Análise de Wavelet na detecção e diagnóstico de oscilações em malhas de controle de processo

Abstract

One of the main causes of control loop performance degradation are the oscillations, which have a negative effect on the performance of these loops and may force the plant to operate at less than optimal conditions. One of the fundamental steps for the evaluation of industrial control loop performance is the detection and diagnosis of these oscillations, also motivated by the growing emphasis on security and earnings capacity of the installations. This paper uses wavelet analysis combined with other signal analysis techniques such as the autocorrelation function and the Granger Causality, to make the complete diagnosis of oscillations in control loops processes. Numerical test simulations are presented to demonstrate the effectiveness of the proposed method. First, the techniques are used for the diagnosis of a simple control loop in the format of internal model control. After, the methods are applied in a catalytic cracking unit operating under model predictive control (MPC). The results show the potentiality of the proposed methodology to real applications.