研究目的
To present a sophisticated method based on artificial neural networks (ANN) for diagnosing, detecting, and precisely classifying faults in solar panels to avoid a reduction in the production and performance of the photovoltaic system.
研究成果
The ANN method proved to be more effective than the thresholding method for diagnosing PV module defects, achieving a high detection accuracy. The methodology can be generalized for grid-connected photovoltaic installations or large-scale photovoltaic plants. Future work could focus on real-time detection and diagnostic capacity of large-scale systems.
研究不足
The thresholding method cannot distinguish all faults, leading to potential false alarms. The ANN method requires a large database and periodic training for accurate fault detection.
1：Experimental Design and Method Selection:
The study proposes a method to detect various faults in PV modules using the Multilayer Perceptron (MLP) ANN network. The ANN requires a large database and periodic training for accurate output parameter evaluation. A comparison with the classic thresholding method is conducted to evaluate the approach's accuracy and performance.
2：Sample Selection and Data Sources:
A KC130GHT PV photovoltaic module, containing 36 PV cells with a power rating of 130 W polycrystalline silicon, is chosen for analysis. Simulations are performed using MatLab/Simscape tool.
3：List of Experimental Equipment and Materials:
Matlab SIMULINK for simulation, KC130GHT PV module.
4：Experimental Procedures and Operational Workflow:
The study involves simulating the PV system under normal and faulty conditions, generating I-V and P-V characteristics, and applying two algorithms for fault detection: a threshold-based approach and an ANN-based approach.
5：Data Analysis Methods:
The performance of the ANN approach is analyzed based on a comparison with the classical method of thresholding and other papers. The classification confusion matrix is used to evaluate the ANN's fault detection accuracy.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
