Prediction of Conveying Velocity of a Flexible Vibratory Feeding System Using Artificial Neural Networks

摘要

In this paper, artificial neural network approach is proposed for the prediction of the conveying velocity of a flexible decoupled vibratory feeding system. A multi-layered BP (Back-Propagation) network is built using NeuralWorks Professional PLUS/II. The inputs to the neural network are the phase difference of the track vibration signals of the two orthogonal axes and the acceleration amplitude of the track vibration, and the output is the conveying velocity. To ensure satisfactory prediction, the learning parameters of the network must be carefully selected. All the training data and test data are obtained from experiments, which are carried out on a sensor-based flexible vibratory feeding system. Satisfactory results are obtained with prediction errors of less than 5%.