毕业设计论文翻译.pdf

Improving Manufacturing Efficiency at Ford Using Product Centred Knowledge Management M B Raza T Kirkham R Harrison Q Reul Loughborough University Loughborough LE11 3TU UK The University of Nottingham Nottingham NG7 2RD UK Vrije Universiteit Brussel B R Muhammad lboro ac uk Thomas Kirkham nottingham ac uk R Harrison lboro ac uk Abstract Western manufacturing is under pressure to produce high quality customized products particularly in large manufacturing such as the car industry at low costs Here the development of such product customization requires the adoption of innovative agile manufacturing techniques To date this innovation has focused on the improved process development between the dif ferent stages of manufacturing Product Lifecycle Management PLM However in terms of implementing the application data management techniques have lagged behind often leaving these processes disjointed and lacking in automation This paper proposes an improved model based on innovation in the manufacturing PLM Building on existing work in the use of ontologies for knowledge management the paper applies these techniques to PLM The implementation has been applied to develop a case study around a Ford production line The prototype presents an innovative approach to PLM and tested using a state of the art Web Service infrastructure implemented on a Ford Powertrain test rig 1 Introduction Agile manufacturing relies on rapid configuration of assembly lines to support high levels of customization in product design The adoption of this type of specialist production is vital for the future of manufacturing in de veloped economies 1 In order to adopt the agile ap proach knowledge of product design needs to be quickly adapted to machine and line creation This can be achieved by greater integration of production and enterprise knowledge into the manufacturing process To date the better integration and management of this knowl edge has been a goal in large scale manufacturers such as Ford Motor Company However in production environments as in many other enterprise level computing environments legacy systems and vendor specific technologies persist These systems slow down and break up the integration process making it hard to achieve enterprise level agile techniques to sup port manufacturing assembly processes Without an over all view on integration and standardization the communi cation and knowledge capabilities of agile manufacturing cannot be realized The Business Driven Automation BDA project alongside the GAIN project at Loughborough University in partnership with Ford Motor Company focuses on ad dressing these challenges This paper presents a method by which knowledge can be better managed in automated production systems using the Ford Powertrain automated manufacturing environment as an example The approach focuses on the enablement of existing PLM systems to support enterprise wide agile manufac turing processes This is done by the development of PLM services to use knowledge based engineering tech niques using ontologies The project has modeled on tologies to capture existing data relations present at Ford in the manufacturing assembly processes Integration of this new level of knowledge has then been achieved with live production data using Web Services to illustrate agile processes in action on models of Ford production lines 2 Related work 2 1 Product Lifecycle Management PLM PLM has its roots in the initial integration of com puters with the manufacturing design process Data cre ated from CAD software has facilitated designers in the electronic creation reuse and manipulation of product models 2 To aid the integration of design data Product Data Management PDM systems have emerged to pre sent a means by which distributed access to design data from design teams can be achieved 3 However this combination of electronic design data and often localized software management has proven inadequate for the de mands of increasingly stream lined business processes Pressure has soon formed on PDM to integrate with other elements of the enterprise including non engineering ar eas such as sales marketing and supply chain manage ment as well as external agents like customers and suppli ers 4 Data use in PDM has therefore moved from a focus on product design to a need to present this data in order to enhance the wider manufacturing process 5 However as PDM systems have developed in specific organizations and around certain environments they tend to be vendor specific and hard to integrate in a generic manner Thus PDM can now be seen as a legacy element of a wider Product Lifecycle Management PLM process that en compasses all elements of the product manufacturing process To deal with such complexity PLM has largely devel oped around vendor specific or project product specific environments 6 PLM distinguishes itself from other enterprise application systems such as ERP SCM and CRM by presenting ways to enable effective collabora tion among networked participants rather than systems specifically in the product design process 7 8 The en ablement of PLM integration within and cross organiza tions in recent years has seen it move into the domain of service orientated computing 9 Using web services the concept of creating a loosely coupled PLM environment is seen as vital to support increasingly globalised manu facturing processes 10 However the shift to more flexible PLM implementa tions is a challenge of both data integration and manage ment As discussed current PLM systems are document oriented with no customizable data models and tied to vendor specific and legacy systems of