Optimal Design for Multiple Variable Steam Temperature System Based on Auto Disturbance Rejection Control

The multi-variable reheat steam temperature process with steam-steam heat exchanger in thermal power generation units is a long time-delay, coupled multi-variable system with variable parameters. Thus, the related gain matrix method is applied in the decoupling analysis and the auto disturbance rejection control (ADRC) is used in the design of the controller. However, there are so many parameters that should be tuned as for auto disturbance rejection control. Moreover, without fixed method, tuning process and effects depend on experience. Therefore, it is necessary to optimize these parameters. In optimization theory, the quantum particle swarm optimization (QPSO) adopts coding mechanism of probability amplitudes to expand the traversal capability of solution space, but the ability of searching global optimal value depends greatly on the choice of initial parameters. In addition, it is easy to get into the local optimum and to be premature convergence. In order to solve problems above, firstly, chaotic sequences are used to initialize the origin angle position of particles; secondly, mutation algorithm is introduced which can effectively increase diversity of population, and also can avoid premature convergence. The test results of function optimization show that the proposed algorithm has better optimized effects. This new kind of quantum particle swarm optimization proposed in this paper is used to optimize parameters of auto disturbance rejection control and applied to a multi-variable coupled process with steam-steam heat exchanger in a domestic coal-fired power generation unit. Computer simulation results show that the control strategy is effective. The control algorithm has merits of simple structure, strong anti-interfere ability, and is easy to realize in Distributed Control System (DCS), so it has hopeful application prospect.