Multivariable systems exhibit complex dynamics because of the interactions between manipulated and controlled variables. In this paper, a control scheme for controlling reactor temperature and pressure in CSTR is implemented. The controller design is distributed into two parts: Initially, the judicious pair of loop configuration is determined using Relative Gain Array technique. Then, a decentralized PI controller is designed to minimize the interaction effects using auto tuning algorithm for two different cases. The servo and regulatory problems confirm the effectiveness of the proposed design when one of the manipulated parameters pressure is fixed and both the manipulated parameters are varied. MATLAB and SIMULINK tools are used for system modeling and controller design. The performance indices of the Multiple Input Multiple Output based decentralized PI controller for CSTR system is also evaluated and presented.

Multivariable systems parade complicated dynamics because of the interactions between input and output parameters. In this paper, a method is presented for controlling the multivariable processes in a laboratory reactor system mimicking the real time deep-sea conditions. Relative Gain Array method based decoupling controller for deep-sea parameters such as pressure and temperature was designed in order to minimize the interaction effects between the parameters. Pressure in the ocean increases by about 1 atmosphere for every 10 meters of depth, the amount of pressure experienced by many deep-sea organisms is extreme and temperature varies from -2 o C to 300 o C based on the location, depth, hydrothermal vent etc. For this purpose, a real time data has been generated from the Deep-sea Microbial Culture facility system and system transfer functions are identified. Various manipulated variables interacting and influencing the MIMO system are analyzed and identified. Using MATLAB system identification tools, the suitable numbers of manipulated variables involved in the transfer function model are identified and the best fit was evaluated using various methods. The operation of the MIMO based decoupled PI controller for deep-sea laboratory system was studied for reference tracking and cases of disturbance rejection. The performance indices of the controller are calculated, compared, analyzed and presented. Simulation results endorse the efficiency of the proposed control system.