- By:
- R pesante colón, Marcos; I cruz salamán, Alberto; Cruz figueroa, Dylan; Sundararajan, Aditya ; Ferrari Maglia, Maximiliano F
- Page Number:
- 1-5
- Book Title:
- 2024 IEEE Power & Energy Society General Meeting (PESGM)
- Publication Date:
- March 12, 2026
- Publisher Location:
- IEEE, New Jersey, United States of America
- Conference Name:
- 2024 IEEE PES General Meeting (PESGM)
- Conference Location:
- Seattle, Washington, United States of America
- Conference Sponsor:
- IEEE
- View DOI Listing:
- https://doi.org/10.1109/PESGM51994.2024.10689007
Abstract
Photovoltaic (PV) generation is a critical component of microgrids, but its accurate modeling is challenging due to the complex and dynamic interactions between solar irradiance, temperature, and PV system installation. This paper develops a multilayer perceptron (MLP) model that inputs solar irradiance and temperature to estimate the PV generation, and it compares the proposed data-driven model’s performance to two well-known physical models: the single-diode model and the inverter model. The results demonstrate that all the models can reach high levels of accuracy. However, the MLP model outperforms the physical models on average by 4.5 to 6.6 percent in R squared scores and 220 to 290 Watts in RMSE scores, and it does not require physical system parameters. Moreover, the data-driven model can overcome the limitations of the lack of real-time PV generation data.
