采用调控流图与数据挖掘相结合的方法研究了原油蒸馏过程故障诊断问题。利用调控流图方法进行建模,通过对调控流图模型的定性仿真获得每个故障的传播途径及对每个观测变量的影响,形成故障一监测变量关系表,最后利用数据挖掘技术建立从监测变量推断故障的故障诊断模型。由于该模型结构简单,可以在生产过程中直接应用,具有速度快、判断直接的特点,在一定程度上解决了复杂过程实时故障诊断中的组合爆炸问题。
Fault diagnosis in a crude oil distillation process has been studied using Regflow graphs (RFG) com bined with data mining. First the RFG model of the process was constructed. Then qualitative simulations were completed, which showed the propagation path of each fault and the effects of the faults on each observed variable. The simulation results can be used to establish a fault diagnosis model by incorporating data mining methods. The simple structure of this fault diagnosis model makes the diagnosis fast and direct, which can partly solve the combi natorial explosion problem in the real-time fault diagnosis of complex processes.