Robust Stability in Multivariable Neural Network Control Using Harmonic Analysis
Robust stability and performance are the two most basic features of feedback control systems. The harmonic balance analysis technique enables to analyze the stability of limit cycles arising from a neural network control based system operating over nonlinear plants. In this paper a robust stability analysis based on the harmonic balance is presented and applied to a neural based control of a non-linear binary distillation column with unstructured uncertainty. They develop ways to describe uncertainty in the form of neglected nonlinear dynamics and high harmonics for the plant and controller respectively.