Research On Fault Diagnosis Of Photovoltaic Power Generation Based On DSP

As a clean and inexhaustible energy,solar energy is rapidly promoted and applied.Grid-connected photovoltaic(PV)generation system is the major application of solar energy,which mainly includes PV array,grid-connected inverter and utility grid.Among them,PV array is the core of the entire power generation system.Inevitably,various faults will occur in PV array due to the harsh external working environment.These faults can severely reduce the output power of the PV array and even cause a fire.Therefore,it is necessary to implement fault diagnosis on the on-line monitoring system for PV array.Accordingly,an online monitoring system for PV array based on DSP with fault diagnosis is designed.Firstly,a fault location method based on the current of PV string is proposed.At first,a current sliding window(CSW)is applied according to the current of PV string.Then,the local outlier factor(LOF)of each current point in the CSW is calculated,and the outliers is judged by comparing their LOF with the predefined threshold.Once the number of consecutive outliers exceeds the preset parameter,the PV string can be judged as fault.Simulation and experimental results show that the proposed fault location method can locate short-circuit,partial shadow and open-circuit fault strings under various irradiance.Secondly,the fault classification method is studied based on strong classifier of BP_Adaboost.The current of PV string,voltage of PV array,irradiance and temperature are selected as input attributes for fault classification model.The algorithm obtains the classification results of 10 BP neural networks firstly,and then combines these 10 BP neural networks into the final output with corresponding weights.The weight of BP neural network is directly proportional to its classification accuracy.The fault classification method can identify the normal.short circuit,partial shade and open-circuit conditions of PV array.The feasibility and accuracy of the fault classification algorithm is both verified by simulation and experimentation.Finally,a PV array fault on-line monitoring system is developed with the DSP as the system’s control core.The whole system can realize fault string location,fault classification and display functions.System collects the string current,voltage of PV array,irradiance and temperature through the PV data acquisition circuit.Then the DSP processor converts the PV data to digital format,locates faulty string and sends results and PV data.Finally,a PV fault on-line monitoring center is designed by using MATLAB to realize the classification and display functions.The test results indicate the designed system can realize the fault location,working condition identification of PV array and online monitoring of PV system.Consequently,the technologies in this study can provide some ideas for the development of online monitoring and fault diagnosis of PV arrays.