STEP FORM DETECTION: A COMPARISON OF CURVE FITTING TECHNIQUES

H. Jokinen, H. Ihalainen, M. Moisio

Abstract

Controlling of process flow is problematic if the valves do not behave ideally. The non-ideal behaviour of a valve causes the valve position to differ from the action given by the control system, which produces discontinuities in the process flow. This is shown as stick-slip steps in the measurement signals. A curve fitting technique is used in step form detection. In this paper we compare this curve fitting technique to well-known Haar wavelet method and present their ability in reliable step form detection.

Keywords

signal processing, stick-slip steps, Haar wavelets

Download

IMEKO-WC-2000-TC10-P265.pdf

DOI

-

IMEKO TC

TC10 - Measurement for Diagnostics, Optimization and Control

Back to the proceedings