Forecasting solar power output in Ibadan: A machine learning approach leveraging weather data and system specifications

2025-08-10

AI Summary

Accurate solar irradiance forecasting is critical for solar power planning and grid stability in tropical monsoon regions like Ibadan, Nigeria, yet conventional methods suffer from high dependency on expensive ground-based pyranometers. Method: This study proposes a physics-informed, data-driven two-stage irradiance forecasting framework. Stage one integrates clear-sky models with cloud-type classification to improve hourly global horizontal (GHI), direct normal (DNI), and diffuse horizontal (DHI) irradiance predictions. Stage two employs PVLib to estimate photovoltaic power output using meteorological inputs, cloud conditions, and system-specific parameters. Random forest, CNN, and LSTM models are comparatively evaluated. Contribution/Results: Random forest achieves the best performance, yielding annual normalized root-mean-square errors (nRMSE) of 0.19 (GHI), 0.33 (DNI), and 0.22 (DHI); during the dry season, GHI nRMSE drops to 0.12. The framework significantly reduces reliance on ground measurements and offers a scalable, low-cost solution for high-accuracy solar forecasting in tropical monsoon climates.

Technology Category

• Intelligent Robots: State Estimation
• Planning, Routing, and Scheduling: Plan Execution and Monitoring
• Machine Learning: Classification and Regression

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

This study predicts hourly solar irradiance components, Global Horizontal Irradiance (GHI), Direct Normal Irradiance (DNI), and Diffuse Horizontal Irradiance (DHI) using meteorological data to forecast solar energy output in Ibadan, Nigeria. The forecasting process follows a two-stage approach: first, clear-sky irradiance values are predicted using weather variables only (e.g., temperature, humidity, wind speed); second, actual (cloudy-sky) irradiance values are forecasted by integrating the predicted clear-sky irradiance with weather variables and cloud type. Historical meteorological data were preprocessed and used to train Random Forest, Convolutional Neural Network (CNN), and Long Short-Term Memory (LSTM) models, with Random Forest demonstrating the best performance. Models were developed for annual and seasonal forecasting, capturing variations between the wet and dry seasons. The annual Random Forest model's normalised Root Mean Square Error (nRMSE) values were 0.22 for DHI, 0.33 for DNI, and 0.19 for GHI. For seasonal forecasts, wet season nRMSE values were 0.27 for DHI, 0.50 for DNI, and 0.27 for GHI, while dry season nRMSE values were 0.15 for DHI, 0.22 for DNI, and 0.12 for GHI. The predicted actual irradiance values were combined with solar system specifications (e.g., maximum power (Pmax), open-circuit voltage (Voc), short-circuit current (Isc), and AC power (Pac)) using PVLib Python to estimate the final energy output. This methodology provides a cost-effective alternative to pyranometer-based measurements, enhances grid stability for solar energy integration, and supports efficient planning for off-grid and grid-connected photovoltaic systems.

Problem

• Predict hourly solar irradiance components using weather data
• Forecast solar energy output with machine learning models
• Enhance grid stability for solar integration in Ibadan

Innovation

• Two-stage forecasting with clear-sky and cloudy-sky irradiance
• Machine learning models: Random Forest, CNN, LSTM
• PVLib Python for solar system energy estimation