Energy efficient convex-lifting-based robust control of a heat exchanger

Abstract

This paper analyses control performance and energy savings reached by application of four robust control approaches implemented on the laboratory plate heat exchanger. The output temperature of the cold medium is the controlled variable and the flow rate of the hot medium is the manipulated variable. Closed-loop control of the heat exchanger aims to ensure offset-free setpoint temperature tracking. First, the computationally demanding robust model predictive control (RMPC) strategy is investigated. Improvements ensured by implementing three modifications of the convex-lifting-based robust control methods are analysed. These modifications are based on the (i) non-tunable, (ii) tunable, and (iii) multiple tunable robust positive invariant (RPI) sets. Designed robust control approaches significantly improved control performance. For considered control setup, the settling time was reduced up to 70%. Energy savings were increased by 82%, when compared to RMPC. Moreover, the considered methods are proper for real-time implementation as they significantly reduce computational demands. The designed robust convex-lifting-based methods are suitable for industrial hardware with limited computational requirements.