t的液压机滑动块稳健优化设计

外文资料翻译译文1250T的液压机滑动块稳健优化设计摘 要为了解稳定变化下影响的液压机性能如载荷，材料性能的波动，零件和工作条件的物理尺寸，早期的设计过程中需要进行优化在本文中，滑动块，这是1250吨水压机的关键受力部件的设计参数，由邻居栽培遗传算法（NCGA）结合六西格玛强大的设计方法进行了优化通过模拟的有限元软件ABAQUS和多学科优化软件ISIGHT，结构设计，坚固性提高达99％，压力和原结构超重的问题都解决了它证明了六西格玛稳健设计法结合NCGA可以在设计初期提高结构性能的稳定性关键词：液压机，滑块，六西格玛的稳健设计法，邻里栽培遗传算法1 引言由于经济的快速发展，它是工业市场，一直是激烈如昔驱动制造商生产更强大和可靠的产品的竞争尽管执行情况变得越来越复杂，产品必须足够健壮以满足消费者的需求液压机的工作状况不佳和负载情况相当复杂在本执行中，液压机将满足负载，材料特性和所有意外的外部噪声的波动作为这样的结果，最佳的设计参数，这使产品对噪声不敏感的影响，也就是，该产品是健壮的，应该在早期设计阶段考虑在本文中，一个六西格玛稳健设计方法结合NCGA施加到设计用于搜索所述滑动块的最佳参数。

遗传算法是一种算法，模拟生物的遗传和进化作为遗传算法可以找到一个帕累托最优集合在多目标优化一项试验中，它是非常有效的液压机具有两个目的，最大压和体积的优化然而，仅与遗传算法不能推导参数，同时优化的平均性能和降低的性能差异因此，一个六西格玛稳健设计方法被引入到基于由遗传算法获得的参数的设在第2节中，NCGA算法，稳健优化设计专门六西格玛稳健优化方法的说明然后在第3，参数化建模和集成ISIGHT原始模型在第4节介绍在最后两节，多目标优化结果和六西格玛稳健优化设计，结果获得2 多目标优化方法的稳健设计法2.1 邻居种植遗传算法（NCGA）多目标优化同时解决了多个目标，通过多目标遗传算法（MOGA），其中发现在一19个优化的多目标优化问题的多重最优解，并获得有效竞争的目标之间的权衡关系的方法【1】 渡边慎也等人【2】引入了一个新的多目标遗传算法叫附近种植遗传算法（NCGA），其中除了其他MOGAS，如NSGA-II【3】和SPEA2 4的机制【4】下面的步骤说明NCGA哪里的总体流程PT：搜人口代位置：在归档一代第1步：初始化：产生初始种群P0人口规模为N.设置T = 0计算健身的价值最初的个人P0。

副本进入P0 A0档案大小也N.第2步：启动新的一代：设置T = T + 1第3步：生成新的搜索人群：PT = AT-1步骤4：排序：铂的个体被按照聚焦物镜的值进行排序聚焦的目标是在每一个所改变例如，当有三个目标，第一目标集中在第一代和第三目标被聚焦在第三代第一个目标是在第四​​代再次集中步骤5：分组：铂被划分成由两个个体的组这两个人被从选择顶部到排序的个体的底部第6步：交叉和变异：在一组，执行交叉和变异操作来自两家母公司的个体，产生两个孩子的个体这里，父代个体被消除步骤7：评价为：所有的个体的目标而得步骤8：组装：所有的个体被组装成一组并且这成为新的Pt步骤9：革新档案：组装Pt和AT-1一起在N个人从2N个人选择以降低个体的数量，则执行SPEA2的相同的操作（环境选择）在NCGA，这种环境的选择是作为一个选择操作施加步骤10：终止：检查终止条件如果满足，模拟结束如果不是，则模拟返回到步骤2在NCGA，大多数遗传操作的一组由两个个体被执行2.2 六西格玛方法优化2.2.1 确定性优化和鲁棒优化的区别坚固的设计是一个方法，它对立的确定性数学优化其典型地产生了推进设计约束边界的限制，而使在制造，材料和设计[5]很少或没有空间的不确定性的优化设计。

确定性优化和鲁棒性优化之间的差异示于图1确定性优化实现目标函数的最优解，而是将设计的变化，这可能会导致意想不到的性能敏感另一方面，由坚固的设计得到的溶液不仅适度在最优方面中的稳健性方面良好，但也良好，即，可靠的解决方案的分散体是窄针对设计变量的分散[1]确定性优化评价标准可以配制成：其中，y表示的质量性能; Y= E{Y}是期望或平均质量特性的值;δ=（γ - y）为y的标准偏差，它表示的质量性能方差;δ=（γ - ÿ）是质量性能，表示灵敏度的绝对偏差稳健设计的基本理念是开发稳定生产这与性能[6]的变化最小，针对产品的进程强大的设计方法被广泛使用，因为它们不仅可以提高产品和流程的质量，而且减少的性能变化而不消除变异的来源图1 确定性优化和鲁棒优化比较2.2.2 六西格玛方法在80年代初，田口玄一的方法走红，作为质量改善特别是在汽车行业的一个工具田口提供了一种专注于稳健的设计，达到成果在六西格玛运动【6】六西格马开始与目标提供产品和服务，这似乎没有缺陷形成消费者的目光【7】从标准差的统计概念和符号西格玛（σ）六西格玛方法的起源通常用来表示标准偏差缺陷数的严格的数学解释将导致不超过十亿分之二不良六西格玛标准。

然而，六西格玛过程假定一个分布，可以推卸中心高达+/-1.5σ，因此与传统的六西格玛相关的缺陷率的部分实际上是每百万3.4份除了数学计算，六西格玛过程中还涵盖了企业理念，涉及管理，财务，客户互动和过程的可预测性方面，它向着零缺点【8】的目标作出贡献在一个六西格玛稳健设计方法中，目标函数，同时含有均值μf和标准ΣF偏差，如下应最小最小化：这里ωu和ωδ是权重因子此外，以下不等式应该满足作为西格玛水平的制约则：这里，n表示西格玛水平和LSL/ USL分别表示下部/上部规格限制3 机产品仿真3.1 水压机模型聚焦液压机是1250吨至该楔形预压方法应用于帧类型因为方便装载和更合理紧迫的优点，楔形预紧方法比螺栓预紧的更好它是主体部分是它的身体承受所有的变形力时，机器运行机体主要由左右列，上梁，滑块，垫板，移动工作台和液压缸结构框图如图2所示1.柱2.上梁3.滑块4.移动工作台5.垫板图2 液压机的结构图3.2 有限元分析的前处理材料的性能如下，Q235：密度=7.86E3，弹性模量=206Gpa，屈服强度=为235Mpa，极限强度=425MPa，泊松比=0.29据认为，所述滑动块的移动是同步的液压缸和所述滑动块的固定点是相同的与滑动块和液压缸，固定的限制，这限制整个六个自由度的连接点，施加到3的表面连接所述滑动块，并在分析液压缸。

简化的模型已经采取由于大量的小的特点，对分析结果没有显著效果 1.42MPa压力被装载在滑动块的下表面上，并且4导向面被设置为表面到表面，它允许没有运动但方向相切的面连接滑块的剖视图示于图3图3 滑块的剖视图3.3 有限元分析（FEA）并ISIGHT集成滑动块是一个移动部件和一个重要应力成分当在满负荷条件下，系统的最大压达为25MPa基础研究，滑动块被视为静止状态，因此，静态分析取入本文由有限元分析软件ABAQUS和最大应力，这是达338.2MPa，发生在滑块和液压缸的交界获得滑块的冯米塞斯应力的结果如图4所示，滑块中心的应力比其它地方大因此中央加强筋厚度应优化图4 原始模型的应力分布ABAQUS的芯被控制以自动执行预处理和后处理步骤，分析所述计算的结果，作为二次开发，经由编程Python语言考虑到滑盖机型的简化，我们直接建模的ABAQUS7变量：X1〜X7加强计划如图5所示，我们推导出多目标功能如下：图5 的设计变量最小化：123457V = X 1+ X2+ 2 X 3+3X4+0.6X5+0.5X6则：2X1= X2X 5= X640≤x2≤6020≤x3≤3040≤x4≤6020≤x5≤3020≤x6≤30这里最大化是最大效应力在结构和V是总宽度。

