Ppredictioon Of Grrinding Ssurfaceroughnness Andd Experimental Researcch

Titanium alloy are becoming popular materials in many areas for their high performancessuch as high specific strength, good corrosion resistance. But its high performances also makethem difficult to grind. So the grinding mechanism is needed to study, and the prediction modelbetween the machining quality and grinding parameters is needed to build.The thesis selected TC4titanium alloy to study the relationship between the surfaceroughness, the grinding force in Y axis and the grinding wheel speed, feed speed as well asgrinding depth. On the basis of experiment, the regression model and BP neural network modelwere established. The experiment was divided into the single factor experiment and orthogonalcomposite experimental. The effect of the single factors and two composite factors on thegrinding force in Y axis and the surface roughness were studied. The experiment data wasanalyzed in regression way, and the regression model was made and verified. By analyzing themechanism of grinding adhesion wear, the adhesion phenomena in grinding process of titaniumwas explained, different processing parameters on grinding force and surface roughness wereanalyzed, mathematical model of wheel adhesion rate was established. It showed that thegrinding force had a similar pattern with the surface roughness, which was inversely proportionalto the wheel speed, and directly proportional to feeding speed and grinding depth. The effect ofwheel speed was largest, followed by the grinding depth, the last was the feeding speed.The thesis selected BP neural network to make a prediction model of surface roughness andgrinding force in Y axis. Comparing with the regression model, the BP neural network was betterthan regression model in multi parameter prediction, while the regression model was moreprecise than BP neural network.