NC Ultrasonic Machining Efficiency: Neural Network-Based Modeling and Simulation
To explore the impact of abrasive granularity, feed pressure and cutting feed speed on NC ultrasonic machining efficiency, a technological test was carried out, and based on the test results, back propagation (BP) neural network model was established and validated by simulation. The validation process showed that when relative error is less than ±10%, only two samples among 18 tested have larger errors. By the utilization of the BP network for training, correct fitting rate of machining efficiency target can be reached up to 88.9%. Our study indicates that (i) the output of the network is well fitted with the test data, (ii) the established model has good generalization ability to reflect the laws of NC ultrasonic machining process, and (iii) the model is suitable as a prediction tool for NC ultrasonic machining efficiency.
Yungang Li, Pengcheng Wang, Liqun Ai, Xiaoming Sang and Jinglong Bu
G. Y. Zhong and Y. Q. Wang, "NC Ultrasonic Machining Efficiency: Neural Network-Based Modeling and Simulation", Advanced Materials Research, Vols. 291-294, pp. 406-410, 2011