在本文中，多目标优化软件ISIGHT平台集成了ABAQUS为了获得反应结果，液压机滑块模型文件被转移到ISIGHT然后传递到ABAQUS求解器来操作的有限元分析通过ISIGHT软件提供的NCGA算法具有遗传操作上设计变量和根据设定的周期数的目标时，流程图示于图6图6 ISIGHT运行流程4 鲁棒性比较结果在NCGA确定性优化算法，人口规模为5，代数为10，交叉型为1，交叉率0.6，突变率是0.006和迭代步数为150的最大米塞斯和减少V的迭代过程分别是在图8（a）所示和（b）图8 (a). 最大MISES NCGA迭代过程 (b). 第V NCGA迭代过程最低在表1中，它表明，最大应力降低30.54％，而宽度（重量）减少11.64％因此，我们可以得出结论，多目标优化结果达到减轻重量和减少压力的目的表1的模型优化结果的比较模型加强筋参数（mm） 最大屈服强度（Mpa)V的最小（mm)原来X2=50.00,X3=25.00,X4=50.00,X5=25.00, X7=25.00338.20327.50优化后X2=40.11,X3=24.90,X4=43.09,X5=29.74, X7=24.54234.90234.90然而，当单个六西格玛鲁棒检查被施加到多目标优化的结果，示出在表2该最大应力σ水平仅为1.55和相应的鲁棒值是77.89％，V的σ水平是8和相应的稳健值只是89.34％。

很显然，在最大应力的坚固性不能满足设计要求表2确定性优化结果六西格玛稳健检查参数名称均值标准偏差标准差成功概率X2(mm)40.100.500.800.5792597094X3(mm)25.000.500.670.565784238X4(mm)43.090.506.290.9999999997X5(mm)29.730.508.000.9999456297X7(mm)24.540.508.000.9567236452最大屈服强（Mpa）223.452.501.550.7779265486速度减少（mm）253.322.008.000.8934565246基于所述NCGA优化结果，六西格玛鲁棒性优化被应用，以2740号steps.The鲁棒性优化结果示于表3它表明，在优化最大应力是223.17MPa和最小V是300.63虽然最小V的有轻微的增加，最大压力减小，对最重要的，最大应力和最小诉的两个σ水平是8与99.9999891％和99.9999347％分别健壮值表2确定性优化结果六西格玛稳健检查参数名称均值标准偏差标准差成功概率X2(mm)40.071.381.230.782589585X3(mm)32.101.082.860.995811758X4(mm)53.601.803.670.999757071X5(mm)33.650.675.060.976558476X7(mm)23.970.802.770.999999591最大屈服强（Mpa）223.172.458.000.999999891速度减少（mm）300.631.968.000.9999993475 结论为了分析和优化滑动加强件的应力分布，三维模型被构造和优化方法，它结合了一个多目标遗传算法，即NCGA，与六西格玛稳健设计方法，是采取。

结果表明，该方法不仅可以减少滑块的最大应力和重量也达不到能够满足的稳健设计的要求的期望的鲁棒性此方法是一般适用于其它机械结构的优化问题致谢作者感谢文君张教授对许多宝贵的建议和很多为本文来源材料附外文资料原文Proceedings of the 4th ICMEMInternational Conference on Mechanical Engineering and MechanicsAugust 10−12, 2011, Suzhou, P. R. ChinaThe Robust Optimization Design of 1250T Hydraulic Press Slide BlockFan FAN1, Baochun LU1, Guojun YANG1, Yongzheng SONG2, Yulan DING31School of Mechanical Engineering, Nanjing University of Science and Technology, Jiangsu,China2Nantong Forging Equipment Co., Ltd, Jiangsu, China3Heavy forging equipment Engineering Research Center of Jiangsu Province, Jiangsu , ChinaAbstract: To stabilize the hydraulic press performance under the effects of variation such as the fluctuation of loading, material properties, physical dimensions of parts and operating conditions, the early design process needs to be optimized. In this paper, the design parameters of slide block, which is the key stress part of the 1250 ton hydraulic press, are optimized by Neighborhood Cultivation Genetic Algorithm (NCGA) combined with Six Sigma robust design method. By simulated in the finite element software ABAQUS and multi-disciplinary optimized software ISIGHT, the robustness of the structure design increases up to 99%, and the problems of stress and overweight of original structure are solved. It is proved that the Six Sigma robust design method combined with NCGA can improve the stability of structure properties during the early design stage.Keywords: Hydraulic Press, Slide Block, Six Sigma Robust Design Method, Neighborhood Cultivation Genetic Algorithm1 IntroductionSince the rapid development of the economy, it is the competition of the industrial market which has been as intense as ever that drives the manufacturer to produce more robust and reliable product. Notwithstanding the performing situation becomes more and more complex, the product has to be robust enough to meet