英文翻译

专 业： 机械设计制造及其自动化 姓 名： 学 号： 外文出处：CHINESE

ENGINEERING

JOURNAL OF MECHANICAL

附 件： 1.外文资料翻译译文；2.外文原

文。

指导教师评语： 签名： 年 月 日

附件1：外文资料翻译译文

基于案例推理的二级齿轮减速器的系统设计

摘要： 二级齿轮减速器的设计是一个繁琐和耗时的过程。为提高效率和智能化的设计过程，基于案例推理(CBR) 技术被应用到二级齿轮减速器的设计。首先，分析了当前二级齿轮减速器的设计方法以及描述CBR 的原则。第二，根据减速器的特点，研究了CBR 的三个关键技术，提供了相应的方法，其中有如下：(a) 面向对象的表示方法、 (b)与归纳法标示最相近的案例相结合的检索方法 (c) 与基于可人工修订规则相结合的案例适应算法。此外，为改善案例检索的正确性，出现了一个新的决定特征权重的方法和相似性的公式，这是一个具有层次 分析法和粗糙集理论的组合权重方法。最后，根据上述分析结果，由 VC + +、 UG 和 Access 2003 开发了关于 CBR 二级齿轮减速器设计系统。这项研究提供了一种二级齿轮减速器的设计新方法。如果上述开发的系统应用于设计二级齿轮减速器，不仅提高设计效率，还可使设计人员从单调乏味的设计过程中出来，目的是把更多的精力投入创新设计方面。研究结果充分反映了CBR 技术在机械零件设计的可行性和有效性。

关键词： 二级齿轮减速器、 基于案例推理、 特征权重，相似性 1.介绍

二级齿轮减速器的传统设计方法是一个耗时的过程。即使是现在将CAD用在其设计过程中，传统的设计方法的到现在还未消除。为了改变这种情况，许多人都做了减速器设计方法研究，主要包括减速器参数化设计和专家系统。这两种方法有各自的缺点：参数化设计中每个部件的关系需要充分考虑，因为一个参数的改变可以导致的减速器其他参数或结构的变动，这就需要设计者熟悉减速器，可以使用设计软件，包括参数化设计功能。此外减速器的专家系统的知识获取是非常困难的，因为一些专家的知识是很难规则的表达。基于案例推理可以改善这两种的不足之处。基于CBR中已存在的案例已经完成一个新的案例，因此获得新案例可以通过修改某些部分或直接使用已存在的案例，与案例获取相比规则获取更容易，因为初级知识是在 CBR [1-2]。事实上，通常案例提供比规则或模型更多的信息。许多研究人员为CBR在工程中的应用做了大量的工作。Sun制定关于 CBR 的智能化夹具设计系统。Kwong通过CBR系统来介绍新的方法来确定适当的注塑参数。Liu提出一种检索算法，它与聚类技术集成在一起，以此在案例库中定位相似的案例，用一个案例库说明在机械设计中的CBR系统的可行性。Peter在CBR基础上开发智能知识系统，是为了支持预处理阶段在接触力学领域中的工程分析。Xiong基于CBR提供应用和创意概念设计方法，体现出工业设计理论。开发的系统帮助设计人员简化概念设计过程，并为后续产品的开发提供基

础。然而上文所述内容重视关于CBR理论研究优于CBR 技术与机械产品的组合。所以采取二级齿轮减速器为例，将讨论关于CBR机制部分。 CBR 2 关键技术

CBR 是一种相似或类比的方法。当CBR系统解决一个新的问题时，它从前期的案例中检索一个或多个与新问题最相似的案例，并修改以满足新条件。根据CBR的流程图，基于CBR产品设计系统的发展，需要解决一些问题，其中包括案例描述、 案例检索、 案例修改、案例库维护，其中案例描述、 案例检索和案例修改称为三项关键技术。案例描述是以电脑能够接受的传统符号将案例变成代码以及数据结构的过程。案例描述的一些方法包括帧、 对象、 述语、 语义网络和规则等，其中帧和对象最常用。案例检索是找出一个相互关联的案例或由特征参数和相似性找出更多类似案例的过程。注释 [9]将案例检索划分成三个部分：特性鉴定、初步匹配和最佳选择。本篇论文将案例检索划分成四个部分：关键特征提取、特征识别、初步匹配和最佳选择。数据集的数据挖掘通常包括多个特征属性，一些属性是与数据挖掘无关。那些不相关的属性会影响数据挖掘的效率。删除那些不相关的属性可以提高数据挖掘的效率，使得数据挖掘的结果容易理解。关键特征提取的目的是选择关键特征建立从案例库的有效索引。特征识别方法的目的是选择新案例的关键特征。特征识别往往可引用关键特征提取。因为它包含了充足的特征属性，所以二级齿轮减速器的关键特征提取更重要。这些典型的案例检索方法包括相似索引、 感应索引、知识引导、脉络知识索引和模板检索。实例修改过程是最好的匹配的情况下修改以满足新的设计要求，最常用的方法包括人工修改，专业性修改及非专业性修改。 3.在CBR 上设计二级齿轮减速器 3.1 案例库建设 3.1.1 案例描述

