Spectral Analysis of Sampled Nuclear Reactor Plant Data with Heterodyne Digital Filtering Techniques

Abstract

A design procedure for the digital algorithm of a heterodyne power spectral density analyzer using digital components has been developed. The heterodyne filtering, a straightforward technique to yield directly the spectral densities, avoids a number of the disadvantages inherent in the method of determining spectral densities of stochastic signals using the Fourier transform of precalculated correlation functions.

For the design of the required digital low-pass filters in this system, the method of bilinear conformal transformation of the continuous s domain onto the discrete z domain has been used to generate recursive filter difference equations.

The complete digital analyzer, consisting of a group of digital filters with contiguous passbands, has been implemented on a process computer and tested with real plant data at the OECD Halden Reactor.