Multiple parametric fault diagnosis using computational intelligence techniques in linear filter circuit

This paper presents parametric fault diagnosis in analog circuit using machine learning algorithm. Single and multiple parametric fault diagnosis based on Simulation before test approach are considered in this study. A benchmark State Variable Filter circuit has been used as an example circuit for experiment validations. Extreme learning machine based autoencoder (ELM-AE) is proposed in this paper for feature reduction and classification. The transfer function of the benchmark circuit is simulated by performing Monte-Carlo analysis, the features are obtained to form fault dictionary for single and double faults in such way that the dictionary contains unique values for faulty and fault free configuration of the circuit. The input test patterns from the fault dictionary which are required for testing are reduced to improve the classification by autoencoder. Autoencoder is used as pre-processor to learn and reduce the features before performing classification by the extreme learning machine (ELM) algorithm. The results obtained by the proposed algorithm is compared with the other ELM empirical models like extreme learning machine, kernel extreme learning machine and an evolutionary algorithm called self-adaptive evolutionary extreme learning machine to analyze the performance of the proposed algorithm. The experimental results show that the ELM-AE contributes the higher diagnosis accuracy than other ELM models referred in this paper.