A Fault Detection Algorithm for Battery Panels in Large Photovoltaic Power Stations

Abstract

A fault detection algorithm for a battery panel in a large-scale photovoltaic power station, wherein a manipulation command is directly sent by an upper computer; an acquired current of each branch and voltage of each panel at a current moment and environmental parameters comprising light intensity and temperature are computed after data is input; wherein by using a horizontal comparison algorithm for locating a single failed panel, a historical comparison algorithm for degradation judgement of the whole plant, and a standard value updating algorithm used as reference, and by using a record table of single faults (112) of an auxiliary algorithm, a record table of historical averages (113) and a look-up table of standard values (111), single fault locating and overall degradation or deposition judgement are conducted at the same time. The fault detection algorithm can realise on-line real-time monitoring, is easy to implement, and can avoid misjudgement due to soft errors. Moreover, the look-up table of standard values (111) is built in an adaptive mode, and can be adapted to different geographical and environmental conditions, thereby having greater universality and higher precision.

Description

technical field [0001] The invention belongs to the technical field of photovoltaic power generation, and relates to a fault detection algorithm for a battery panel in a large-scale photovoltaic power station, in particular to a method for detecting and locating electric panel faults and judging overall aging or dust accumulation in a large-scale photovoltaic array. Background technique [0002] With the continuous development of new energy technologies, the application of solar photovoltaic power generation has become more and more extensive. The prior art discloses that the photovoltaic array is the core component of the photovoltaic power generation system, and it is composed of several photovoltaic cell panels connected in series and parallel. According to the data, the current common photovoltaic power plants are composed of large-scale photovoltaic arrays placed outdoors. Due to environmental factors, photovoltaic panels are very likely to have individual failures, agi...

AI Technical Summary

Problems solved by technology

According to the data, the current common photovoltaic power plants are composed of large-scale photovoltaic arrays placed outdoors. Due to environmental factors, photovoltaic panels are very likely to have individual failures, aging or dust accumulation. Due to the large number of photovoltaic panels, it is often impossible to detect them in time. Faults, cannot be replaced in time, seriously affecting the power generation efficiency of photovoltaic power plants

[0003] Based on the above situation, technicians in the industry use some fault detection methods of photovoltaic panels to detect by collecting parameters such as voltage and current of photovoltaic panels. However, practice shows that the data of collected panel parameters cannot directly draw the conclusion of the fault situation , it is also necessary to carry out reasonable and effective diagnostic analysis on the data in order to obtain accurate fault location and fault type

There are still following limitations in the fault detection algorithm of the photovoltaic power generation system of the prior art, as: in the technology of wherein a kind of category, judge by the method of current difference (document , the 14th The National Academic Conference on Equipment Monitoring and Diagnosis, 2010), find out the branch circuit containing faulty components by screening out the adjacent columns with excessive current difference, and analyze the voltage value of each photovoltaic module in the branch circuit, and locate Faulty photovoltaic cell components; this algorithm can be realized by real-time measurement, and only needs to measure voltage and current information, the calculation amount is small, and it is easy to implement. Situations such as occlusion are prone to misjudgment, and only horizontal comparisons between branches are carried out, which lacks analysis of the overall working conditions of the power station in combination with environmental parameters, and cannot reflect the overall aging or dust accumulation of the power station; , use known data for training, obtain a complete database, and then compare the algorithm with the measured data. Chinese patent (application number 201410449777.6) discloses a method for fault judgment using fuzzy clustering algorithm, which is characterized by a large amount of previous work Establish a fault knowledge base, and then collect the fault alarm information of photovoltaic power plants as samples to be compared. By comparing the membership degrees, the fault type with the highest membership degree is selected as the fault type represented by the sample to be tested. The problem with this algorithm is that it needs to be judged in advance Only when a certain set of data is fault data can the membership degree algorithm be used, and the fault cannot be directly judged by electrical parameters such as current and voltage that are easy to test; and the algorithm requires a large amount of training data in advance to establish a fault knowledge base for comparison. There are many types of faults, and the simulation of faults is more complicated, which will consume a lot of resources; and, because the collection of fault knowledge base is limited, and the update of fault knowledge base takes a long time, it is difficult to adapt to various geographical environments. After the geographical environment changes, the content of the fault knowledge base will face inaccurate problems; there are also types of technologies that use neural network models for fault judgment algorithms, with maximum power point current, maximum power point voltage, short-circuit current, and open-circuit voltage And other electrical parameters are used as the input parameters of the algorithm, and the fault model output is obtained through the BP neural network, thereby judging the fault type (document "A four-parameter photovoltaic module online fault diagnosis method", Proceedings of the Chinese Society for Electrical Engineering, May 5, 2014, Volume 34, Issue 13), (document "Research on Fault Diagnosis of Photovoltaic Array Based on BP Neural Network", Power System Protection and Control, August 16, 2013, Volume 41, Issue 16), this method can be directly passed through electrical The parameters determine the type of fault, but the existing problems are: the algorithm itself is relatively complex, with a large amount of calculation, and also requires a large amount of fault data for training; in addition, the short-circuit current and open-circuit voltage input into the algorithm cannot be measured in real time, and the photovoltaic array needs to be temporarily stopped Work, affect the power generation efficiency of the power station, etc.

Images