A user interactive tool for assessment of performance ratio for commercial solar photovoltaic system: Leveraging exergy and energy based inputs

Abstract

The practice of prediction and early estimation of Performance Ratio (PR) for grid integrated Photovoltaic (PV) system is critical for power reliability, techno-economic viability and regulatory compliance for plant owners and grid system operators. Current approaches for its estimation remain as a mathematical framework and can be employed only when the set of dependent monitored attributes are made available. Also, these derived system inputs are often challenging to assess or priorly estimable. Nevertheless, a classified approach that relies on pre-estimable factors concerning the electrical and thermal behaviour of PV plants can effectively and accurately assess its on-field performance. So, the presented investigation develops a user-friendly deep-learning based predictive tool for prediction/short-term estimation of PR encompassing novel thermo-electric attributes namely failure mode-based power degradation rate (R_d) and thermal exergy loss. The proposed approach is derived from a lager sample of minute-based observations ranging for an annual duration, belonging to a realistic 191.9 kW_p PV plant situated at Khopoli, India. The developed optimized Long Short-Term Modeler (LSTM) operates with a training and testing accuracy of 91.68 % and 90.61 % respectively. This is further transformed into a user interactive tool employing Tkinter in python. The predictor exhibited a highest prediction accuracy with least Mean Absolute Percentage Error (MAPE) of 0.0183 on comparing it with benchmark-based models like normalized ratio method, PVsyst, corrected PR and an existing model learning approach. It is also validated for a roof-top PV facility at Bengaluru, India and Koprübaşı, Turkey showing an MAPE as low as 5.81 % and 1.48 % respectively, in comparison to existing methodologies. So, the proposed PR analyser increases user interaction and is an accurate tool benefiting stakeholders in Solar PV Industry.