FROM SIMULATION TO INVESTMENT EVIDENCE: MULTISCALE VALIDATION AND TRANSPARENT TECHNO-ECONOMIC ASSESSMENT OF UTILITY-SCALE AND INSTITUTIONAL SOLAR PV SYSTEM

Abstract

Photovoltaic project assessment requires an evidence framework in which simulated energy yield, measured generation, and economic performance remain explicitly traceable to their underlying data and assumptions. This study develops such a framework through a comparative assessment of two photovoltaic configurations in Pakistan: a 100 MWp utility-scale system (QASP/LSPP) and a 405 kWp institutional system (UOL/SSPP). The analysis integrates PVsyst energy predictions with reported generation records and project-level economic indicators, while distinguishing directly reported quantities from independently reconstructed financial metrics. For QASP/LSPP, twelve paired monthly observations covering September 2015–August 2016 are evaluated using correlation, goodness-of-fit, mean absolute error (MAE), root mean square error (RMSE), and paired statistical testing. The annual PVsyst estimate of 155,276 MWh is 2.48% lower than the reported generation of 159,216.63 MWh. Despite this close annual agreement, monthly correspondence is only moderate (Pearson r = 0.690; R² = 0.476), with MAE and RMSE of 1,312.24 and 1,508.84 MWh, respectively; paired tests do not indicate a statistically significant systematic difference at the 5% level. For UOL/SSPP, the reported annual generation is 700 MWh compared with 688 MWh from PVsyst, corresponding to a 1.71% model shortfall. However, only three dated paired observations are available, precluding a statistically comparable monthly validation. The reported project-level LCOE values are Rs 4.26/kWh for QASP/LSPP and Rs 5.53/kWh for UOL/SSPP. To examine the reproducibility of these economic indicators, an independent lifecycle reconstruction using a 25-year project horizon, an 8% discount rate, linear degradation to 80% of initial output by year 25, and the documented cost inputs yields LCOEs of Rs 12.32/kWh and Rs 9.96/kWh, respectively. The substantial divergence between reported and reconstructed LCOE values demonstrates that economic conclusions are highly sensitive to the treatment of degradation, discounting, operating expenditure, tariff assumptions, financing structure, and other omitted cost components. The findings therefore establish an important distinction between annual energy agreement and temporal model fidelity and demonstrate that photovoltaic techno-economic claims are reproducible only when measurement records, simulation outputs, and financial assumptions are maintained within a transparent and auditable evidence chain.