Research On Fractional Order PI~?D~? Control And Its Applications

Fractional order PI~?D~? controller is a generalized form of the integer order PID controller. Fractional order PI~?D~? controller is better than integer order PID controller because of the two more parameters, with a greater range of regulation. For large delay, large inertial thermal process, fractional order controller has better control performance and strong robustness. With the wide use of the fractional order controller, the design of the controller and tuning of the parameter become more and more difficult. To solve this problem, referencing tuning method of integer order PID controller parameters, an engineering setting method for the parameters of the fractional order PI~? controller and an engineering setting method for the parameters of the fractional order PI~?D~? controller with adjustable differential are proposed in this paper. The main contents include:1. Based on an improved genetic algorithm, an optimal tuning method for fractional order PI~?D~? controller parameters is proposed. In view of the problems existing in the application of the basic genetic algorithm, the improved performance index to optimize the parameters of the controller is carried out. Finally, simulation experiments are provided to study the control performance of the fractional order PI~?D~? controller;2. An engineering setting method for the parameters of the fractional order PI~? controller is proposed. Engineering setting neural network model is established. Input variables are the lag time of the process and the inertial time. Output variables are controller parameters of fractional order PI~? controller. The model can change inputs in order to get the parameters of the fractional order PI~? controller;3. An engineering setting method for the parameters of the fractional order PI~?D~? controller with adjustable differential is proposed. The differential input share of the fractional order PI~?D~? controller is used as an input variable in the engineering setting model. Engineering setting neural network model is established. Input variables are the lag time of the process, the inertial time and the differential share of the controller. Output variables are controller parameters of fractional order PI~?D~? controller. The model can change the differential input share in order to adjust the size of the differential effect of fractional Order PID controller;4. The two kinds of means of engineering setting for parameters in fractional order controller are applied to the control superheated steam temperature and reheating steam temperature. The simulation researches on the practicability of the two kinds of methods of engineering setting for parameters.