研究目的
To quickly and accurately diagnose the type of failure of the PV array and implement online monitoring of the PV array using BP neural network.
研究成果
The BP neural network fault diagnosis model designed and trained by the proposed method has high precision, with an accuracy rate of 97% when the number of hidden layer nodes is 5. However, due to the complexity of the PV array's working environment, further improvements are needed.
研究不足
The complexity and variability of the PV array output characteristic curve information seriously affects the accuracy of the PV array intelligent diagnosis algorithm. The intelligent fault diagnosis algorithm based on BP neural network needs further improvement.
1：Experimental Design and Method Selection:
The BP neural network is proposed for PV array fault diagnosis, utilizing a network search method during training and the K-cross-validation method for selecting the number of hidden layer nodes.
2：Sample Selection and Data Sources:
A 3×4 photovoltaic array was built to collect data under normal, insufficient irradiation, open circuit, and short circuit conditions.
3：List of Experimental Equipment and Materials:
PV array IV curve tester (model IVT-30-1000 by Kewell Company) was used for data acquisition.
4：Experimental Procedures and Operational Workflow:
Data was collected under various conditions, normalized, and then used to train and test the BP neural network model.
5：Data Analysis Methods:
The model's accuracy was tested using a separate test set, and the number of hidden layer nodes was optimized using network search and K-cross-validation.
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
