Simultaneous application of adaptive PID controller and smith dead-time predictor rule in nonlinear water level control in Neka power plant

• Published in: 2011 6th IEEE Conference on Industrial Electronics and Applications pp. 1342 - 1347
• Main Authors: Mehrafrooz, A., Sorkhkolaei, M. R., Yazdizadeh, A.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2011
• Subjects: Adaptation models; Biological neural networks; Control systems; Delay; Fluid level Control; Fluids; Multi-variable Systems; Multiple-Input-Multiple-Output Systems; Neural Networks; Neurons; Nonlinear Control; PID Controllers; Pumps
• ISBN: 9781424487547; 1424487544
• ISSN: 2156-2318
• DOI: 10.1109/ICIEA.2011.5975796

In this paper, a novel adaptive PID-like controller is introduced. Compared to the conventional methods, better responses are achieved by using the proposed method. The proposed controller is based on neural networks technology and is applied to different kind of black box systems, linear or nonlinear systems and time variant and/or time invariant systems. Generally speaking, it is known that classical tuning methods for PID controllers provided unsatisfactory results for industrial plants where the time delay exceeds the dominant lag time; that is the reason for studying alternative strategies. In this context the most popular rule is Smith predictor. Generally, dead time is a variable parameter and leads to changing the model of system during process of control. Therefore we need to estimate this variable parameter using an online algorithm. We have done and proven this estimation via Smith rule for different systems with first order or second order models. In order to show the actual performances of mentioned methods, we have run them on the level control of tanks in water refinement process of Neka power plant which generally is a very nonlinear and delayed system. In order to control this system we apply our neural networks controller and Smith predictor rule simultaneously. Simulation results in this chapter show the perfect performance of our adaptive controllers and rules.