the demands of consumers. The hydraulic press working situation is poor and load situation is quite complex. During the performance, the hydraulic press would meet the fluctuation of load, material properties and all unexpected external noise. As a result of this, the optimum design parameters, which make the product insensitive to the effects of noise, that is, the product is robust, should be considered in the early design stage.In this paper, a Six Sigma robust design method combined with NCGA is applied to the design for searching optimal parameters of the slide block. The Genetic Algorithm is an algorithm that simulates the heredity and evolution of living things. As the GA can find a Pareto-optimum set with one trial in multi-objective optimization, it is very effective for the optimization of the hydraulic press which has two objective, maximum press and volume. However, only with the genetic algorithm cannot we derive parameters which simultaneously optimizing the mean performance and minimizing the performance variance. Therefore, a Six Sigma robust design method is introduced into the design based on the parameters obtained by the GA.In section 2, the NCGA algorithm, robust optimal design specifically Six Sigma Robust Optimization Method are stated. Then the original model in section 3, parametric modeling and ISIGHT Integration in the section 4 are presented. In the last two sections, the multi-objective optimization results and the six sigma robust optimal design results are obtained.2. Multi-objective Optimization Method and Robust Design Method2.1 Neighborhood cultivation genetic algorithm (NCGA)Multi-Objective optimization solves the multiple objectives simultaneously via the methodology of multi-objective genetic algorithm (MOGA), which finds multiple optimal solutions of multi-objective optimization problem in one optimization and obtains the trade-off relation between competing objectives effectively【1】. Shinya Watanabe et al.【2】introduced a new multi-objective genetic algorithm called neighborhood cultivation genetic algorithm (NCGA), which has the crossover mechanism in addition to the mechanisms of other MOGAs, such as NSGA-II【3】and SPEA2【4】.The following steps illustrate the overall flow of NCGA wherePt: search population at generationAt: archive at generation.Step 1: Initialization: Generate an initial population P0. Population size is N. Set t = 0. Calculate fitness values of theinitial individuals in P0. Copy P0 into A0. Archive size is also N.Step 2: Start new generation: set t = t + 1.Step 3: Generate new search population: Pt = At−1.Step 4: Sorting: Individuals of Pt are sorted according to the values of the focused objective. The focused objective is changed at every generation. For example, when there are three objectives, the first objective is focused in the first generation and the third objective is focused in the third generation. The first objective is focused again in the fourth generation.Step 5: Grouping: Pt is divided into groups consisting of two individuals. These two individuals are chosen from thetop to the bottom of the sorted individuals.Step 6: Crossover and Mutation: In a group, crossover and mutation operations