Photovoltaic array fault diagnosis method based on composite information

Abstract

The present invention discloses a photovoltaic array fault diagnosis method based on composite information, and belongs to the technical field of fault diagnosis. The method comprises: collecting working state composite information data of a photovoltaic array, and preprocessing the working state composite information data, wherein the working state composite information data comprises working state image data of the photovoltaic array and working state text data of the photovoltaic array; training a pre-established deep convolutional neural network fault classification model by using the working state image data of the photovoltaic array, and after the training is completed, obtaining an image fault classification model; training a pre-established support vector machine-based fault classification model by using the working state text data of the photovoltaic array, and after the training is completed, obtaining a text fault classification model; and fusing the image fault classification model and the text fault classification model by using a logistic regression algorithm to obtain a fusion model, training the fusion model by using the working state composite information data of the photovoltaic array, and after the training is completed, obtaining a photovoltaic array fault diagnosis model based on composite information.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis, in particular to a photovoltaic array fault diagnosis method based on composite information. Background technique [0002] With the rapid increase of fossil fuels such as coal, oil and natural gas, the non-renewable resources will be exhausted one day, and the combustion of fossil fuels will produce a large amount of harmful gases, which is very harmful to the human living environment. Therefore, as a renewable resource, solar energy has become the most ideal renewable energy source because of its inexhaustibility, inexhaustibility, cleanness and environmental protection, and not limited by geographical factors. [0003] The development of solar photovoltaic technology has brought huge economic benefits, but in practical applications, solar photovoltaic arrays will produce various types of failures due to manufacturing or environmental reasons. At present, there are mainly three types of ...

AI Technical Summary

Problems solved by technology

The time domain reflection method is similar to the radar detection method. The input signal is used to enter the input line. When there is an impedance mismatch, a reflected signal will be generated. The fault is detected by comparing the input signal with the reflected signal; the intelligent algorithm collects a large amount of fault data for intelligent detection. The algorithm provides training, although the effect is relatively good, but the data collection has become its biggest obstacle; although the power comparison method is simple, it cannot locate the fault, but can only judge whether the fault occurs; the electrical characteristic detection method uses voltage and current sensors to detect Signal analysis to achieve fault diagnosis, which requires a large number of sensors to achieve signal collection, so there are great limitations

The infrared image diagnosis method is based on the fact that there will be a certain temperature difference between normal and abnormal conditions after a solar panel fails, and the infrared image can just reflect the temperature difference characteristics of the solar panel. At the same time, the infrared image can not only realize the fixed-point detection of the fault but also It is easy to collect, but the infrared image can only judge whether the solar panel is faulty, and the infrared image has no way to identify the type of fault

To sum up, whether it is based on infrared images or text data such as current and voltage, the utilization of fault information has certain limitations, is not comprehensive enough, and has low accuracy.

Images