PID autotuning via two-stage safe Bayesian optimization for heat pump control: Simulation and experimental results

16:31-16:43, January 27 @ 2BC

Talk/ Overview

Autotuning industrial controllers offer advantages over manually tuned controllers by providing a high-level control performance over the lifetime of the system and reducing the need for maintenance and re-commissioning. These controllers are particularly useful for systems whose parameters vary over time, for example, due to aging or changing of some components of the system. In this work, we describe an algorithm for autotuning of the most common industrial controller, the proportional-integral-derivative (PID) controller, and we demonstrate its application to a heat pump system that is part of a heating system of a vertically integrated neighborhood NEST at EMPA Duebendorf in Switzerland. 

Talk/ Speakers

Bratislav Svetozarevic

Postdoctoral Researcher, Empa

AMLD / Global partners