are performed. From two parent individuals, two child individuals are generated. Here, parent individuals are eliminated.Step 7: Evaluation: All of the objectives of individuals are derived.Step 8: Assembling: All the individuals are assembled into one group and this becomes new Pt.Step 9: Renewing archives: Assemble Pt and At−1 together. The N individuals are chosen from 2N individuals. To reduce the number of individuals, the same operation of SPEA2 (Environment Selection) is performed. In NCGA, this environment selection is applied as a selection operation.Step 10: Termination: Check the terminal condition. If it is satisfied, the simulation is terminated. If not, thesimulation returns to Step 2. In NCGA, most of the genetic operations are performed in a group consisting of two individuals.2.2 Robust optimization with six sigma method2.2.1 Difference between deterministic optimization and robust optimizationRobust design is a methodology which opposites to the deterministic mathematical optimization which typically yields the optimal design that pushes hard to the limits of design constraint boundaries, leaving little or no room for uncertainty in manufacturing, materials and design【5】. The difference between deterministic optimization and robust optimization is shown in Fig.1. The deterministic optimization achieves the optimal solution of objective function but is sensitive to the variation of design, which may lead to unexpected performance. On the other hand, the solution obtained by the robust design is not only moderately good in terms of optimality but also good in terms of the robustness, that is, the dispersion of robust solution is narrow against the dispersion of the design variable【1】. The evaluation criterion of deterministic optimization can be formulated as:Where y denotes the quality performance;y = E{y}is the expectation or mean value of quality characteristic;δ = ( y − y)is the standard deviation of y, which denotes variance of quality performance;δ = ( y − y)is the absolute deviation of quality performance, which denotes sensitivity. The basic philosophy of robust design is to develop processes that consistently manufacture products which target with minimal variation of performance[6]. Robust design methods are widely used because they can not only improve the quality of products and processes but also minimize the variation of performance without eliminating the sources of variation【5】。

2.2.2 Six sigma methodologyIn the early 1980’s, the methods of Genichi Taguchi became popular as a tool for quality improvement especially in the automotive industry. Taguchi provided a method focusing on robust design that reached fruition in the Six Sigma movement【6】.Six sigma starts with the target to deliver products and services which seem no defect form the eyes of consumers【7】. The Six Sigma method origins from statistical concept of standard deviation and the sigma symbol (σ)typically used to denote standard deviation. A strict mathematical interpretation of defect counts would result in a six sigma criteria of not more than 2 parts per billion defective. However, the Six Sigma process assumes a distribution which can shift off center as much as +/-1.5σ, and therefore the defective part rate traditionally associated with Six Sigma is actually 3.4 parts per million. In addition to the mathematical measurement, the Six Sigma process also encompasses a corporate philosophy involving aspects of management, finances, customer interaction, and process predictability, which contribute toward the goal of zero defects【8】. In a six sigma robust design method, the objective function, containing both mean value µ f and standard σf deviation, should be minimized as follows.Minimize:Where ωu and ωδ are weighting factors. In addition, following inequalities should be satisfied as constraints on sigma level.Subject to:Here, n denotes the sigma level and LSL/USL denote the lower/upper specification limits, respectively.3. Machine Product Simulation3.1 Hydraulic press modelThe focused hydraulic press is frame type of 1250 ton to which the wedge preload method is applied. Because of the advantages of convenient loading and more reasonable pressing, the wedge preload method is better than the one of bolt preload. It is the main body component which is its body that endures all deforming force when the machine is operating. The body mainly consists of left and right columns, upper beam, slide block, bolster plate, moving bolster and hydraulic cylinder. The structure diagram is shown in Fig.2.3.2 Pre-Process of finite element analysisMaterial properties are as follows, Q235: density = 7.86E3, elastic modulus = 206Gpa, yield strength = 235MPa,ultimate strength = 425MPa, Poisson ratio = 0.29. It is considered that the movement of the slide block is synchronous with the hydraulic cylinder and the fixed point of the slide block is identical with the junction point of the slide block and the hydraulic cylinder, fixed constraints, which restrain whole six degrees of freedom, are applied to the 3 surfaces connecting the slide block and the hydraulic cylinder in the analysis. Simplified model has been taken due to the large number of small features which has no significant effect on the analysis result. 1.42MPa pressure is loaded on the lower surface of slide block, and the 4 guiding face are set to connection of surface to surface which allows no movement but the direction tangent to the face. The cross-sectional view of the slide block is shown in Fig.3.3.3 Finite element analysis (FEA) and ISIGHT integrationThe slide block is a moving part and an important stress component. When in full load condition, the maximum press of system is up to 25MPa. For fundamental research, the slide block is regarded as static state, therefore, static analysis is taken in this paper. The Von Mises stress result of slide block is obtained by the FEA software ABAQUS and the maximum stress, which is up to 338.2MPa, occurs at the junction of the slide block and hydraulic cylinder. As is shown in Fig.4, the stress of the slider centre is larger than other places. Therefore the central stiffener thickness should be optimized.The ABAQUS core is controlled to automatically perform the pre-process and post-process, analyzing the compute result, as the secondary development, via programming Python language. Considering the slider model is simplified, we modeled