基于贝叶斯网络的制粉系统安全控制方法

Safety Operation Control for the Coal Pulverizing System Based on Bayesian Network

制粉系统作为燃煤火电机组重要的辅助系统,其状态直接影响整个机组的安全稳定运行。针对制粉系统中最易出现的堵煤异常工况,提出基于贝叶斯网络的安全控制方法。首先,利用专家知识及数据信息离线建立诊断贝叶斯网络模型及安全控制贝叶斯网络模型;其次,在线获取异常工况现象层的节点等级作为证据,通过诊断模型推理,在线诊断异常工况;再次,诊断产生异常工况后,在线获取当前异常工况现象层的实时节点等级进行证据更新,通过安全控制模型,推理产生实时的安全控制决策。最后,利用仿真验证了所提方法的有效性。

Coal pulverizing system is an essential auxiliary system for coal-fired power plant,and it affects the safe and stable operation of the entire unit.Aiming at the coal blockage of abnormal condition,a safety operation control method based on Bayesian network is established.Firstly,the diagnostic model and safety operation control models based on Bayesian network are established offline which using expert knowledge and data information.Then,the node level of the abnormal working condition phenomenon layer is obtained as evidence,and the abnormal conditions are identified online based on the diagnostic model.After then,the evidence is update by real-time node level,real-time security control decisions are made by the safety operation control model.The simulation results shows the proposed method has highly accuracy and validity.