基于全连接神经网络的旋挖钻机缓冲平衡阀故障诊断

回转装置是旋挖钻机的主要结构之一,而缓冲平衡阀是回转装置的重要部件。针对缓冲平衡阀故障诊断中故障类型复杂多变、早期微弱故障特征难以提取等问题,提出了一种基于全连接神经网络的故障诊断方法。以VAA-B-SICN-PDRM-VF-150平衡阀为研究对象,针对平衡阀故障数据难以获取的问题,进行加速退化实验,获取故障数据;然后,根据阀的特性建立故障评价指标,结合全连接神经网络方法,建立了平衡阀故障诊断模型,并进行模型训练,验证故障诊断模型的准确性。结果表明,基于全连接神经网络的故障诊断模型能够快速、准确地检测出缓冲器平衡阀的故障,从而避免发生事故导致停工停产。

Fault Diagnosis of Cushion Balance Valve of Rotary Drilling Rig Based on Fully Connected Neural Network

The rotary unit is one of the main structures of rotary drilling rigs, and the buffer balance valve is an important part of the rotary unit. For the problems of complex and variable fault types and difficult extraction of early weak fault features in buffer balance valve fault diagnosis, this paper proposes a fault diagnosis method based on fully connected neural network. This paper takes VAA-B-SICN-PDRM-VF-150 balancing valve as the research object, and conducts accelerated degradation experiments to obtain fault data for the problem of difficult to obtain fault data of balancing valve. Then, the fault evaluation indexes are established according to the characteristics of the valve, combined with the fully connected neural network method, the balancing valve fault diagnosis model is established, and the model training is carried out to verify the accuracy of the fault diagnosis model. The results show that the fault diagnosis model based on the fully connected neural network can quickly and accurately detect the fault of the buffer balancing valve, thus avoiding the accidental shutdown and production stoppage.