Original Article Proceedings of IDMME - Virtual Concept 2010 Bordeaux, France, October 20 22, 2010 HOME Yasser Rhahli 1, Magali Bosch-Mauchand 1, Bernard Anselmetti 2, Benoît Eynard 1 (1) : Université de Technologie de Compiègne CNRS UMR6253 Roberval BP 60319, rue du Docteur Schweitzer, 60203 Compiègne Cedex, France Phone : +33 (0)3 44 23 44 23 {firstname.lastname}@utc.fr Abstract: The issue of product tolerancing data management during its lifecycle is not completely solved and need to take into account the multidisciplinary nature of the engineering process, which includes design, analysis, manufacturing and control. Nowadays, Information Technology vendors provide a various Computer-Aided Tolerancing (CAT) systems which aid designer to address tolerancing challenges. But each system is used in a specific stage of tolerancing process and the results provided can t be used or explored by another tool. Moreover, these systems are not always integrated into CAD systems. This situation raises some questions that would need to be paid more attention in the future, especially: how to manage tolerancing data used or generated during tolerancing tas ks, how to ensure traceability of tolerancing data during product lifecycle, how to integrate CAT systems to CAD/CAM systems, how to integrate tolerancing data to PLM platforms. Key words: Tolerancing process, CAT systems, Product Lifecycle Management, Process Engineering. 1- Introduction Nowadays, the companies are under increasing pressure to keep their market safe in a global economic context. To improve their ability to launch products to market faster, and reduce products overall cost, they continue to improve their development process in particularly their tolerancing process. For aerospace and car manufacturers tolerancing task has become an important issue in development process because it affects both product performance requirements and manufacturing costs. Moreover, this task is carried out by several experts coming from different departments (Designer, Manufacturer, Controller...) which exchange and share tolerancing data and other information (functional drawing, design drawing, machining drawing, machine capabilities, etc.). Therefore, effective capture of information, and its extraction, recording, exchange and sharing become progressively important. These lead companies to review and improve their methodologies in managing the exchange and sharing information throughout the tolerancing process. (2) IUT Cachan, Univ, 61, Avenue du Président Wilson, 94235 Cachan Cedex, France IUT Cachan, Univ, 9, avenue de la division Leclerc, 94230 Cachan, France Phone: +33 (0)1 47 40 29 71 anselm@lurpa.ens-cachan.fr The goal of the paper is not to propose a novel method or approach to solve this problem, but to give a survey of the development status of tolerancing task. There are three main purposes: first, describe tolerancing process; second, analyze CAT systems, in order to understand its limits and its usefulness in tolerancing task; finally, formalize the expected requirements of industrial grade in order to improve future CAT systems and their integration with CAD/CAM systems, and therefore, with PLM p latforms. 2- Problem statement Tolerancing process is a sequence of dependent tasks. Each one uses the data produced by the previous tasks and generates another data for the next tasks. But, the sub-process related to each task is still unknown by other experts (designer, process engineer, manufacturer.). Recently, the need for each expert to know how these results were generated has been raised. Especially, all information as follows is required : - Calculating hypothesis - Computer Aided-Tolerancing Systems used - Problem solving methodology used - Data used in calculation - Results and risk associated Currently, each e xpert has his own methodology in managing tolerancing data throughout his perimeter and he shares only the results in a Product Data Management system (Fig 1), and therefore, in a Product Lifecycle Management platform. Moreover, there are many aspects which should be taking into account as follows: - Several CAT systems used for different goals; - Interoperability between CAT systems; - CAT systems not fully integrated into CAD systems; - Tolerancing process is not clearly defined; - Tolerancing data is not managed properly by a data management system which covers the whole Paper Number -1- Copyright of IDMME - Virtual Concept
