Practical Research And Application Of GPC In Large-Lag Temperature Control System

Temperature control is widely used in many industrial areas, such as the plastic extruder's stuff container, and the process of thermodynamic engineering in power plant. The temperature control effect of extruder's stuff container influences the capability and quality of extrusion products directly. Thus, the dependability of temperature control over stuff container has become a key point of product molding. Meanwhile, temperature control has also become an important part of extruder product manufacture. The super-heated steam temperature is the maximum temperature in the whole steam channels in the process of thermodynamic engineering in power plant. It will bring on dangerous factors if the steam temperature is too high or too low. So, the super-heated steam temperature must be controlled to a required range.Large-lag, large-inertia and nonlinearity are the main features of extruder and power plant, and the object controlled models are uncertain and are easily disturbed by environmental factors, meanwhile, these are also characteristics of most of the temperature controlled objects. So it is difficult to achieve an ideal control effect due to those features. Now PID controller is the most widely used in temperature control. It is very simple and has good stability as well as dependability, and can get a good effect for general temperature control. But when the parameters of the model changed in the production process, the control effect became worse and the system would be overshoot and oscillations because the PID controller can not self-tune its parameters.So the generalized predictive control (GPC) algorithm is utilized in this paper for the temperature control. GPC is suitable for systems with large-lag and non-minimum phase systems those are open-loop unstable. It can realize adaptive control through system identification on line. Due to the large computational work of the GPC general algorithm, we take two methods to improve the real-time performance: applying GPC implicit algorithm and inversing a matrix by recursion. Meanwhile, combining GPC implicit algorithm and PID together eliminates the bad influences that inaccurate initial controller parameters bring. There are theoretical research and simulations in this paper. Besides, experiments for temperature control are also done by B&R PCC controller. Simulations and experimental results have both shown that the implicit GPC-PID controller can deal with the problems of large-lag systems, and have excellent robust and self-adaptation performance.