基于聚类分析和状态估计的制粉系统故障预警

摘要

Condition monitoring can greatly reduce the maintenance costs of power generation equipment.In this paper,a new condition monitoring methods based on the nonlinear state estimation for the generator set was presented.A new and improved matrix in memory can achieve a good coverage of the normal operation status of the milling system.When a large difference between the estimated value and the actual measurement value,the existence of the fault was indicated.The cluster center was extracted from the normal condition monitoring parameters,the similarity was calculated,and work condition of the fan was pronounced.The real-time data of the power plant were used to verify the method.The simulation results show that this method can effectively achieve early failure warning.%状态监测可大大降低发电设备的维护成本.将基于非线性状态估计的状态监测方法用于发电机组制粉系统.采用一种改进的内存矩阵的构造方法,能够很好地覆盖制粉系统的正常运行空间.当估计的模型和实际测量值之间相差很大时表明状态异常.通过提取正常工况下的监测参数的聚类中心,计算估计值与聚类中心的相似性测度值,确定风机工作状态.根据国内某机组某直吹式制粉系统的实时数据验证,仿真结果表明此方法可以有效地实现早期故障预警.