Study On Cutting Tool Wear Condition Monitoring Based On Particle Swarm Optimized Neural Network

On present days, technology has been improved and the degree of automation is getting higher and higher, machinery manufacturing industry as an important symbol of the level of industrialization of country, has been developing. Metal cutting is the most widely method used in manufacturing industry, and the tool is the most basic and essential factor of production in processing. The key is to ensure the tool work normally, improve it’s performance and monitoring it’s condition.In this paper, there are two sensors, the force measuring instrument of Kistler9257B. vibration sensor of8702B50M1. used to sense the signals. In signal processing and condition recognition, mainly used the MATLAB platform, with the cutting force and vibration signals of the tool in different condition and the parameters of time, frequency and time-frequency as eigenvector condition recognition, so that we can get the tool wear characteristics. Through the study of particle swarm optimization, it has some advantages such as simple concept, easy realization, quick convergence rate and few parameters need to be tuned. So, it can be applied to optimize BP neural network in recognition. Relative to the traditional BP neural network, the BP neural network of particle swarm optimization is better.Finally, the Tool Wear Condition Monitoring System is based on the modern signal processing method and the BP neural network of particle swarm optimization. Following experiment shows that by taking full advantage of the processing of cutting force and vibration of the effective information, and used the BP neural network of particle swarm optimization in condition recognition, identification accuracy and speed can be greatly improved. The developed method is proved to be practical and significant for the cutting tools monitoring in the producing process.