Abstract:Effective machine learning regression models are useful toolsets for managing and planning energy in PV grid-connected systems. Machine learning regression models, however, have been crucial in the analysis, forecasting, and prediction of numerous parameters that support the efficient management of the production and distribution of green energy. This article proposes multiple regression models for power prediction using the Sharda University PV dataset (2022 Edition). The proposed regression model is inspired by a unique data pre-processing technique for forecasting PV power generation. Performance metrics, namely mean absolute error (MAE), mean squared error (MSE), root mean squared error (RMSE), R2-score, and predicted vs. actual value plots, have been used to compare the performance of the different regression. Simulation results show that the multilayer perceptron regressor outperforms the other algorithms, with an RMSE of 17.870 and an R2 score of 0.9377. Feature importance analysis has been performed to determine the most significant features that influence PV power generation.

What problem does this paper attempt to address?

The paper attempts to address the problem of accurate prediction of power generation in photovoltaic (PV) systems. Specifically, the authors aim to improve the accuracy of PV power prediction by comparing various regression models to support effective energy management and grid stability. Below is a summary of the main content of the paper: ### Research Background and Problem 1. **Background**: - PV systems are considered an important form of alternative energy due to their ability to generate clean energy. - Accurate PV power prediction is crucial for effective power management, planning, and ensuring grid stability. 2. **Problem**: - PV power generation is influenced by multiple factors, including climate, solar radiation, and system efficiency. - An effective method is needed to accurately predict PV power generation to support decision-making for grid operators and energy suppliers. ### Research Methodology 1. **Dataset**: - The dataset comes from PV panels installed on the rooftops of multiple buildings on the Sharda University campus, spanning from January 21, 2022, to December 31, 2023, with data recorded every 15 minutes. - The dataset includes seven features: hour, power (kW), irradiance (W/m²), wind speed (km/h), ambient/panel temperature (°C), and performance ratio (PR). 2. **Regression Models**: - The authors used various regression models for PV power prediction, including Linear Regression (LR), Support Vector Regression (SVR), K-Nearest Neighbors Regression (KNR), Decision Tree Regression (DTR), Random Forest Regression (RFR), Gradient Boosting Regression (GBR), and Multi-Layer Perceptron Regression (MLP). 3. **Performance Evaluation**: - Metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), R² score, and plots of predicted vs. actual values were used to evaluate the performance of different regression models. ### Main Contributions 1. **Data Collection**: - Real-time series data were collected from PV panels installed on the rooftops of multiple buildings on the Sharda University campus. 2. **Model Comparison**: - A comparative study and performance analysis of various regression models were conducted to identify the best model for PV power prediction. 3. **Feature Importance Analysis**: - Key features affecting PV power generation were identified through feature importance analysis. ### Results - Simulation results show that the Multi-Layer Perceptron Regressor (MLP) performed the best among all models, with an RMSE of 17.870 and an R² score of 0.9377. ### Conclusion - The study validated the superior performance of the Multi-Layer Perceptron Regressor in PV power prediction by comparing various regression models, providing strong support for effective management of PV systems and grid stability. ### Future Work - Further optimization of the model to improve prediction accuracy. - Exploration of more features and data preprocessing techniques to enhance the model's generalization capability. Through this research, the authors hope to provide a reliable method for PV power prediction, thereby promoting the effective utilization and management of clean energy.

Comparative Analysis Using Multiple Regression Models for Forecasting Photovoltaic Power Generation

Improving Photovoltaic Power Prediction: Insights through Computational Modeling and Feature Selection

Analysis Of Solar Power Generation Forecasting Using Machine Learning Techniques

Machine Learning Approaches for Short-Term Photovoltaic Power Forecasting

A Comparison of the Performance of Supervised Learning Algorithms for Solar Power Prediction

The performance comparison of machine learning methods for solar PV power prediction

Deep and Machine Learning Models to Forecast Photovoltaic Power Generation

Comparison of machine learning methods for photovoltaic power forecasting based on numerical weather prediction

Solar Power Forecasting Using Support Vector Regression

Enhanced solar photovoltaic power prediction using diverse machine learning algorithms with hyperparameter optimization

Solar Radiation Prediction in PV Power Systems: a Comparison of Deep Learning Models Using Big Data

Forecasting of photovoltaic power generation and model optimization: A review

Investigating photovoltaic solar power output forecasting using machine learning algorithms

A review and evaluation of the state-of-the-art in PV solar power forecasting: Techniques and optimization

Forecasting Solar Power Generation Utilizing Machine Learning Models in Lubbock

Machine learning models for renewable energy forecasting

Machine Learning Based Solar Photovoltaic Power Forecasting: A Review and Comparison

A hybrid machine learning forecasting model for photovoltaic power

Prediction of daily photovoltaic plant energy output using machine learning

A New Data Driven Long-Term Solar Yield Analysis Model of Photovoltaic Power Plants

Comparative Analysis of Machine Learning Algorithms for Solar Irradiance Forecasting in Smart Grids