Investigation of Data Inaccuracies in the Database of the Solar Photovoltaic Monitoring System

Solar energy is sustainable and never runs out; thus, it's starting to contribute to electricity production. A modern power system's reliable and inexpensive operation depends on an accurate PV generation estimate. Weather factors like temperature, humidity, wind direction, and precipitation affect solar PV electricity generation based on location. There are worries about sensor, communication, or database-based issues that could make the forecasting model less accurate. These problems make integrating green energy into the grid risky. This paper used RMSE, R2, MAE, and MSE to fill in missing values in NREL meteorological data and Thiagarajar College of Engineering units for voltage, current, frequency, and power. MICE Imputation using LightGBM and Iterative Imputer, Backward Fill, Forward Fill, KNN, and Linear Interpolation were used to test on hourly and weekly datasets collected over the year to generate missing data.