| High-speed machining technology has been rapidly developed since 1990 s. As a typical fundamental and representative position of the advanced manufacturing technology, it is gradually widely used in aerospace, aviation, automotive, mold and other fields. In general, molds and parts that processed by high speed milling are materials that are more difficult to process and have large amount of material removal, low working efficiency and shorten tool life and has negative effects on surface quality and fatigue life of mold.Surface integrity is one of the important indicators to measure surface properties, and it has a very important influence on the fatigue life of parts. Therefore, the research on the influence and interaction mechanism between and the surface integrity and fatigue life in high speed milling has an important theory value and practical significance. In this paper, high-speed cutting of mold is investigated from two aspects of theory and experiment and put forward the mathematical models, aiming at providing guidance for high-speed machining of die steel.Research mainly from the following four aspects: firstly, toroidal cutter is applied in the high-speed milling experiments of SKD61 die steel. The influence rule of spindle speed, feed rate, axis depth, radial width and tool angle on the surface roughness is investigated through the orthogonal test method. The mathematical prediction model of roughness is put combined regression analysis. Secondly, the rules and mechanism of technical parameters on surface integrity is investigated. The results shows that low surface roughness, low residual stress and small machining surface hardness of SKD61 die steel processed by high speed milling could be obtained. Thirdly, the fatigue experiment was carried out on the test prototype, the result shows that surface residual compressive stress can extend the fatigue life of parts significantly. In a certain extent, the fatigue life of parts will vary rapidly with the change of surface roughness and surface hardness. Finally, the prediction model of fatigue life was built according to the surface integrity The fatigue life can be predicted by the prediction model according to the BP neural network, its errors of predicted and tested data range from 2.3% to 15.8%. |
