The Research Of Predictive Control In Main Steam Temperature Control System

The main steam temperature system of power plant has the characteristics of large inertia, large delay, time variation and disturbance factors. With the increase of the unit capacity, the main steam temperature system will be more complex. The the main steam temperature has a direct impact on the safety and economic operation of the whole power plant, so the control requirement of the system is very strict. At present, the PID controller is simple and easy to operate, widely used in power plant. For the complex main steam temperature system, the traditional PID control, which is derived from the feedback mechanism, is often caused by the inability to adjust in time. And in the changeable environment, the fixed controller parameter control effect is not ideal.The research content of this paper is to study the predictive control method in the main steam temperature system. According to the characteristics of the main steam temperature system, Generalized Predictive Control algorithm is used as the research direction. In this paper, we first study the recursive particle swarm optimization algorithm, the particle swarm optimization algorithm and the recursive algorithm, which is composed of recursive particle swarm optimization algorithm. Secondly, the model is selected as the controlled object, based on the operating environment of the actual valve limit, in the case of hard constraints, the use of recursive particle swarm parameter estimation, selection based on the diophantine equation of the Generalized Predictive Control algorithm to control, get a good control effect. However, the generalized predictive control needs to solve the complex diophantine equations and the simulation control of the hard constraints, not only the amount of calculation, but also with the industry will be very different. In this case, the further reference to the generalized predictive control of a constrained input, the algorithm selection of CARMA models are commonly used as the basis of the research, combined with constrained input softness parameter, soften the valve opening changes. The derivation is simple, the calculation is small, at the same time, avoid the input of the surge in the situation, more close to the actual industrial. The results show that the generalized predictive control based on constraint input has good robustness and control ability, and is of great practical value.