IV characteristic and deep residual network-based photovoltaic array fault diagnosis method

Abstract

The invention relates to an IV characteristic and deep residual network-based photovoltaic array fault diagnosis method. The IV characteristic and deep residual network-based photovoltaic array faultdiagnosis method comprises the steps of firstly, building a model array by Simulink, and acquiring electrical data and environmental data under various working conditions; secondly, removing abnormaldata in original simulation data, acquiring an original I-V curve, performing sampling, and splicing one-dimensional characteristic to two-dimensional characteristic used as total characteristic of afault; thirdly, dividing sample data to a training set, a verification set and a test set, designing a network structure of residual convolution neural network with dimensional conversion and a training parameter of a training algorithm Adam, and performing sample training to obtain a DT-ResNet fault diagnosis training model; and finally, detecting and classifying photovoltaic power generation arrays under the test set in a to-be-tested working condition by the DT-ResNet fault diagnosis training model, and diagnosing a fault type. The method has the advantages of high accuracy, rapid convergence, high robustness, wide generalization capability and the like, and fault detection and classification accuracy of the photovoltaic power generation array can be effectively improved.

Description

Technical field [0001] The invention relates to a photovoltaic power generation string fault detection and classification technology, in particular to a photovoltaic array fault diagnosis method based on IV characteristics and a deep residual network. Background technique [0002] As a clean and renewable new energy source, solar energy has received extensive attention and use in recent years. According to the latest announcement of the World Energy Organization, global photovoltaic installed capacity and power generation have increased annually. As of the end of 2017, the global installed capacity of photovoltaic power plants has reached 399,613MW, and the global photovoltaic power generation has increased to 442,618GW. However, currently installed and deployed photovoltaic power plants often encounter problems such as violent installation or irregular installation. In addition, photovoltaic power plants work in harsh outdoor environments all year round, and they are easily aff...

AI Technical Summary

Problems solved by technology

[0005] At present, there is no research on applying dimensionally transformed residual convolutional neural network algorithm to fault diagnosis and classification of photovoltaic power generation arrays in published literature and patents

Images