tolerancing process. This situation makes it difficult to ensure the traceability between all data used and generated during tolerancing process, and therefore, between functional requirements and tolerances and/or geometrical specifications. In the industry, tolerancing process often starts during detailed design phase when product geometry is totally defined (Fig 3), and therefore any modification of a part dimension becomes a very complex and an expensive task. Furthermore, it is very difficult to keep traceability between functional requirements and functional tolerancing. Figure 3 Tolerancing process throughout Design process Figure 1 Data flows between tolerancing activitie s experts 3- Tolerancing task Basically, Tolerancing consists in defining coherent dimensional and geometrical constraints to ensure a good product performance. Salomons et al. [S1, S2] has divided tolerancing process to four major tasks (Fig 2): tolerancing specification, tolerancing analysis, tolerancing synthesis and tolerancing verification. Each task has a particular goal. Tolerancing specification is the task of specifying tolerances, defining the tolerancing types and related tolerancing values. Tolerancing analysis is a method which is used to verify the proper execution of the assembly after tolerances have been specified. It s used as bottom-up approach going from tolerancing value for each isolated part to the value of a functional requirement by calculating the total resultant on the assembly level. On the contrary Tolerancing synthesis is an up-bottom approach which consists in determining tolerancing value of each part in order to satisfy a functional requirement. Tolerancing verification defines inspection planning and metrological procedures for functional requirement in order to check the product conformity and to verify assumptions made by designers. There are some existing approaches dealing with tolerancing task in the whole product process [T1, K1] (survey papers). Ballu et al. [B1] has proposed GASAP approach for modeling as soon as possible parts, assembly and tolerance specifications in a CAD system. Mathieu et al. [M1] has proposed GEOSPELLING approach which is presented as the basic of a complete and coherent tolerancing process. Dantan et al. [D1] has proposed an integrated approach called Integrated Tolerancing Process (ITP), which ensures the continuous transition from functions to functional tolerances. They suggest to begin tolerancing process early in the design process in conceptual stage and progressively all along engineering tasks, taking into account, product functionalities and tolerances (Fig 4). Figure 4 Integrated Tolerancing Process [D1] 4- Current status of CAT systems Figure 2 Tolerancing process In the last few years, CAT industry has known a significant improvement and over the same period a number of CAT systems have been emerged. Each system is used for a special task in a specific stage of design process. In this section, the main CAT systems used in the current tolerancing practices are analysed. The information is obtained from CAT software vendors and some works related Paper P26-2- Copyright IDMME - Virtual Concept
to this subject. However, there are many existing CAT systems developed in the research area, Anselmetti et al. [A1, A2] has developed QUICK GPS (Geometric Product Specification), this CAT system generates automatically part tolerance specifications which is based on junction analysis between parts within the designer methodology. In the same time, designers used empirical approach based on their experience to define tolerancing values. But, the introduction of CAT systems has changed the rules. It has offered to designers a scientific approach based on modelling and analysing. Currently, CAT systems are designed to carry out two main tasks: tolerancing analysis and tolerancing synthesis. For tolerancing analysis aspect, they are based on standard arithmetic models (Worst case) or statistical models (Root Sum Square), and advanced simulation algorithm as Monte Carlo. The Monte Carlo simulation performs assembly tolerancing analysis using a random number generator which selects values for each manufactured variable, based on the type of statistical distribution assigned by the designer. The final result will more realistic than standard tolerancing analysis method. For tolerancing synthesis aspect, several CAT systems propose the possibility to identify main contributors, and therefore, the most influence link of a tolerancing chain. This functionality is very important for designers to optimize tolerancing values in order to reduce manufacturing costs. 4.1- CAD/CAT system integration The main CAT systems are not fully integrated to CAD systems (for examp le Ce-Tol provided by Sigmetrix is not integrated to Pro/ENGINEERING) except the one that were developed by CAD providers or their partners (3DCS for CATIA). In spite of being fully integrated CAT systems use only geometric data based on DMU. Then, designers have to specify manually all information needed for modelling his system (parts, links, geometries, dimensions, dimensional and geometric tolerances). In the further stages of tolerancing process, designers use a specific system to generate Geometric Dimensions & Tolerancing annotations of DMU (Functional Tolerancing & Annotation of Dassault System, Product and Manufacturing Information of UGS Siemens PLM). These systems are fully integrated with some CAT systems (FT&A with 3DCS, PMI with NX Quickstack) and therefore, they generate automatically GTA of DMU which is not the case for the other CAT systems. Table 1. CAT system characteristics CAD system Tolerance analysis Method Tool Provider CATIA SolidWorks NX Pro/ENGINEER Worst case Statistic Objective Input Output Mecamaster Assembly Mecamaster X X X Validate design concepts and define frontier tolerances Anatole EADS X X X 3DCS DCS X X X Validate NX Quickstack UGS Siemens PLM em-tolmate Tecnomatix X X X X X Ce-To l6 sigma Tolerance Manager X Sigmetrix X X X X PCO Innovation X X X X X GA IA EADS X Unknown Unknown assembly of tolerance part Rough geometry, Building philosophy (kinematic scheme) Complete geometry with GD&T, assembly sequence, Manufacturing capabilit ies Influence analysis of each part deviation on one functional requirement Main contributors of tolerance chains Paper P26-3- Copyright IDMME - Virtual Concept
