为提高控制系统在线辨识能力,实现NOx、SO2 的超净排放,研发了脱硫脱硝除尘一体化设备的数字化控制技术.首先将脱硫率、脱硝率设定值作为设备控制系统的输入,采用减聚类算法优化RBF神经网络,构建动态RBF辨识器,使控制输出值无限接近脱硫率、脱硝率设定值;然后将获得的Jacobian信息输入人工神经网络PID控制器,整定控制器参数,实现脱硫率、脱硝率精准控制.实验结果表明:改进后的RBF辨识器具有较强的抗噪声干扰能力,可实现脱硫脱硝除尘一体化设备的数字化控制,控制时间更短、超调量更小,平方积分误差(ISE)、绝对积分误差(IAE)、控制增量累计平方和(CSCI)等指标值均更低,且NOx、SO2 排放浓度低于设定值.
In order to improve the online identification ability of the control system and realize the ultra - net emission of NOx and SO2, the digital control technology of desulfurization, denitrification and dust removal integrated equipment is studied. The set values of desulfurization rate and denitrification rate are taken as input of the device control system, and the RBF neural network is optimized by using the reduced clustering algorithm, and a dynamic RBF identifier is constructed to make the control output value infinitely close to the set values of desulfurization rate and denitrification rate. Jacobian information obtained is input to the artificial neural network PID controller and the controller parameters are adjusted to set the achieve precise control of desulfurization rate and denitrification rate. The experimental results show that the improved RBF identification instrument has strong anti-noise interference ability, and can realize the digital control of desulfurization, denitrification and dust removal integrated equipment. The control time is shorter, the overshoot is smaller, the ISE, IAE and CSCI index values are lower, and the NOx and SO2 emission concentrations are lower than the set values.