Study And Implementation Of Nonlinear Analog Circuit Fault Diagnosis Based On Volterra Kernel

Now life, analog circuit widely exists in various fields of people’s life, therefore,the fault diagnosis of analog circuits also has been praised by experts research andattention. Nonlinear analog circuits of analog circuit get a lot of obstacles in theaspect of fault diagnosis because of its complexity and diversity. At present, althoughthere are a lot of ways about fault diagnosis theory, but most of all are not enoughand perfect. In this paper, we study and design the nonlinear analog circuit faultdiagnosis system based on the theory of the intelligent diagnosis which is practicaland the accuracy.This article mainly research and design two aspects carries about the intelligentdiagnosis theory and fault diagnosis system. Research on intelligent diagnosis theorymainly includes the calculation of Volterra of nonlinear analog circuits, then selectand extract the feature of the Volterra and generated sample set. In this paper, wedesign an improved BP neural network, which has faster search speed and globaloptimization ability, using the generated sample set to improve neural network fortraining. Fault diagnostic instrument with double channel source collection plan,which using DDS chip to applying the test circuit with single tone signal andmultiple frequency mixed sound signals; Acquisition channel using100M and6Mhigh and low speed acquisition plan, which fully meet the various need of actualmeasurement and collection of data with a high speed USB chip CY7C68013istransmitted to the PC software to carry on the corresponding data processing andcalculation, then take the corresponding diagnosis through the trained neural network,thus completing the process of fault diagnosis. In this paper verified the diagnosissystem using the experimental data and result of diagnosis through the instance test.This article design improves neural network algorithm of the intelligentdiagnosis, which is the improved BP neural network having adaptive fast global optimization characteristics and contributed a lot for improving the diagnosticefficiency. The nonlinear analog circuit fault diagnosis system of this paper designhas characteristic such as high practicability and good accuracy etc. which can makeactual fault diagnosis for a lot of different for analog circuits.