4.2- CAT system throughout design process CAT software editors propose a large scale of systems to guide designers to specify and evaluate tolerancing part and assembly. Each system covers a limited perimeter of design process (Fig 5). This situation leads designers to use different systems in order to carry out their tolerancing analysis. CAT systems can be classified into two major categories according to three points: analysis objective, data used and generated data (Table 1). The first category (Mecamaster Assembly and Anatole) is used to validate design concepts and define frontier tolerances. Thus, the analysis is based on few elements such as rought geometry, building philosophy (kinematic scheme) in order to generate influence analysis of each part deviation associated to one functional requirement. The second category (3DCS, NX Quickstack...) is used to validate assembly of toleranced parts. Thus, complete geometry with GD&T, assembly sequence, manufacturing capabilit ies are used in order to identify main contributors of a tolerancing chains. Figure 5 CAT systems throughout design process 4.3- CAT systems benefits and limits CAT systems own two major benefits as follows: Speed of calculation: software performance calculating time is very quick. And therefore, it permits to carry out a complex assembly like a gear box in little time. Choice between several analysis approach: arithmetic analysis (Worse Case) and statistical analysis (Root Sum Square and Monte Carlos) As the same time, they are many aspects that would need to be paid more attention to in the future CAT system, especially: Manual modelling: designers still specify all information regarding their system (parts, links, geometries, dimensions, dimensional and geometrical tolerances) wh ile CAT should use all CAD data. Each system covers a limited perimeter of design process. Solids are considered infinitely rigid: current systems cannot be used for non-rigid solids which limits the usefulness of CAT tools. Interoperability: each system uses its own language and display results regarding a specific structure which cannot be used or explored by another system. Integration with CAD systems: ma jor systems still do not use all information contained in CAD file. Manual annotation of the geometric specifications and the tolerances. Manual modification of nominal dimensions of DMU. 5- Toward integration of Tolerancing Data Management (TDM) in a PLM system. Tolerancing task is carried out by several e xperts coming fro m different departments (Designer, Manufacturer, Controller...). For each e xpert, it is essential that the produced information should be easily availab le. Du ring preliminary design, the information produced should be documented simultaneously with tolerancing work and all the collected information should be recorded and capitalized in an integrated information processing system. This information should easily used by other experts. In this way, an efficient collaborative working approach can be created inside the company between all experts involved in tolerancing process. Thus, TDM system becomes progressively crucial for tolerancing task, and therefore for design task. TDM must be fully integrated in a PLM p latform to ensure the traceability of tolerancing data within product lifecycle. There are some existing approaches dealing with data model for a TDM system [F1, Z1]. S. C. Feng et al. and X. Zhao et al. have proposed two data models supporting tolerance data, those models should be improved to take into account PLM context. Product Lifecycle Management Product Data Management Tolerancing Data Management Figure 6 Tolerance Data Management (TDM) 6- Data modelling method for TDM system Nowadays, it is necessary to elaborate a data model which represents tolerance information with a unified, Paper P26-4- Copyright IDMME - Virtual Concept