directly in the ABAQUS with 7 variables:x1~x7 . The plan of stiffener is shown in Fig.5 and we derive multi-objective functions as follows:Minimize: Subject to: Here maximises is the max Von Mises stress in the structure and V is the overall width. In this paper, the multi-objective optimization software ISIGHT platform is integrated with ABAQUS. In order to obtain response result,the hydraulic press slider model file is transferred into ISIGHT and then delivered to ABAQUS solver to operate finite element analysis. The NCGA algorithm provided by ISIGHT software has genetic manipulation on the design variables and objectives according to the set of cycle number, the flow chart is shown in Fig.6.4. Results Comparison on RobustnessIn NCGA deterministic optimization algorithm, population size is 5, number of generations is 10, crossover type is 1,crossover rate is 0.6, mutation rate is 0.006 and number of iterative steps is 150. The maxmises and minimize of V in iteration process are shown in Fig.8 (a) and (b) respectively.In Table 1, it is shown that the max stress decreased 30.54% and the width (weight) decreased 11.64%. Therefore, we can conclude that multi-objective optimization results reach the goal of reduce weight and decrease stress.However, when a single six sigma robust check is applied to the multi-objective optimization results, it is shown in Table2 that theσ level of max stress is only 1.55 and the corresponding robust value is 77.89%, theσ level of V is 8 and the corresponding robust value is just 89.34%. Obviously, the robustness of the max stress cannot meet the requirement of design.Based on the NCGA optimization results, the six sigma robust optimization is applied, with 2740 numbers of steps.The robust optimization results are shown in Table3. It is shown that the optimized max stress is 223.17MPa and the minimum of V is 300.63. Though the minimum of V has a slight increase, the max stress decreases, to the most important, the both σ levels of max stress and minimum of v are 8 with a robust value of 99.9999891% and 99.9999347% respectively.5. ConclusionsIn order to analyze and optimize the stress distribution of the slide stiffener, a 3D model is constructed and an optimization method, which combines a multi-objective genetic algorithm, i.e. NCGA, with the six sigma robust design method, is taken. The results show that the method could not only reduce the max stress and weight of slide block but also reach a desired robustness that meet the requirement of robust design. This method is general applicable to the optimization problem of other mechanical structure.AcknowledgementThe authors thank Professor Wenjun Zhang for many valuable suggestions and much of the source material for this paper.References[1] Koji Shimoyama, Akira Oyama and Kozo Fujii, 2005, A New Efficient and UsefulRobust Optimization Approach–Design for Multi-Objective Six Sigma, IEEE Congress on Evolutionary Computation, p.950-957[2] Shinya Watanabe, Tomoyuki Hiroyasu, and Mitsunori Miki, 2003, Multi-objective Rectangular Packing Problem and Its Applications, Computer Science, Springer Press, p.565-577[3] K. Deb, 2001, Multi-Objective Optimization Using Evolutionary Algorithms, Wil。