这篇论文使用面向对象层的案例表述模型，其向上层的软件提供一种统一的面向对象的数据模型，即提供所有类型的面向对象概念、 数据结构、 维修操作和灵活的扩展，并提供保护至底底层具体数据库的关系模型，同时通过映射原理、合理运行、有效管理来建立一个透明对象转换机制。映射原理中的对象模型是在上层的面向对象的数据模型和底层的关系数据库模型之间转换的关系。对象模型层、 数据库和其他功能模块之间的关系如图所示。

二级齿轮减速器是一个复杂的集合，为了方便的案例表示，二级齿轮减速器被划分成五个组件类和两个部件类。五个组件的类由高速齿轮组、 低速齿轮组、 高速轴和轴承、 中间轴和轴承、 低速轴和轴承，而两个部件类是覆盖件和箱体。这些组件类可以分为部件类。最后，每个部件类映射到其基础关系数据库中的表。

因此，二级齿轮减速器的面向对象的知识模型见图2。 3.1.2 案例库建设

根据二级齿轮减速器在图3中的表示模型，二级减速器特征属性值和其组件被映射到基础数据库，从而建立案例库。在基础数据库中，每个表包含\"案例编号\"的字段。这种关系通过减速器及其部件与表格之间映射建立起来。图4显示二级齿轮减速器在数据库软件Access中的存储模型。由于篇幅有限，论文中没有列出在 Access 中各部件的存储模式 3.2 案例检索

本篇论文应用一种将就近检索和编制索引相结合的方法，因为二级齿轮减速器的设计是一个具有悠久历史的经验过程，二级齿轮减速器包括很多案例。因此，它能更好地使用编制索引进行粗糙的检索，并使用就近原则进行最后检索。图5显示了详细的检索过程。接下来该论文将描述子进程和基于CBR二级齿轮减速器设计系统的案例检索方法。 3.2.1关键特征提取

原始案例库的特征提取是关键特征提取处理过程的前奏。第一步数据预处理。第二步关键特征选择。 3.2.2 初步匹配

初步匹配是从案例库中选择一组相互关联到当前的设计案例的过程。该过程是通过索引网络建造的所有关键特征和决策信息增益计算实现的。如图 6 所示是建立索引网络。案件的初步匹配是基于索引网络完成的。 3.2.3 最优选择

最优选择是从获得初步匹配的案例中选择最好案例的进程。最好案例由就近原则选出，所以它对于计算关键特征的权重和案例的相似性是不可或缺的。 3.3 实例修改

本论文使用规则修复和人工修复相结合的方法修改案例。实例修改流程图见图 7。规则包括经验公式，计算公式。例如，一个小齿轮的宽度通常是大齿轮宽度增加5-10 毫米。这些是评估设计可行性的主要措施，仅作为检查齿轮接触疲劳强度、 齿轮弯曲疲劳强度等主要公式。 4 系统的发展

基于 CBR 二级齿轮减速器的设计系统开发使用 VC + +，UG 和 Access2002 作为工具，其中包括五个模块的文件管理、 案例检索、案例修改，进入 UG、帮助。文件管理的模块用来管理有关案例的相关文件。案例检索模块用于检索案例和查看案例检索信息。实例修改模块用于通过案件编号修改案件检索。进入 UG 的模块用来从案例中修改参数和通过由ODBC 技术和 UG工具建成的子系统完成

