一种基于运行数据的SCR串级脱硝控制系统参数优化方法

A Parameter Optimization Method for SCR Cascade Denitration Control System Based on Operating Data

SCR脱硝系统具有严重的非线性、大惯性、大延迟和时变性等特点,在系统实际运行时,常规PID控制器的控制效果并不理想。提出一种基于数据驱动控制且带有数据库更新的控制器参数优化方法。该方法在利用改进KNN算法在历史数据库中查询与当前状态较为相近的K个数据组,并对其进行加权平均得到适合当前状态下的控制器参数。利用Hurst指数判断控制性能,通过扩展虚拟参考整定(EFRIT)算法对得到的控制参数进行优化并在线完成数据库的更新,保持数据库的时效性。以SCR脱硝系统的典型工况作为测试对象,实践结果表明,相比于常规PID控制器,当系统负荷增加时,该控制策略下的调节速度增快;系统负荷减小时,该控制策略下的超调量减小,具有更好的控制效果。

The SCR denitration system has the characteristics of strong nonlinearity,large inertia,large delay and time-varying.When the system is running,the control effect of the conventional PID controller is not ideal.A controller parameter optimization method based on datadriven control with database update was proposed.The method used the improved KNN algorithm to query K data groups that were close to the current state in the historical database,and weighted them to obtain the controller parameters suitable for the current state.The Hurst index was used to judge the control performance,and the obtained control parameters were optimized by the extended virtual reference tuning(EFRIT)algorithm,and the database was updated online to maintain the timeliness of the database.The typical working conditions of SCR denitration system was taken as the test object.The results show that,compared with the conventional PID controller,when the system load increases,the adjustment speed under this control strategy increases;when the system load decreases,the overshoot under this control strategy is reduced and has better control effect.