Research On Fault Detection Of PV Array Based On Composite Data Fusion

There are different messages generated by the photovoltaic power system when it is running. These messages reflect the running state of the system from every aspect. The faults of the photovoltaic power system are divided into three kinds, the first is sliver of the solar panel, the second is aged deterioration, and the third is hot spot. The most important problem to keep the photovoltaic power system in normal state is state detecting and fault diagnosis. At present, the common technologies of fault diagnosis are the infrared picture analysis and the compare of difference value about current and voltage. However, the detection accuracy and efficiency are in a low level when using these technologies.According to the contrast and analysis of the photovoltaic array’s connecting structure, this paper proposed a kind of new photovoltaic array’s connecting structure which is named Sensor Network Total Cross Tied Array, SN-TCT for short. According to the framework of fault diagnosis based on complex data fusion, this paper applies the algorithm of complex data fusion to the fault diagnosis and monitoring of photovoltaic power system. This paper divides the fault diagnosis system into two parts, one is local fusion module, and the other is the global fusion module. We combine the local fusion part and the global part into a whole system. Adaptive weight data fusion was adapted in the local fusion, this algorithm was used to obtain the exact value of the variables, such as the value of voltage, current, the temperature of the environment and the solar illumination. The evidence theory was used in the global fusion, we merge the difference information between the measured value and the theoretical value, the measured value and the average value by side respectively, including the electrical information (voltage and current), solar illumination, environmental temperature and the speed of wind. Finally we reach the goal that recognizing the faults which include the sliver of the solar panel, the aged deterioration, and hot spot. The veracity and reliability were improved in the PV system.A photovoltaic cell is used to verify the effectiveness of our proposed method through the fault signatures which are collected by the data collecting card. What’s more, we finished the hardware circuit of the faults diagnosis system of the large scale photovoltaic system using Field Programmable Gate Array. Under the condition of full exposure and overshadow, we did an experiment, the results show that our method has high accuracy and reliability for the fault diagnosis system, it can locate the faults point of the photovoltaic array quickly.