Hot Spot Diagnosis And Dust Quantity Prediction Of Photovoltaic Modules

At present,wind energy and solar energy are the most valuable renewable energy sources.Solar energy has a high degree of competitiveness in power generation compared to other energy sources.Photovoltaic modules were used to complete photoelectric conversion,which has hot spots and ash problems during operation.Hot spots can cause local photovoltaic modules to reduce power generation efficiency,in severe cases,burn components.The problem of ash accumulation can lead to a decline in power generation efficiency of large-area photovoltaic modules,causing significant economic losses.In response to the above problems,the research on hot spot diagnosis and dust quantity prediction of photovoltaic modules was carried out.The specific research contents are as follows.Firstly,the ameliorate isolation forest(AIF)algorithm was proposed to solve the problem of hot spot diagnosis.The gray relational projection algorithm was used to select the similar day.Since the various feature quantities of the hot spot have different dispersion and characteristic characteristics,an improved isolation forest algorithm was proposed to analyze the hot spot feature quantity.The grey relational projection method selects the sample data of similar days,an ameliorate isolation forest algorithm to determine whether each branch has an abnormal situation.The greater the probability of the branch with the higher score,the greater the probability of branch failure.By comparing the isolation forest algorithm(IF),local outlier factor algorithm(LOF),ameliorate isolation iforest algorithm(AIF),the ameliorate isolation forest algorithm(AIF)was proved to have best accuracy and stability.Secondly,in order to improve the hot spot diagnosis of photovoltaic modules,a fuzzy control algorithm based on grey correlation-entropy weight and fuzzy control(GEFC)was proposed to improve the accuracy of hot spot diagnosis.The gray correlation-entropy weight algorithm was used to calculate the weight of each feature quantity on the hot spot,then the ratio of the hot spot fault voltage/current was different to other faults are used.Lastly,the fuzzy control algorithm was used for hot spot diagnosis.The experimental results show that compared with the fuzzy control algorithm,the GEFC algorithm improves the accuracy of photovoltaic hot spot diagnosis.Thridly,the problem of large-scale power generation efficiency caused by ash accumulation of photovoltaic modules was settled,an indefinite time cleaning strategy for photovoltaic modules was formulated.The wavelet neural network prediction model was used to predict the amount of dust accumulated by the photovoltaic modules,the prediction results combined with the weather conditions in the next few days,then the next cleaning time was determined.Finally,the spot-hot diagnosis and ash cleaning of PV modules system hardware configuration and software development were completed based on B&R AS4.0.AIF algorithm and EGFC algorithm were applied.The experiment verifies the correctness and effectiveness off the hot-spot diagnosis algorithm and the economics of photovoltaic power generation was improved at same time.