Control of heat exchangers in series using neural network predictive controllers
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology, Institute of Information Engineering, Automation and Mathematics, Radlinského 9, 812 37 Bratislava, Slovak Republic
E-mail: * email@example.com
Abstract: The paper reveals three applications of neural network predictive control (NNPC) to a system of four heat exchangers (HEs) in series with counterflow configuration to save energy expressed by cooling water in the system of HEs cooling the distillation product. Neural networks (NNs) are used at first in conventional NNPC and subsequently, neural network predictive controllers (NNPCLs) are employed as a master controller in a cascade control, and as a feedback controller in the control system with disturbance measurement. Neural-network-predictive-control-based (NNPC-based) feedback control systems are compared with PI controller based feedback control loop. Series of simulation experiments were done and the results showed that using NNPC-based cascade control reduced cooling water consumption. This control system also significantly reduced the settling time and overshoots in the control responses and provided the best assessed integral quality criteria compared to other control systems. NNPC-based cascade control can also be interesting for industrial use. Generally, simulation results proved that NNPC-based control systems are promising means for the improvement of HEs control and achievement of energy saving.
Keywords: heat exchanger; neural network predictive control; neural-network-predictive-control-based cascade control; neural-network-predictive-control-based control system with disturbance measurement
Full paper in Portable Document Format: acs_0356.pdf
Acta Chimica Slovaca, Vol. 13, No. 1, 2020, pp. 41—48, DOI: 10.2478/acs-2020-0007