【作者】 王泽武；

【导师】 王乘； 樊建平；

【作者基本信息】 华中科技大学 ， 系统分析与集成， 2007， 博士

【摘要】 环件轧制技术是借助于轧环机使环件产生壁厚减下、直径扩大、截面轮廓成形的局部塑性加工工艺,具有节能、节材、生产率高、生产成本低、产品范围广等显著特点,在机械、冶金、化工、能源、航空航天等许多工业领域中得到日益广泛的应用,成为轴承环、齿轮环、法兰环、火车车轮、燃气轮机环等各种无缝环件零件的先进制造技术和主要加工方法。近20年以来,随着有限元理论逐步完善和相关有限元软件广泛推广,基于通用有限元软件采用计算机模拟环件轧制过程代替传统依靠工程师经验的试错法成为研究的热点。然而环件轧制不仅具有普通的平板轧制、异步轧制、多道次轧制的性质,而且涉及到芯辊和锥辊直线进给运动、主辊旋转轧制运动、抱辊导向运动以及环件自身的转动和直径扩大运动,因此在同一数值模拟仿真模型中既要模拟环件复杂的金属流动,又要实现各轧辊运动轨迹的实时变化,一直成为环件轧制数值模拟研究的难点。为了跟踪这些研究热点,解决这些难点问题,本文在前人研究工作的基础上开展了以下几个方面的研究工作。(1)首先基于通用动力显式有限元软件LSDYNA建立了环件轧制数值模拟仿真模型,采用质量缩放等技术有效降低计算时间,克服了环件轧制数值模拟过程对计算机性能和长时间计算的要求;然后采用大型结构分析有限元软件ANSYS的APDL参数化语言,建立了环件轧制过程中复杂数控系统仿真模型,实现了环件轧制运动参数实时调用修正,克服了在环件轧制金属流动模拟过程中不能求解含未知变量的问题;最后对这两个模型集成建立了环件虚拟轧制自动修正有限元仿真模型(简称AMFE),从而实现环件金属流动和轧环机数控系统耦合实时仿真。(2)基于AMFE虚拟轧制仿真模型对某矩形截面环件的一个生产周期轧制过程进行了数值模拟研究,计算结果和德国SMS Wagner-Banning公司的ROLLTECH实验经验数据、美国P. V. Ranatunga博士的UBET计算数据进行了对比分析,结果显示AMFE模型仿真数据与上述数据基本吻合,从而验证AMFE仿真模型的可靠性,同时AMFE模型仿真结果还显示了环件轧制过程任意时刻的应力、应变和位移云图,以及环件轧制实时扩展和缺陷生成的动态过程,而这些数据是环件轧制工程师急需想要而其它研究方法很难得到的,验证AMFE仿真模型具有精度好、功能强、效率高等特点。该环件实时动态虚拟轧制方法的成功实现,突破了传统环件轧制过程二维、三维局部或者三维瞬态的研究。(3)基于AMFE环件虚拟轧制仿真模型,进一步对环件虚拟轧制过程优化进行探讨,针对于不同情况,提出两种优化方案,一是采用MATLAB神经网络工具箱对环件轧制规程、初始毛坯进行优化分析,二是以缩短轧制时间为优化目标,建立环件虚拟轧制过程优化仿真模型,对优化变量、优化策略、优化过程进行了详细分析。并以某矩形截面环件虚拟轧制仿真模型为研究对象,采用两种方案分别对环件轧制过程进行优化,对优化变量(材料模型、质量缩放、网格细化、芯辊进给、锥辊进给、磨擦系数等)进行了优化对比分析,最后对环件虚拟轧制过程整体优化分析,得出了合适的轧制规程、轧制毛坯及其他轧制工艺参数。这些问题的研究超越了前人仅仅停留在如何建立环件轧制三维仿真模型问题上的研究,从而拓宽了环件轧制数值模拟技术研究领域。(4)为了把AMFE环件虚拟轧制仿真模型延展到异形截面环件轧制过程的研究,论文对异形截面环件虚拟轧制过程特点进行了分析,对一些关键技术提出了解决方案,分别建立了Φ500型径向轧环机、RAW200/160-5型径轴向轧环机的AMFE虚拟轧制仿真模型,在国内外首次对汽车后桥从动伞齿轮锻件、飞机发动机涡轮机匣锻件和600MW核反应器壳体大型锥形环件三种异形截面环件的一个生产周期内的轧制过程进行了虚拟轧制仿真实证分析,仿真计算结果显示与实际生产过程基本吻合,从而验证了AMFE环件虚拟轧制仿真模型可应用于异形截面环件的虚拟轧制,可用于检验异形截面环件轧制成形制造工艺,可用于对新型环件的轧制工艺和轧制过程的开发,这些对于降低研发成本、缩短研发周期、快速响应市场、实现绿色制造和提升国内外环件轧制水平具有重要意义。更多还原

