Research On Milling Of Alumina Ceramic Plate By Abrasive Water Jet

Alumina ceramics have high toughness,high hardness,high temperature resistance,etc.,which are widely used in industrial production.Using traditional machining alumina ceramics is prone to many disadvantages,such as easy to damage tools,low processing accuracy and large deformation stress in the workpiece.As a technique that is clean and non-polluting and high processing efficiency,the abrasive water jet is naturally very suitable for processing rigging materials.Based on the milling of alumina ceramic plates,the working principle of water jet structural characteristics and abrasive water jet was analazed the oretically.The impact model of the abrasive water absorbing to the workpiece and milling was explored.To perform single factor experiment,five process parameters the jet pressure,nozzle moving speed,standoff,horizontal feed amount,and abrasive flow rate were selected.Their impact on milling process performance was analyzed.Through orthogonal experiments and range analysis,5 parameters affecting the material removal rate and milling depth as well as the primary and optimal combination was found.The effect of process parameters on the surface quality of milling was investigated.The surface roughness can be reduced by secondary milling process with low-jet pressure and fine abrasive.The parameters of the nozzle movement and the lateral feed direction affecting the surface roughness was explored.The Minitab software is used to obtain the main order and optimal combination of process parameters affecting surface roughness.On the basis of the abrasive water jet to remove the crispy material model,after considering the material parameters and jet parameters,a single milling depth mathematical model is established,and the maximum error is 11.6% by experimental verification.The surface roughness model in both directions is established in the direction of nozzle moving and the lateral feed direction.Based on the theory of artificial neural network,and MATLAB programming language,the BP neural network model was trained through 100 sets of sample data.By verification of 9groups of experimental data,the maximum error of the expected milling depth and the actual milling depth is 6.1%.It indicate that the prediction of multi-parameter nonlinear system for abrasive water by artificial neural networks is feasible.