Study On Cutting Model For Abrasive Water Jet Based On Artificial Neural Network

Because of its unique "cold, soft" processing characteristics, abrasive water jet technology is being used increasingly in manufacturing area, and becoming one of the fastest-growing special processing technologies in the past 30 years. The cutting model for abrasive water jet is one of the topics on the study of abrasive water jet technology. It has a great significance on achieveing the precise machining with abrasive water jet.The research topic on the establishment of abrasive water jet cutting model using the neural network technology is put forward on the basis of overview of the current development of abrasive water jet cutting model home and abroad.The main characteristics,the cutting system, the cutting principles and the factors which impact the cutting performance of abrasive water jet were studied in detail. The analysis shows that the relationship between the cutting quality and cutting parameters of abrasive water jet is highly nonlinear, so it is difficult to establish a complete cutting model using traditional mathematical methods.Based on the theoretical research of abrasive water jet cutting and neural network modeling techniques, a cutting model for abrasive water jet based on artificial neural network was established in the Matlab environment by design of neural network topology and network training algorithm. With the HJ300 precision abrasive waterjet cutting machine, the plexiglass cutting experiments were carried out using the orthogonal experiment method. The 25 set of representative datas which can reflect the laws of abrasive water jet cutting were obtained to train and test performance of the network model. The results showed that the network model has good fitting ability and generalization ability. It can be used for the speed forecast of abrasive water jet cutting.Combined with the neural network technology, Matlab computing technology and Delphi software developing technology, an abrasive water jet cutting speed prediction unit which can integrate with the existing abrasive water jet cutting machine tool's CNC System was developed due to the achievement of the communication between the Matlab and the Delphi language using the Matlab interface technology. Predicted results showed that the prediction unit can achieve the fast and accurate prediction of the cutting speed in non-Matlab environment. It lays the foundation for the precise machining with abrasive water jet and the intelligent control for abrasive water jet cutting.The verification experiment using abrasive water jet showed that the cutting model of artificial neural network for abrasive water jet can meet the needs of the improvement of the abrasive water jet cutting quality. It has a high accuracy and practicabile uses. The cutting model for abrasive water jet based on artificial neural network has the advantages of predicting rapidly and accurately,operating easily and intuitively,and broad scope of application.It is a reliable model for abrasive water jet cutting. It has a highly practical value and good promising in application.