Photovoltaic module fault diagnosis and prediction method

Abstract

The invention discloses a photovoltaic module fault diagnosis and prediction method. The method comprises the steps of collecting photovoltaic string output power and irradiance data; converting an output power vector and an irradiance vector into frequency domain data; calculating an output power-irradiance amplitude ratio of each frequency point in the frequency domain data, when a certain faultoccurs, selecting a frequency band of which the output power-irradiance amplitude ratio is obviously increased as a characteristic frequency band of the fault, and setting an occurrence threshold ofthe fault; carrying out fault state division on a photovoltaic string according to the fault occurrence threshold value to realize fault diagnosis of the photovoltaic string; calculating to obtain a transfer frequency matrix and a transfer frequency matrix with different step lengths and weights corresponding to the transfer frequency matrix according to the fault state change of the photovoltaicstring; and according to the transfer frequency matrix, the weight corresponding to the transfer frequency matrix and the fault state of the photovoltaic string in each day, calculating the probability that the photovoltaic string is in different fault states in the next day, and predicting the fault state of the photovoltaic string in the next day according to the probability.

Description

technical field [0001] The disclosure belongs to the technical field of photovoltaic power generation, and in particular relates to a photovoltaic module fault diagnosis and prediction method. Background technique [0002] In actual situations, there are few types of data collected by photovoltaic modules, only current and voltage. However, the types of photovoltaic modules are the same and there are many, that is, the output power value (the product of current and voltage) has a great similarity with the change of irradiance. At present, in terms of photovoltaic module fault diagnosis: it is only possible to judge whether a photovoltaic module is faulty based on a simple power threshold setting. For example, the average power of all modules is used as the threshold. It can be considered that a fault has occurred; the problem with this method is that it cannot identify the specific type of fault; in terms of fault prediction: because the power output of photovoltaic modules ...

AI Technical Summary

Problems solved by technology

[0002] In actual situations, there are few types of data collected by photovoltaic modules, only current and voltage. However, the types of photovoltaic modules are the same and there are many, that is, the output power value (the product of current and voltage) has a great similarity with the change of irradiance. At present, in terms of photovoltaic module fault diagnosis: it is only possible to judge whether a photovoltaic module is faulty based on a simple power threshold setting. For example, the average power of all modules is used as the threshold. It can be considered that a fault has occurred; the problem with this method is that it cannot identify the specific type of fault; in terms of fault prediction: because the power output of photovoltaic modules will be affected by many factors, such as basic environmental factors: including irradiance, Ambient temperature, wind speed, ambient humidity, etc.; other external factors: including shading, bird droppings, sand and dust coverage, etc.; and irreversible factors of long-term operation aging: including hot spots, delamination, delamination, etc.; It has great randomness, therefore, it is impossible to distinguish and predict the failure of photovoltaic modules by using traditional methods

Images