Title: Neural network parametric modelling of abrasive waterjet cutting quality

Authors: Lin Yang, Jun Song, Biaohui Hu

Addresses: School of Mechanical and Electrical Engineering, Chongqing Jiaotong University, Xuefu Road 66, Chongqing 400074, PR China. ' School of Computer and Information Engineering, Chongqing Jiaotong University, Xuefu Road 66, Chongqing 400074, PR China. ' School of Mechanical and Electrical Engineering, Chongqing Jiaotong University, Xuefu Road 66, Chongqing 400074, PR China

Abstract: Abrasive Waterjet (AWJ) can provide very effective means for shape cutting of difficult-to-machine materials. One of the principal deficiencies of AWJ cutting process is the wavy striation on the generated cut surface in relatively thick workpieces or high traverse speed. It is therefore essential to predict the main AWJ processing parameters to achieve a desired cutting quality. But many aspects about this technology are neither fully understood nor have they been accurately modelled. In this paper, based on the experimental data, an Artificial Neural Network (ANN) parametric model of AWJ cutting process was developed and used in CNC machines. The predicted results indicated that the model could identify the cutting quality to a high desirable accuracy, and the ANN can be used as an appropriate method for prediction of cutting parameters in AWJ systems.

Keywords: abrasive waterjets; AWJ; waterjet cutting; surface quality; parametric modelling; artificial neural networks; ANN; process parameters; CNC machining.

DOI: 10.1504/IJAT.2007.015384

International Journal of Abrasive Technology, 2007 Vol.1 No.2, pp.198 - 207

Published online: 12 Oct 2007 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article