Synthetic Dataset for Photovoltaic Fault Diagnosis Based on Simulated I–V Curve Images and GASF Images

This dataset provides synthetic photovoltaic (PV) fault diagnosis data generated using a MATLAB/Simulink-based digital twin simulation framework. It comprises 35,000 simulated I–V curve images, 35,000 corresponding Gramian Angular Summation Field (GASF) images, and associated environmental metadata for each sample. The dataset covers seven photovoltaic operating conditions: Normal, Shading, Hotspot, Crack, Short Circuit, Global Aging, and Partial Aging. Each class contains 5,000 samples, resulting in a total of 35,000 samples. The dataset consists of: gasf_images.zip: 35,000 GASF images organized by fault class. iv_images.zip: 35,000 simulated I–V curve images organized by fault class. environmental_metadata.csv: Environmental metadata including sample identifiers, fault labels, irradiance (W/m²), and temperature (°C). The dataset is intended to support research in photovoltaic fault diagnosis, machine learning, deep learning, and renewable energy analytics.