3D 建模。 5 结论

（1）面向对象的知识模型描述了二级齿轮减速器的案例，它能更方便和准确的传递减速器相关知识，并有益于案例检索

(2)根据二级齿轮减速器设计的实际情况，提出了具有层次分析法和粗糙集理论的组合权重法，这可以提高案例检索的正确性。

(3)关于 CBR 二级齿轮减速器的设计系统研制成功。该系统使二级齿轮减速器的设计更加方便、 快速、智能。方法可以用作CBR 上的其他组件发展的参考。

(4)在案例库中足够的案件是在 CBR 上运行的二级减速器设计系统的前提条件。如果满足特殊二级减速器的设计要求，则有关的案件必须添加进入案例库。

附件2：外文原文

Design System of the Two-step Gear Reducer on Case-based Reasoning

Abstract: The design of the two-step gear reducer is a tedious and time-consuming process. For the purpose of improving the efficiency and intelligence of design process, case-based reasoning(CBR) technology was applied to the design of the two-step gear reducer. Firstly, the current design method for the two-step gear reducer was analyzed and the principle of CBR was described. Secondly, according to the characteristics of the reducer, three key technologies of CBR were studied and the corresponding methods were provided, which are as follows: (a) an object-oriented knowledge representation method, (b) a retrieval method combining the nearest neighbor with the induction indexing, and (c) a case adaptation algorithm combining the revision based on rule with artificial revision. Also, for the purpose of improving the credibility of case retrieval, a new method for determining the weights of characteristics and a similarity formula were presented, which is a combinatorial weighting method with the analytic hierarchy process(AHP) and roughness set theory. Lastly, according to the above analytic results, a design system of the two-step gear reducer on CBR was developed by VC++, UG and Access 2003. A new method for the design of the two-step gear reducer is provided in this study. If the foregoing developed system is applied to design the two-step gear reducer, design efficiency is improved, which enables the designer to release from the tedious design process of the gear reducer so as to put more efforts on innovative design. The study result fully reflects the feasibility and validity of CBR technology in the process of the design of the mechanical parts.

Key words: two-step gear reducer, case-based reasoning(CBR), weights of characteristics, similarity 1 Introduction

The traditional design method of the two-step gear reducer is a time-consuming process. Even now CAD is used in its design process, restriction of the traditional design method is not eliminated up to now. For the sake of changing this kind of situation, many people have done research on the design method of reducer, mainly including the parametric design of reducer and expert system(ES).These two kinds of methods have respective shortcomings: for parametric design, every part’s relation need to be fully considered, because a parameter’s change can cause a change of a reducer’s other parameters or structure, which requires designer to acquaint with a reducer and be able to use a design software including the function of the parametric design. Furthermore knowledge acquirement is very difficult in reducer’s ES, because some expert knowledge is very difficult to express with rules. Case-based reasoning(CBR) can improve these two kinds of shortcomings. A new case is finished on the base of an existed case in CBR, so the new case can be achieved by modifying some parts or directly making use of the existed case, and case acquirement is easier than rule acquirement, because the primary knowledge is cases in CBR[1–2]. In fact, case usually provides more information than rule or model[3]. Many researchers have

done a lot of work for applications of CBR in engineering. SUN, et al[4], developed an intelligent fixture design system on CBR. KWONG, et al[5], introduced a approach to determine proper injection moulding parameters by developed CBR system. LIU, et al[6], proposed a retrieval algorithm integrated with the clustering technique to locate the similar cases in the case-base and gave a case-base to illustrate the feasibility of the CBR system in the mechanical design. PETER[7] developed an automated knowledge-based system on CBR for intelligent support of the preprocessing stage of engineering analysis in the contact mechanics domain. XIONG, et al[8], provided an applied and creative conceptual design method based on CBR that embodies the industrial design knowledge. The system developed abbreviates the conceptual design process, help designers, and provides a base for the following development of product. However the works described above paid attention to the theoretical research on CBR superior to the combination of CBR technology with mechanical product. So, taken the two-step gear reducer as example, the whole process of mechanical parts on CBR will be discussed. 2 Key Technologies in CBR

The CBR is a kind of similar or analogical method. When a CBR system solves a new problem, it retrieves one or more cases from the antecedent cases that are the most similar to the new problem, and modifies the cases to satisfy the new situation[4]. The flowchart of the CBR is shown in Fig. 1. According to the flowchart of the CBR, the development of the product design system on CBR needs to solve some problem, including case description, case retrieval, case modification, case study and case base maintenance where case description, case retrieval and case modification are called three key technologies. Case description is the action that cases are coded to data structure accepted by computer with some conventional signs. Case can be described by some methods including frame, object, predicate, semantic network and rules, etc., among which, frame and object are most commonly used. Case retrieval is a process finding out one interrelated case or more similar cases by characteristic index and similarity[9]. Ref. [9] divides case retrieval into three parts: characteristic identification, preliminary match, and best selection. The paper divides case retrieval into four parts: key characteristics extraction, characteristic identification, preliminary match, and best selection. Dataset of data mining often includes many characteristic attributes, and some attributes are irrelevant to data mining. Those irrelevant attributes influence efficiency of data mining. Removing those irrelevant attributes can improve efficiency of data mining and make the result of data mining easier to understand. The purpose of key characteristics extraction is to select key characteristics to establish valid index from the case base. The purpose of characteristic identification is to select key characteristics of new case. Characteristic identification can be often made reference to key characteristic extraction. Because it includes a plentiful characteristic attributes, the key characteristic extraction of the two-step gear reducer is

more important. These typical methods of case retrieval include nearest neighbor, induction indexing, knowledge-guided, neural indexing on knowledge and template retrieval[10]. Case modification is the process modifying the best- match case to meet

the new design requirements, and the most used methods include artificial modification, knowledge-intensive modification and knowledge-lacking modification[10].

