盾构施工典型故障诊断初步研究

Preliminary study of typical fault diagnosis in shield tunneling

随着盾构技术的广泛应用,出现了一些典型事故,有必要在盾构施工过程中引入故障诊断技术,以避免类似事故的再次发生。从本质上讲,盾构施工过程中的故障诊断技术是个模式分类问题,可以借助BP前馈神经网络来实现。结合广州地区的生产实例,在对典型故障简单分类的基础上,对具体应用BP网络实现盾构机的故障诊断进行了分析和探讨。算例表明,应用BP神经网络进行盾构施工过程的故障识别与诊断是可行的。当然,为进一步提高故障诊断的效果,应加强对典型故障数据的积累并提高故障间的可分离度。

With the widespread use of shield tunneling technique,many accidents appear,and it is necessary to introduce fault diagnosis technique in shield tunneling to avoid similar accidents appear again.In essence,the fault diagnosis technique in shield tunneling is the problem of pattern recognition,which can be realized by means of the BP feed forward neural network.Combined with the production examples in Guangzhou region,the practical application of BP feed forward neural network to realize the fault diagnosis is analyzed and discussed.Computation example shows that it is feasible to use BP feed forward neural network to realize the fault diagnosis.Of course,the accumulation of the data of typical faults should be strengthened and increase the degree of separation among faults in order to the increase the effectiveness of the fault diagnosis.