Research of Reheated Steam Temperature Multi-Model Predictive Control for 1000MW Ultra-Supercritical Unit

In this paper, targeted at reheated steam temperature's large time delay and strong nonlinearity characteristics in power plant, a multi-model predictive control strategy is proposed in light of the nonlinear characteristics of steam temperature control. This strategy utilizes the linearized model of nonlinear process at multiple working points to divide the process into several subspaces. A relatively accurate fixed model can be found in each subspace. Based on this, a global approximation model of complex object could be obtained. Simulation and application results show no matter for stable or load varying process, reheated steam temperature can be maintained in the required control range. It is much superior to the traditional PID control method.