【Abstract】 Ring rolling technology is a specialized partial plastic rolling process in which a pre-designed cross-section is formed using a rolling mill. Very often, the wall thickness of the ring is decreased with increase in the diameter. Due to its competitive advantages, such as high efficiency and low cost, this technology has been widely used in mechanics, metallurgy, energy, aerospace, chemical engineering. Nowadays, ring rolling is one of the major advanced methods for production of seamless annular-shaped components, such as bearing races, gear ring, flange ring, railway wheel, turbine ring and pressure vessel ring. The development trend of ring rolling nowadays is to achieve high productivity, precision, complexity and flexibility in production. The traditional trial-and-error method can not respond rapidly to the market change and realize the green manufacture.During the past 20 years, with the rapid development of finite element theory and the availability of commercial finite element codes, the trial-and-error approach has been replaced gradually by numerical simulation. However, the ring rolling process carries the characteristics of plate rolling, asynchronous rolling and multi-way rolling. Moreover, it also relates to the feed movement of the mandrel and the axial rolls, the rotation of the main roll, the movement of the guide rolls as well as the rotation of the ring itself with the expansion of its diameter. Therefore it is difficult to determine the metal flow process using conventional methods. Particularly, the forming process is affected real-time by the change of each roller movement, which is the main barrier in the virtual ring rolling research. To overcome this barrier and address the research focuses, the following tasks have been completed on the basis of some former research work:(1) Firstly, a numerical metal flow model for the ring rolling process was created by using the general dynamic explicit code LSDYNA. In the simulation model, the mass shrink technology was also adopted to reduce the running time, therefore the model released the general requirement of high computer performance and long computation time. Secondly, the ring rolling control process was modeled using a complex numerical simulation technique that made use of the finite element software ANSYS with APDL (ANSYS Parametric Design Language). With the real-time transfer and the modification of movement parameters, the numerical control model could overcome the difficulties in determining the unknown parameters. Lastly, the two simulation models were integrated to form a self-adjusted ring rolling finite element simulation tool (AMFE) for realizing real-time coupling simulations for both the ring metal flow and the numerical control system.(2) Based on the AMFE, the rolling process of a ring with rectangular cross-section was simulated for one complete cycle of production. The simulation results were compared with ROLLTECH experimental data reported by the German SMS Wagner-Banning company and the UBET results reported by P. V. Ranatunga. The AMFE was hence validated as good agreement between these results had been found. Moreover, the AMFE simulation also provided dynamic information on the stress, strain and displacement contours of the ring as well as the damage evolution process. Apart from many other advantages, such as high calculation efficiency and versatility, the information provided by the AMFE is very important to conduct a ring rolling process, but usually difficult to be determined experimentally. Therefore, this project is amongst the first attempt to simulate the entry ring rolling process and overcome the limitations in conventional transient studies.(3) Two optimization schemes were firstly developed and incorporated into the AMFE simulation model. One was to utilize the MATLAB neural networks toolbox to optimize the rolling process and the initial billet structure of the ring. The other one was to create the optimum model for the ring rolling process, in which the rolling time was taken as the target function and the optimization variables, the strategies as well as the process were analyzed. With the former rectangular cross-section ring being taken as the research object, the rolling process was optimized separately by using the two proposed schemes and the optimization variables were analyzed in detail. Eventually, the whole process of ring rolling was optimized with the optimum rolling process parameters and rolling billet structure found. The contributions of this research do not only create a 3D finite element model, but also provide a computational tool for optimization of a ring rolling process.(4) Further research work was conducted to study the ring rolling process of a profiled cross-section ring using the AMFE simulation model. The rolling characteristics of a profiled ring were firstly analyzed, and some solution schemes were proposed accordingly. The AMFE models for theΦ500 radial ring rolling machine and RAW200/160-5 radial-axial ring rolling machine were also created. The whole dynamic rolling process within a complete production cycle was simulated for three kinds of complex profiled rings: a rear axle bevel gear blank, an aero-engine turbine casing blank, and a great conical ring of 600MW nuclear reactor shell. The simulation results matched well with the actual production process, and thus it was proved the validity of the virtual rolling model. The method for studying the profiled ring rolling process may also be used to examine the feasibility of the profiled ring rolling process, and develop a rolling process for a new cross-section profiled ring. The deliverables of this project contribute to save the R&D cost, shorten the overall cycle time, respond more rapidly to the market requirement, and help realizing the green manufacture objective.更多还原