ABSTRACT
This chapter considers, as a case study, the problem of fault diagnosis of electric power transformers. It analyzes feedforward neural networks (FNNs) and explains their use in nonlinear systems modeling. FNNs with Gauss–Hermite activation functions are introduced and their distinctive properties, such as orthogonality of the basis functions and invariance to Fourier transform, are explained. The chapter presents basic principles of signals spectral analysis and explains the use of Fourier transform in calculating a signal’s energy content and power spectral density. It analyzes modeling issues for condition monitoring of electric power transformers and explains the use of neural networks with Gauss–Hermite basis functions in modeling the thermal dynamics of electric power transformers. The chapter also shows how these neural networks enable spectral analysis of the hot-spot temperature signal and how based on this signal’s spectral content one can perform fault detection and isolation for the power transformer.
