【作者】 张子义；

【导师】 邹德宁；

【作者基本信息】 西安建筑科技大学 ， 冶金工程， 2005， 硕士

【摘要】 优化设计了高钨、钒高碳高速钢轧辊材料;针对离心铸造高速钢轧辊制造问题,对高速钢轧辊的变质处理和热处理技术进行了深入研究;采用模式识别的偏最小二乘法(PLS)与反传人工神经网络算法(BPN)相结合,在MATLAB平台上开发了复合轧辊热处理保护工艺网络模型;制造出了性能优良的高速钢轧辊。 高速钢轧辊经钾—稀土复合变质处理后,不仅碳化物和基体组织明显细化,还出现共晶碳化物断网现象,产生了较多的小块状和颗粒状碳化物。复合变质高速钢轧辊加热到1050℃时,碳化物网状组织全部消失,大部分碳化物变成了球状组织。变质处理可以明显改善高速钢轧辊力学性能和耐磨性,提高导热性,降低热膨胀系数。 高速钢轧辊热处理工艺不同于普通高速工具钢,经1025～1075℃喷雾淬火和500～550℃回火,具有良好的力学性能和耐磨性,高速钢轧辊经两次回火可以满足要求,三次回火后,硬度反而降低。 采用控制轧辊内层冷却技术,在热处理条件下,可获得内外层组织和性能呈梯度分布的高速钢复合轧辊。采用PLS—BPN算法,对保温板厚度、保温板导热系数、保温涂料厚度、保温涂料导热系数与轧辊内层硬度样本集的学习,建立了复合轧辊热处理保护工艺网络模型;用于预测离心铸造高速钢复合轧辊热处理时,保温材料参数对轧辊内层硬度的影响;模型具有较高的精度和很好的泛化能力,而且运算速度快,可方便的用于指导实际生产。 研制出的高速钢轧辊硬度超过66HRC,淬硬层深度超过30mm,辊面硬度差小于1.5HRC,耐磨性优于高铬铸铁和高镍铬无限冷硬铸铁轧辊,接近硬质合金轧辊。更多还原

【Abstract】 High-speed steel (HSS) roll containing high carbon, tungsten and vanadium has been optimizing developed. In view of the manufacturing Technologies of Centrifugal Casting High Speed Steel Roll, the modification and heat treatment of HSS roll have been studied intensively. Combining with the PLS algorithm of pattern recognition and BPN algorithm of neural net, the model in protecting process in the heat treatment of compound HSS roll was developed based on the MATLAB commercial software. High quality of HSS roll has been successfully manufactured.The carbides and matrix of cast HSS modified by potassium and rare earth (RE) were fined and broken network eutectic carbides were achieved, and many little lumps and grain carbides were produced as well. When modified HSS roll was heated at 1050℃, network eutectic carbides disappeared and many spherical carbides came out. The mechanical properties and wear resistances of modified HSS roll were distinctly improved, modification was bound to increase the thermal conductivity and decrease the hot expansion coefficient of HSS roll.The heat treatment of HSS roll was different from that of common high-speed tool steel. Fog quenching at 1025-1075℃ and tempering at 500-550℃, HSS roll turned out to have excellent mechanical properties and wear resistance. Two times temper was able to satisfy the requirement of HSS roll, while three times temper decreased the hardness of HSS roll.With the controlling technique of roll inner layer cooling, compound roll of different structures and hardness at inner layer and outer layer were obtained under heat treatment conditions. With the utilization of PLS and BPN algorithm to learn samples collected among the heat insulation plate thickness, the plate thermal coefficient, the coating thickness, the coating thermal coefficient and the inner layer hardness of HSS roll, the neural network model in protecting process was obtained in the heat treatment of compound HSS roll. Themodel could be used to predict the effects of insulting materials parameters on the inner layer hardness of roll during heat treatment for Centrifugal Casting High Speed Steel Roll. The neural network model turned out to be precise, efficient and great reliable in the application.The hardness of HSS roll was over 66 HRC, the depth of hardened layer excesses 30 mm, the variation of hardness in the roll surface was smaller than 1.5 HRC, and the wear resistance of HSS roll was better than that of high chromium cast iron roll and high nickel chromium indefinite hilled cast iron roll, close to that of powder metallurgy (PM) hard alloy roll.更多还原