Title :
Gradient optimization of control law for counterflow heat exchanger
Author_Institution :
Silesian Univ. of Technol., Gliwice, Poland
Abstract :
This paper deals with a problem of finding of suboptimal control law for a counterflow heat exchanger. The counterflow heat exchanger is a non-linear, multidimensional, distributed parameter system, which frequently appears in industrial and civil installations. The most common task is to control of output temperature of one medium. In previous works of the author [7, 8] the problem of identification and finding suboptimal control signal in open-loop system has been solved using so called Generalized Back Propagation Through Time (GBPTT) method [4]. In this paper in order to decrease influence of disturbances the control law will be introduced. State variables of the model, which serves as a state observer are arguments of the control law. Parameters of the control law are optimized using GBPTT method. Numerical examples are presented.
Keywords :
backpropagation; distributed parameter systems; gradient methods; heat exchangers; identification; multidimensional systems; neurocontrollers; nonlinear control systems; optimal control; optimisation; temperature control; GBPTT method; counterflow heat exchanger; generalized backpropagation through time method; gradient optimization; nonlinear multidimensional distributed parameter system; open-loop system; output temperature control; state variables; suboptimal control signal law; Distributed parameter systems; Heating; Numerical models; Observers; Optimization; Control and Optimization; Industrial Processes; Neural Networks; Process Control; Trajectory Tracking in Non-linear Systems;
Conference_Titel :
Control Conference (ECC), 2001 European
Conference_Location :
Porto
Print_ISBN :
978-3-9524173-6-2
