC-to-AC conversion efficiency was 98.72% and the observed daily-equivalent capacity factor was 15.48%, both computed directly from measured values. Performance Ratio is reported as not evaluable due to the absence of temporally overlapping plane-of-array irradiance data. The dynamic assessment evaluates day-to-day energy temporal variability and historical hourly weather variability from measured June 2026 and 2020–2025 records (mean absolute daily ramp 12.66%, daily energy CV 19.45%), while sub-minute electrical transient characteristics remain not evaluable owing to the lack of sub-minute electrical logging. A Model Observability Score (MOS) and Data Resolution Adequacy Index (DRAI) are computed from an a priori parameter registry. The steady-state model achieves an MOS of 87.5% (7 of 8 applicable parameters), while the dynamic model achieves 25.0% (4 of 16), with the gap concentrated in the electrical-transient domain and attributable to instrumentation limitations. A phased instrumentation roadmap is proposed. The findings provide a transparent, auditable template for determining appropriate monitoring depth in tropical grid-connected PV operations and maintenance practice.