摘要
汽轮机转子在运行过程中由于各种原因会出现振动异常情况,振动烈度随转子转速和负荷升高而呈线性增大,直接影响机组正常运行,处理不当将造成机组动静碰磨,甚至造成转子大轴弯曲。传统的诊断方法试验项目多,诊断时间长,为快速诊断故障类型,查找故障原因,以防故障恶化,用BP神经网络建立汽轮机转子故障诊断模型,对汽轮机转子运行故障进行诊断,诊断结果精确,解决了传统诊断方法费时费力的问题。利用BP神经网络方法,对国产300MW机组转子运行故障进行分析诊断,诊断结果与现场实际揭缸查验一致,论证了利用BP神经网络方法对汽轮机转子运行故障进行预诊断是可行的和准确性,同时诊断结果为机组检修提供了更为准确的依据,并为转子设计和加工制造提供了参考。
The turbine rotor of thermal power appear abnormal vibration during operation due to various reasons. Vibration intensity increased linearly with speed and load increasing, affecting the normal operation of the unit directly. If improper handing the unit will take place rubbing, or even result in a large rotor shaft bending. Tests for traditional diagnostic methods are a little more, diagnostic for a long time. In order to diagnose the fault type quicklyl find the cause and prevent failure deterioration. We make the fault diagnosis model of turbine rotor using BP neural network, diagnose the turbine rotor fault of power turbine. The diagnostic results are accurate. The method solved the problem of time consuming for traditional method. We analyze and diagnosis the rotor operation fault of the domestic 300MW unit use of BP network method in this paper. The diagnostic results is same with the actual mortgage, demonstration of using BP neural network method for thermal power turbine rotor running pre-fault diagnosis is feasible and accuracy. Meanwhile the diagnostic results of the unit overhaul provide a more accurate basis, and provide a reference for the rotor design and manufacturing.
出处
《机械设计与制造》
北大核心
2014年第5期233-236,共4页
Machinery Design & Manufacture
基金
辽宁省教育厅资助项目(L2010382)