印花机液压系统的故障诊断与检测技术研究

Research on Supervision and Fault Diagnosis Technology for Hydraulic System of the Rotary Screen Printing Machine

在现场实测圆网印花机液压系统的故障数据中,液压系统的磁场信号和振动信号具有短时非平稳的特点。在功率谱分析的基础上,采用小波神经网络技术,对各分布式液控阀产生的振动信号、磁场信号进行小波包能量谱分解,从而获得各类故障在不同频带的能量分布特征向量;应用WNN神经网络建立从特征向量到故障模式之间的数学映射模型,并结合Zigbee网络通讯技术对分散数据进行二次融合辨识。实验结果表明:该系统能够完成对圆网印花机液压系统的无线实时监测,故障定位准确,且体积小巧、通讯可靠性高,能够满足企业及时抢修的需要。

In the realtime tested fault data of the hydraulic system of the rotary screen printing system, the characteristics of the magnetic field and vibrating signals of hydraulic system are short time and unstable. On the base of analysis of power spectrum, the small wave neural network technology was adopted to decompose the power spectrum of small wave package of magnetic field and vibra- ting signals generated by the distributed pilot safety valves, so as to get the power distribution characteristic vector in different frequen- cy band of all kinds faults. The WNN neural network was used to establish the mathematic model from the characteristic vector to fault model and perform the second fusion identification for the dispersed data combining the Zigbee network communication technology. The tested results show that this system can realize the on-line monitoring for the hydraulic system of rotary screen printing system. The system has the characteristics, such as precise fault positioning, small volume, high communicating reliability and can satisfy the requirement of fast repairing for enterprise.