Affiliations: [a] School of Mechanical Engineering, Hebei University of Technology, Tianjin, China
| [b] Career Leader Intelligent Control Automation Company, Suqian, Jiangsu Province, China
Correspondence:
[*]
Corresponding author. Lingyu Sun, School of Mechanical Engineering, Hebei University of Technology, Tianjin 300130, China. E-mail: [email protected].
Abstract: The abnormal vibration of the loom spindle will seriously affect the quality of the textile. Based on the inherent embedded control system of the rapier loom, this paper develops an embedded system that monitors and analyzes the vibration signal of the spindle to determine the cause of the spindle failure. The system improves the traditional vibration sensor signal acquisition method, design the sensor peripheral auxiliary circuit and vibration signal conditioning circuit, and design the data storage and communication module so that the system has the characteristics of low cost, strong flexibility and scalability. The embedded algorithm program of Fast Fourier transform is developed, optimized, and is applied to embedded platform, therefore the system can analyze the characteristics of vibration signal in frequency domain. Finally, back propagation neural network (BPNN) is introduced to investigate and match the relationship between the vibration spectrum characteristics and fault types of the loom spindle. The extracted back propagation (BP) learning result is a mathematical mapping formula, which enables the embedded system to analyze and determine the cause of vibration fault by using this formula. System design is conducive to improving the level of production intelligence and reducing personnel costs in the production process.
Keywords: Loom, vibration detection, embedded system, back propagation neural network, Fast Fourier transform
