A Preprocessor for Analog Circuit Fault Diagnosis Based on Prony's Method
The paper presents a new technique for analog circuit fault diagnosis. The circuit under test is simulated both at fault free and different faulty conditions using PSPICE software. The AC response to a set of sinusoidal input frequencies is calculated at selected test nodes. Prony's method is then utilized as a preprocessor to extract an optimal set of features representing nodal voltage waveforms. The resultant features are used to train a back-propagation neural network to identify circuit faults. The potential of the algorithm is demonstrated by a second order active circuit example.
Document Type: Original Article
Affiliations: 1: Electronics and Communication Department, Cairo University, Giza, Egypt. E-mail: firstname.lastname@example.org 2: Engineering Mathematics and Physics Department, Cairo University, Giza, 1221, Egypt. E-mail: email@example.com
Publication date: January 1, 2003