Research On Photovoltaic Panel Cleaning And Hot Spot Detection System

In the long-term operation of photovoltaic power plants,they will be exposed to foreign matter such as dust and bird droppings.The foreign matter accumulated on the surface of the photovoltaic panel not only affects the photoelectric conversion efficiency,but also may cause the hot spot effect of the photovoltaic panel due to its partial shielding.The existing photovoltaic panel operation and maintenance device is too large in size,low in efficiency,poor in applicability,and does not have hot spot detection capability.This paper has developed a photovoltaic panel cleaning and hot spot detection system.The main research contents include:(1)Designed and built a photovoltaic panel cleaning and hot spot detection device.Aiming at the problems of excessive volume and poor applicability of the existing photovoltaic panel operation and maintenance devices,a small operation and maintenance mechanism with a PV array mounting groove as a mobile platform was designed,and an additional cross-line auxiliary operation and maintenance mechanism was designed.Operation and maintenance tasks for multi-row photovoltaic arrays.(2)The design and construction of the photovoltaic panel image acquisition system was carried out,and the infrared image preprocessing algorithm was studied.In view of the high noise and low contrast of hot spot infrared images,the differences between different image denoising and enhancement algorithms are compared.Finally,adaptive median filtering and histogram equalization are used to complete image preprocessing,and high quality hot spot infrared images are obtained.(3)Collect a large number of infrared images containing hot spots of different shapes and sizes and create a data set.Aiming at the low accuracy of Faster R-CNN on the dataset of this paper,the method of migration learning,improved feature extraction network and adjustment of RPN anchor frame selection scheme are used to improve Faster R-CNN,which improves the detection rate of hot spot.Experiments show that the cleaning efficiency of the photovoltaic panel cleaning and hot spot detection system after cleaning the contaminated photovoltaic panels three times,the cleaning efficiency reached 99.5%,and the accuracy of hot spot detection on the test set reached 97.34%.