Artificial intelligence based non linearity prediction and design of compensator for nonlinear process control reactor

Abstract

Continuous Stirred Tank Reactor (CSTR) is primary unit operations in chemical industries. Chemical reactions are either exothermic or endothermic which require energy either to be removed or added to the reactor to maintain constant temperature. Exothermic reactions create safety measure problems and exotic behavior possibility in multiple steady states. Process modeling is essential in the model based process control systems and to capture the effects of nonlinearities. To ensure the successful operation, it is necessary to understand their dynamic characteristics which will ultimately enable effective modelling and design. To describe the dynamic behavior of a CSTR, the knowledge of the functional expressions about the chemical reactions are required for developing the mass, component and energy balance equations. Complicated mathematical models need many parameters like priori information, subjective information, complexity, training, model evaluation, fitness to empirical data, scope of the model and other technical considerations. The determination of reaction rate coefficients in chemical reactions and heat transfer coefficients in industrial processes are diagnostic situation. In such a process, the determination of specific parameter values might be the final goal of the identification. Identification is the process of determination based on the input and output of a system within a specified class of systems. The selection of the class of input signals and the criterion are largely influenced by the a priori knowledge of the process as well as by the purpose of identification. The neural network predictive controller is applied in the proposed work, which is very efficient to identify complex nonlinear systems with partial model information. newline newline newline