unambiguous and human-computer understandable language. This section presents a description of data modelling process steps. 6.1- Requirements of a TDM data model. Defining data requirements is the first step in a data modelling process. The data must not focus on current needs but also on future industrial needs. Thus, it is very important to involve aerospace and automotive industry in this work. Therefore, current research work, international standards, industrial practices, scientific state-of-art, technological state-of-art and new technologies should be considered. Cost estimation related to tolerance specification is an important aspect; it should be taken into account in TDM data model [MS1]. 6.2- Information modelling languages. An information model language specifies the formal s yntax and semantic of data elements and their relationships. Recently, several international standard data modelling language have emerged in order to model and specify product structure: UML (Unified Modeling Language) [O1,BR1], EXPRESS [I1], and XM L [BC1]. UM L remains a popular standard for expressing the product structure. It can be easily applied to the development of a data translator, because it is made for the development of software systems. 6.3- TDM model as an extension of a PDM data model. PDM manages all information related to the product, process and organization information. At the present time, PDM systems do not support product data related to tolerancing data which should be integrated in a future generation of PDM systems. Eynard et al. [E1, E2] have proposed a UML approach for the specifications of a PDM system implementation. This approach can be used to specify a TDM model which can be considered as an extension of a PDM model, and therefore, of a PLM model. 7- Conclusion It is widely acknowledge that aerospace and automotive industries are more than ever obliged to improve their design process particularly their tolerancing process to be competitive. The management of product tolerancing data within its lifecycle is one of the major problems not solved yet well by taking into account the multidisciplinary nature of the engineering process. Currently, CAT vendors are focusing their energy on technical aspects by developing tools that help more designers to solve a specific problem instead of interesting to organizational aspect. Future research should focus on new process, methodologies and tools which helps tolerancing experts on managing the tolerancing data exchange and sharing throughout product lifecycle. 8- References [A1] B. Anselmetti, Generation of functional tolerancing based on positioning features, Computer-Aided Design, 38(8): 902-919, 2006. [A2] B. Anselmetti, R. Chavanne, J.X. Yang, N. Anwer, Quick GPS: A new CAT system for single part tolerancing Computer-Aided Design, 42(9): 768-780, 2010. [B1] A. Ballu, H. Falgarone, N. Chevassus, L. Mathieu. A new Design Method based on Functions and Tolerance Specifications for Product Modelling. CIRP Annals - Manufacturing Technology, 55(1): 139-142,2006. [BC1] S. Boag, D. Chamberlin, M.F. Fernández, D. Florescu, J. Robie, J. Siméon, XQuery 1.0: an XML query language, Technical report, W3C, 2003. [BR1] G. Booch, J. Rumbaugh, I. Jacobson, The Unified Modelling Language User Guide, Addison-Wesley, Reading, 1999. [D1] J-Y Dantan, N. Anwer, and L. Mathieu. Integrated tolerancing process for conceptual design. CIRP Annals - Manufacturing Technology, 52(1):135 138, 2003. [E1] B. Eynard, N Troussier., and B. Carratt, PLM based Certification Process in Aeronautics Extended Enterprise, International Journal of Manufacturing Technology and Management, 19(3-4): 312 329, 2010. [E2] B. Eynard, T. Gallet, P. Nowak, and L. Roucoules, UML based Specifications of PDM Product Structure and Workflow, Computers in Industry, 55(3): 301-316, 2004. [F1] S. C. Feng and Y. Yang. A dimension and tolerance data model for concurrent design and systems integration. Journal of Manufacturing Systems, Vol. 14/No. 6, 1995. [I1] ISO/DIS 10303-11, "Product Data Representation and Exchange - Part 11: The EXPRESS Language Reference Manual," International Organization for Standardization, Geneva, Switzerland, 1993. [K1] M. Khanafer, A. Desrochers, L. Laperrière, Tolerancing assistance methodology in a product life cycle perspective, Anaheim, CA, USA, 2007. [M1] L. Mathieu, A. Ballu. GEOSPELLING/ A common language for specification and verification to express method uncertainty. In: Proc. of 8 th CIRP Seminar on Computer Aided Tolerancing. 2005. [MS1] M. Mauchand, A. Siadat, J.Y. Dantan, N. Perry : A proposal for decision-making system aiming to support cost estimation during the preliminary design, in Proc. of IDMM E-Virtual Concept 2006, Grenoble (France), May 17-19, 2006. [O1] Object Management Group, The Unified Modeling Language, Release 1.1, Reference Manual, 1997. http://www.omg.org/. Paper P26-5- Copyright IDMME - Virtual Concept
[S1] O.W. Salomons, F.J. Haalboom, H.J. Jonge Poerink, F. van Slooten, F.J.A.M. van Houten, and H.J.J. Kals. A computer aided tolerancing tool ii: Tolerancing analysis. Computers in Industry, 31(2):175 186, 1996. [S2] O.W. Salomons, H. J. Jonge Poerink, F. J. Haalboom, F. van Slooten, F. J. A. M. van Houten, and H. J. J. Kals. A computer aided tolerancing tool i: Tolerancing specification. Computers in Industry, 31(2):161-174, 1996. [T1] S. Tich kiewitch, and D. Brissaud. Diverse aspects of tolerancing on an integrated design context. CIRP Annals - Manufacturing Technology, 48(1):107 110, 1999. [Z1] X. Zhao, T.M. Khethara Pasupathy and R. G. Wilhelm. Modeling and representation of geometric tolerances information in integrated measurement processes. Computer in Industry 57 319-330, 2006. Paper P26-6- Copyright IDMME - Virtual Concept