3 Design of the Two-step Gear Reducer on CBR 3.1 Case base building 3.1.1 Case description

The paper uses a case representation model of object- oriented layer[11–12], which provides an uniform object- oriented data model to the upper software, namely, provides all kinds of object-oriented concept, data structure, maintenance operation and flexible expansion, and provides a relation model of shielding concrete database to bottom, and establishes a transparent object conversion mechanism by mapping principle and carries on reasonable and valid management. Mapping principle of object model is a relation of the conversion between a upper object-oriented data model and a bottom relation model of database. The relation among object model layer, database, and other function module is shown in Fig. 2.

The two-step gear reducer is a complicated assembly, for the sake of convenience of case representation, the two-step gear reducer is divided into five component classes and two part classes. The five component classes are made of high speed gear group, low speed gear group, high speed shaft and bearing, middle shaft and bearing and low speed shaft and bearing, and the two part classes are made of the cover and the housing. These component classes can be broken up part classes. Finally, every part class is mapped to its table in the underlying relation database. Therefore, the object-oriented knowledge model on the two-step gear reducer is given in Fig. 3. 3.1.2 Case base building

According to the case representation model of the two-step gear reducer in Fig. 3, characteristic attribute values of the two-step gear reducers and their components are mapped to the underlying database, thus the case base is built. In the underlying database, each table includes a “case number” field. The relation is established by “case number” between the tables mapped by every part and by the reducer. Fig. 4 shows the storage mode of case base of the two-step gear reducers in the database software of Access. Due to the limited space, the paper doesn’t list successively the parts’ storage mode in Access 3.2 Case retrieval

The paper applies a retrieval method of the combination of the nearest neighbor and the induction indexing, because the design of the two-step gear reducer is an experience process with a long history, two-step gear reducer includes many cases. So, it is better to use the induction indexing to have a rough retrieval, and use the nearest neighbor to have a fine retrieval. Fig. 5 shows the detailed retrieval process. Now the paper describes the child process and method to case retrieval of the design system of the two-step gear reducer on CBR in turn. 3.2.1 Key characteristic extraction

The characteristics of original case base are pre- processed before the key characteristic extraction. First step is Pre-processing of data.second step is Key characteristic selection.

3.2.2 Preliminary match

The preliminary match is the process that a group of cases interrelated to the current design case are selected from the case base. The process is realized by index tree built by all key characteristics and decision information gain calculation. The index tree is built up as shown in Fig. 6. The preliminary match of case is fulfilled based on the index tree. 3.2.3 Best selection

The best selection is the process which the best case is selected from the cases acquired by the preliminary match. The best case is selected by the nearest neighbor, so it is indispensable for calculating the weight of the key characteristics and case’s similarity.

3.3 Case modification

The paper uses the combination of the revises on rule and artificial revises to modify case. The flowchart of the case modification is given in Fig. 7. Rules include experience formulas, calculation formulas. For example, the width of a pinion often is that of a gear with the addition of 5–10 mm. And restrictions are main measures of evaluating design feasibility, just as the check formula of the gear tooth contact fatigue strength, the gear tooth-bending fatigue strength, etc. 4 System Development

The flowchart of the design system of the second-class cylinder decelerator on CBR is shown in Fig. 8.

The design system of the two-step gear reducer on CBR is developed by using VC++, UG and Access2002 as tool[17–19], which includes five modules of file management, case retrieval, case modification, entering UG, and help as shown in Fig. 9. The module of file management is used to manage relevant files about cases. The module of case retrieval is used to retrieve cases and look over information of retrieved case. The module of case modification is used to modify case retrieved by case number. The module of entering UG is used to call modified parameter from case base and finish 3D modeling by the subsystem redeveloped by ODBC technology and UG/Open++ tool.

A practical example is given to illustrate the operation of the developed system on CBR as follows. Suppose to design the two-step gear reducer, and its design requirements: transmission power 5.36 kW, transmission ratio 11.06, life-span 310 000 h. The operation process of the system is described as follows. 5 Conclusions

（1）The case of the two-step gear reducer is represented by the object-oriented knowledge model, which is more convenient and accurate to express the knowledge for the reducer, and beneficial to the case retrieval.

（2）According to the actual situation of the two-step gear reducer design, the combinatorial weighting method with AHP and roughness set theory was proposed, which can improve the credibility of case retrieval.

（3）The design system of the two-step gear reducer on CBR was successfully developed. The system makes the design of the two-step gear reducer easier, rapider and more intelligent. The approach can be used as the development of other

components on CBR for reference. （4）That the sufficient cases in case base is the premise condition of running the design system of the two-step reducer on CBR. If the design requirements of special two-step reducer are met, the related cases must be added to the case base.