Application of multivariable 2-DOF PID controller with neural network tuning method to the heat exchange

A heat exchange system such as the boiler of a power plant, gas turbine, and radiator require a high rate heat efficiency. But the efficiency of these systems depend on the control methods. In order to properly apply control equipment to boilers or any other heat process, it is necessary to understand the basic aspects of the process that relate control, interrelationships of the process characteristics, and the dynamics that are involved. But it is difficult to understand these complex dynamics and the tuning method of controller. Generally, PID controllers are used in these systems but they cannot be controlled because of the coupling action and disturbance in the system loop. We study an application of the multivariable 2-DOF PID controller to the fuel flow system through simulation and experiments and a backpropagation leaning algorithm of the neural network is used as it´s tuning methods. The experimental results represent good responses to a change of the setpoint and have a robustness against disturbances.