Classification method of photovoltaic array failure grade based on infrared image

Abstract

A photovoltaic array fault classification method based on infrared images, collecting infrared image information of photovoltaic array systems, including multiple infrared images of photovoltaic modules, each image is divided into multiple image units and the R, G, and B values ​​of each unit, These constitute the tensor of the image information database, and according to the obtained temperature information of the photovoltaic module, the infrared image is fault coded, and the Tensor-V hybrid partial least square method is used to predict the fault level vector through the tensor of the infrared image information database, thereby establishing a tensor The fault level division model of quantity-vector has a strong predictive ability, which can realize fast and effective fault level coding of infrared images of photovoltaic modules, and provide it to the operation and maintenance staff of photovoltaic power plants to arrange maintenance and replacement of photovoltaic modules with problems in time , so as to realize the continuous optimal operation of photovoltaic modules and ensure the economic benefits of photovoltaic power plants.

Description

technical field [0001] The invention relates to a method for classifying fault levels of photovoltaic arrays, in particular to a method for classifying fault levels of photovoltaic arrays based on infrared images. Background technique [0002] As a sustainable green energy, solar energy has the advantages of cleanliness and high efficiency. As a big country in energy production and consumption, the development of solar photovoltaic power generation is an ideal choice for my country to solve energy demand and optimize energy structure. In the process of photovoltaic power generation, photovoltaic cell components convert solar energy into electrical energy, which is the main factor affecting the efficiency of photoelectric conversion. With the rapid development of the photovoltaic industry, problems such as photovoltaic array operation status detection are gradually exposed. Due to the component packaging process and other reasons, when the component has been in operation fo...

AI Technical Summary

Problems solved by technology

This method needs to consider environmental information such as illumination, ambient temperature, and wind speed from different types of sensors. Therefore, the manual data collection in the early stage has a long working cycle and heavy workload, and the data preprocessing is more cumbersome, and the model construction is complicated. Analysis and diagnosis of a single image cannot achieve rapid analysis and diagnosis of a large number of images

Images