Set Membership Estimation of Parameters and Variables in Dynamic Networks by Recursive Algorithms with Moving Measurement Window

Authors:	Brdys Mietek, University of Birmingham, United Kingdom Duzinkiewicz Kazimierz, Gdansk University of Technology, Poland
Topic:	5.4 Large Scale Complex Systems
Session:	Large Scale Complex Systems - Applications
Keywords:	set membership estimation, uncertainty, complexity, networks, water

Abstract

The paper considers a set-membership joint estimation of variables and parameters in complex dynamic networks based on parametric uncertain models and limited hard measurements. The recursive estimation algorithm with moving measurement window is derived that is suitable for on line network monitoring. The window allows stabilising the classic recursive estimation algorithm and significantly improves the estimate tightness. The estimator is validated on a case study water distribution network. Tight set estimates of unmeasured pipe flows, nodal heads and pipe resistances are obtained.