Modelling and real time performance evaluation of a 5 MW grid-connected solar photovoltaic plant using different artificial neural networks

• Published in: Energy conversion and management Vol. 279; p. 116767
• Main Authors: Narasimman, Kalaiselvan, Gopalan, Vignesh, Bakthavatsalam, A.K., Elumalai, P.V., Iqbal Shajahan, Mohamed, Joe Michael, Jee
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.03.2023
• Subjects: ANN; Capacity utilization factor; Cleaning factor; Performance ratio; Shading loss; ANN; Cleaning factor; Shading loss; Capacity utilization factor; Performance ratio
• ISSN: 0196-8904; 1879-2227
• DOI: 10.1016/j.enconman.2023.116767

[Display omitted] •Modeling of 5 MW PV plant and analyzes its performance parameter using different.•ANN model.•Experimentally analyze the performance of 5 MW photovoltaic (PV) utility-scales with one year dataset.•Increase the Performance Ratio (PR) and Capacity Utilization Factor (CUF) values by 48.5% and 9.3% using optimum tilt.•Increase average power gain of plant by 5.4% using a novel spray cleaning. Renewable energy sources have gained greater importance in the past decade owing to the continuous growth in energy demand of developing nations. The modular nature of solar energy has made it possible to utilize it in a variety of applications without any adverse economic impact. In this research, a 5 MW grid-connected solar PV plant was modelled using various ANN networks namely Cascade forward backdrop, ELMAN back prop, Feed forward back prop, Feed forward Distribute Time delay, Layer current and Regression Analysis to identify the impact of various environmental parameters on the PV power generation. Comparatively the Lavenberg- Marquardt model is obtained as a reliable model for smart grid system especially for solar PV power prediction with high R value of 0.9966. Also, the plants real time annual energy production, performance ratio (PR), and capacity utilization factors (CUFs) were analysed under various geometrical condition. The shading losses of the entire PV plant was evaluated using PV SYST software at different tilting conditions (seasonal and fixed tilt). Further, the annual PV generation and energy consumption of the plant is calculated under various load conditions. Out of the total energy consumption of the plant, only 11.2 % of the load will be handled by the 5 MW solar PV plant. Compared with fixed tilt, the proposed optimized tilt increased PR and CUF values to 48.5 % and 9.3 % for the months of November and December, respectively (excluding Sunday maintenance). During Sunday, energy loss was nearly 5248 kWh/day due to improper shutdown of power control units. With proper supply management schemes, the average power saving was nearly 48,000 kWh/year and the total cost saving was nearly INR 38 lakh per year. Owing to proper cleaning (twice in a month), the average power gain was increased to 5.4 %.