Tracking Simulation Based on PI Controllers and Autotuning

Abstract

A mechanism for automatic update of tracking simulator parameters is suggested. A tracking simulator is a simulator, which runs in real time, and which corrects its own behavior (models) by comparing real process measurements to simulator outputs. Typically, a process simulator, no matter if static or dynamic, cannot adapt its behavior to reality, but the tracking simulator has this ability due to its integrated update mechanism. A tracking simulator is commonly used for state estimation of non-linear process models (simulation models). The suggested invention utilizes standard PI controllers, which enables fast update of parameters. A major advantage with the PI controllers is that autotuning can be applied and, hence, no tuning parameters need to be plucked out of the air. With the proposed innovation, it will be easy to extend an ordinary process simulator to a tracking simulator, which can be used for many purposes, including predictive and fault tolerant control, soft sensors, prediction of future plant behavior, parameter estimation, and plant optimization.