| The detection method of cable core resistance under the condition of constant temperature offline steady-state temperature field can effectively solve the problem of cable quality detection.However,due to the uncontrollability of temperature factor during on-line detection in the production process of stranded cable core,the measurement method based on steady-state temperature field can’t be effectively applied.Concerning this issue,this paper takes BV50 cable as the research object,the method of obtaining equivalent temperature by analyzing the temperature field of cable core and using it to detect cable resistance is studied.The purpose is to obtain an effective method for power cable resistance detection under on-line unsteady temperature field in production process.In the production process,the inner temperature of the stranded wire core can’t be known under the on-line unsteady temperature field,and the calculation of equivalent resistance directly by the surface temperature will cause a large error.The resistance needs to be temperature corrected,Therefore,the accurate corporation of the temperature field model is the key to solve the problem of on-line detection of wire core resistance.Therefore,a three-dimensional finite element temperature field model of stranded wire core is established to solve the problem of unmeasurable temperature gradient and inner layer temperature of stranded wire core.Because on-line detection needs rapidity,aiming at the problem of calculation accuracy becomes low and long calculation time of finite element under unsteady state conditions,this paper uses BP neural network optimized by genetic algorithm（GA-BP）to calculate the equivalent resistance of on-line detection stranded wire core,to solve the problems existing in the finite element method.Using the experimental data to calculate the core equivalent resistance of finite element method and GA-BP neural network method,it is found that under the on-line steady-state condition,the average absolute percentage errors of finite element method and GA-BP neural network method are 0.183% and 0.256% separately,and the finite element accuracy is higher than that of GA-BP neural network;Under on-line unsteady state conditions,the average absolute percentage errors of finite element method and GA-BP neural network method are 0.581% and 0.265% separately.The accuracy of GA-BP neural network method is higher than that of finite element method.The calculation time of GA-BP neural network method is much lower than that of finite element method.It shows that the calculation method of on-line core equivalent resistance based on GA-BP neural network has better generality and real-time,and provides a practical and effective method for on-line cable core resistance detection. |
