基于改进粒子群算法的SCR烟气脱硝系统模型辨识

Model Identification of SCR Flue Gas Denitration System Based on Improved Particle Swarm Optimization

针对火电厂燃煤机组中选择性催化还原(SCR)烟气脱硝系统存在非线性、大惯性和大延迟等特点,传统辨识方法存在收敛速度较慢和辨识精度较低的问题。将自适应动态惯性权重、粒子的自适应变异以及自然选择的思想引入到基本粒子群算法中,并将改进后的粒子群算法用于某660 MW超超临界燃煤火电机组SCR烟气脱硝系统的模型辨识。通过机组现场运行数据建立了调节阀位置开度与NH_(3)质量流量、NH_(3)质量流量与出口NO_(x)浓度、入口NO_(x)浓度与出口NO_(x)浓度之间的传递函数模型。研究结果表明:改进后的粒子群算法提高了模型辨识精度,有效改善了基本粒子群算法易陷入局部最优和后期种群多样性丧失的缺陷。模型验证的仿真结果表明了该方法的有效性。

Due to the nonlinearity,large inertia and large delay of selective catalytic reduction(SCR)flue gas denitration system in coal-fired power plants,the traditional identification methods have the problems of slow convergence speed and low identification accuracy.In this paper,the ideas of adaptive dynamic inertia weight,adaptive variation of particles and natural selection are introduced on the basis of fundamental particle swarm algorithm.The improved particle swarm algorithm is applied to model identification of SCR flue gas denitration system of an ultra-supercritical 660 MW unit.Based on the field operation data of the unit,the transfer function models between the position opening of the regulating valve and the mass flow of ammonia,between the mass flow of ammonia and the concentration of NO_(x)in the outlet,and between the concentration of NO_(x)in the inlet and that in the outlet were established.The results of the study show that the particle swarm algorithm optimization improves the model identification accuracy,and effectively improves the defects of basic particle swarm algorithm which is easy to fall into the local optimum and the loss of population diversity in the later stage.Meanwhile,the simulation results demonstrate the effectiveness of the proposed method.