Algorithm Research Of Cable Condition Assessment Based On Oscillatory Waves Detection Signal

With the upgrading of power grids, XLPE power cable gradually become an importantpart of urban power network due to its own advantages. Cable’s reliability is an importantfactor for the safe operation of power grid. However, as the using time prolonged, cableinsulating material is affected by the external environment, and occurring cable insulationdegradation phenomenon. Faulty cable will occurs partial discarge with applied high voltage.The major methods of diagnosing partial discharge are ultrasonic detection, UHF detection,and Oscillation detection. Oscillation detection is the most widely used method owning to itsadvantage of high precision.This paper mainly studies the recognition methods of cable’s insulation fault and themethods of evaluating cable’s health status. Firstly, it proves that the partial dischargeparameters can characterize the health status of the cable. Secondly, it introduces basicprinciples of the Oscillation detection techniques and the measuring methods of partialdischarge which is based on the detection of the Oscillation in detail.Firstly, the paper analyzes the relationship between the characteristics of partialdischarge parameters and type of insulation failure, and proves that q H),( jncan be used asthe feature vector for the recognition of cable insulation fault type. Then, a neural networkclassifier has been established with the q H),( jnas the input and the six type of insulationfaults as the output. The results shows that the neural network classifier has a goodperformance for calssification and it can be used for practical application.Secondly, the paper proposes that the statistical knowledge can be used to analyze theaging states of the cables. Comparing the empirical distribution of aging parameters with fourdifferent theoretical distributions, It founds that lognormal distribution is the best fitteddistribution of the aging parameters.Finally, embedded the insulation fault classification algorithm and the aging stateevaluation algorithm to the intelligent analysis software which is based on the Oscillationdetection system and complete module design, development and integration testing.