Heat Fault Diagnosis Of Transmission Line Joints Based On Infrared Detection Technology

Infrared diagnosis is through the detection of infrared energy radiated from the surface of transmission line and then translates the energy signals into visual signals which are the thermal images to gain the heat characteristics of transmission line. Based on the characteristics obtained and in accordance with the appropriate criterion, we can find out whether there is a heat fault, in what position and how serious the fault is. Heat fault of transmission lines joints is of high failure rate, difficult to find and greatly detrimental. So a safe, fast, efficient heat fault diagnostic method is needed. Infrared diagnosis with the merits of non-contact, passive detection, high-precision, easy implementation, plays a crucial role in the heat fault detection of transmission line joints in recent years.Although there are a lot of advantages, infrared diagnosis also has its own limitations. The research work in diagnosis standardization, diagnosis mathematics, and intelligent diagnosis is extremely lacking. Moreover, in the actual detection environment, the reception of infrared radiation signals of transmission line joints will be interfered by many factors. So the detection accuracy usually can not be guaranteed. These factors will result deviation between the true state of heat features of the joints and the heat characteristics obtained by the infrared thermal imaging system which make the diagnosis still deliberate and empirical.A circuit approximating the heat fault of transmission line joints will be designed and a quantitative research of detection range and load state which are two of the external factors affecting the diagnostic accuracy is made. This work essentially can be classified in diagnosis standardization. First of all, according to the experimental circuit designed, a lot of thermal images will be collected under different detection range and load voltage. Then a reference thermal image whose impact of transmission attenuation and load voltage can be neglected will be selected from the thermal images collected. Finally, based on the maximum temperatures the thermal images show, the temperature compensation curves of distance and load will be plotted.An integrated and effective thermal image denoising and enhancement algorithm will be presented. The algorithm will make a combination between wavelet analysis and traditional spatial and frequency domain analysis. This work can also be attributed to diagnosis standardization. The thermal images collected in actual field are usually blurred and mixed with various types of noise and contrast is low. Denoising and enhancement have been an important aspect in the heat fault detection. Because of multiresolution analysis properties of the wavelet analysis, it is effective for the removal of noise. Spatial and frequency domain analysis can effectively enhance the contrast and smooth the image. The algorithm of denoising and enhancement proposed will absorb the advantages of different analysis domains.A fault identification mathematical model of transmission line joints which can not only determine whether there is a heat fault, but also determine the severity of the failure will be established. This can be attributed to diagnosis mathematics. In order to promote the fault diagnosis from the current experience level to the theoretical level and to reduce the impact of human nature, it is necessary to establish a unified and effective model of fault identification. The mathematical model will be presented on the basis of the temperature compensation aforementioned and the analysis and processing of the thermal image.The innovative work contains three aspects: firstly, the circuit approximating the heat fault of transmission line joints is designed and a quantitative research of detection range and load state affecting the thermal image is made. Secondly, a local area histogram equalization algorithm based on linear smoothing filter is presented and wavelet analysis and traditional space-frequency domain analysis are combined to build a comprehensive and effective thermal image denoising and enhancement algorithm. Thirdly, a fault identification mathematical model proved effective by experiments is established.