摘要
针对直接空冷凝汽器的结构特点,归纳了运行中可能发生的典型故障,包括真空系统不严密、凝汽器积灰、凝汽器结冰等11个故障。确定了能够正确反映上述故障征兆的过程参数,进一步改进了直接空冷凝汽器故障征兆集。在此基础上利用遗传神经网络对直接空冷凝汽器进行故障诊断。该算法利用遗传算法高效、并行、全局搜索特点,解决了神经网络收敛速度慢,容易陷入极小点的问题。最后,将该方法用于某直接空冷凝汽器故障诊断中,结果表明该算法诊断迅速且诊断结果准确。
With the large-scale application of direct air-cooled condenser, the establishment of fault diagnosis system for it has an important significance. Based on the structural characteristics of direct air-cooled condenser, the typical faults that may occur in the operation of condenser were summarized, including dust accumulation of condenser, freez- ing of condenser and so on. Some process parameters that can accurately reflect the signs of the failure are determined, and the fault knowledge library was further improved. On this basis, the Genetic-BP neural network was used for fault diagnosis of the air-cooled condenser. The algorithm can solve the problems of slow convergence rate, easy plunging in- to local minimum for neural network by the high efficiency, parallelism, global searching of genetic algorithm. Finally, the presented method was applied to diagnose the faults in a direct air-cooled condenser. The result shows that it can rapidly and accurately diagnose the fault.
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2013年第3期69-73,共5页
Journal of North China Electric Power University：Natural Science Edition
关键词
故障诊断
征兆提取
BP神经网络
遗传算法
直接空冷凝汽器
fault diagnosis
symptoms abstraction
BP neural network
genetic algorithm
direct air-cooled condenser