integration Even using PLM flawed coordination among teams systems and data incompatibility and complex approval processes are common 11 12 To improve the PLM approach a focus on how data is both managed and captured in the manufacturing assembly processes needs to be adopted This will enable future PLM systems to better integrate in both terms of data and overall production process 2 2 Knowledge Management and Engineering Ontologies are often viewed as allowing more com plete and precise domain models 20 An ontology is commonly defined as a formal explicit specification of a shared conceptualization 22 More specifically an ontology explicitly defines a set of entities e g classes relations and individuals imposing a structure on the domain that is readable by both humans and ma chines As a result the domain knowledge represented in ontologies assists information sharing and re use Within engineering the use of ontologies has great po tential to aid knowledge management in these complex environments Ontologies are not only useful for achiev ing semantic interoperability on the Web but also for large organisations that attempt to coordinate a range of disparate expertise More specifically it will enable dif ferent communities to infer the same meaning when in formation knowledge is exchanged across systems For example Rolls Royce now needs to coordinate informa tion collection from various parts of the organisation since it has shifted its focus from selling products to pro viding services through maintenance service Therefore minimising the maintenance cost through out the life of an engine is fundamental to the success of this market shift As a result this company relies on knowledge stored in repositories created by different parts of the organisation The IPAS Integrated Products and Services project 21 developed ontologies describing products and processes to allow information recorded during the servicing of engines to be represented and shared Within Ford the same benefits can be realized through the use of ontologies particularly from a PLM perspec tive Vendor specific approaches to PLM and integration issues increase with the size of an organisation and cus tom integration often leads to integration issues re emerge 13 Thus to enable real change in the manufacturing as sembly processes the PLM approach needs to be changed from the bottom up The PLM approach of working with existing systems and data formats has often led to the corralling of data in centralized repositories 14 This presents problems as the means by which this data can be both understood and integrated often depends on user retrieval In an ideal world the data formats would be understood and accessed at the source where they are created in an automated way The rapid development of web services based comput ing has not only influenced the adoption of common stan dards to enhance business processes across enterprises but has presented ways of integrating cross organisational processes In order to aid this integration the concept of the Semantic Web has been developed to better aid the integration of varying forms of distributed data using on tologies 15 Linking the ontologies to aid PLM into Se mantic web services will allow greater knowledge of or ganisational data structures and improve processes and productivity at Ford 3 Case study Ford 3 1 Background The PLM system at Ford is a complex aggregation of several domains working collaboratively to manufacture and assemble different variants of vehicles PLM at Ford is managed using the Teamcentre that links product data from various CAD CAM designs stored in a repository Before the data is added to Teamcentre it has been sorted in order to fit into the appropriate reporting structures of the system By applying a more automated approach us ing SWSs at Ford it is envisaged both time and cost dur ing the process can be reduced 3 2 Problem Powertrain assembly systems The automotive assembly plant role in terms of Power train assembly and its relationship with PLM is the focus for this paper The assembly lines such as Powertrain assembly line for automotive engine have a limited ca pacity to produce variety of products A typical assembly process of automotive parts using Powertrain systems involves hundreds of individual parts Simply in terms of reconfiguration a product change and its impact on a change in any one part may cause a rippling effect in the whole assembly processes A key role of PLM should be to detect this effect in order that it can be managed effi ciently which found missing Currently Powertrain re configuration using PLM is product focused This creates problems as once the prod uct design is finalised it is difficult to manage and control the resources needed to re reconfigure the assembly line The management of processes and resources for both product and line needs to start as the product design is finished 4 Implementation 4 1 Environment The testing environment that we used for the runtime element of our design was constructed using Festo auto mation components These components were supplied by Ford and use the same interfaces as the machines on the Powertrain line at Ford plant in Dagenham UK Control of the interfaces was linked to a Web Service enabled control application developed for the Socrades EU Framework 6 project 16 Live data of the execution from the line is available through web service interfaces on the control application 4 2 Ontology A general layout of the line is captured as ontology and an instance of it is defined as Festo Rig with the current layout of the line where each component independent work unit has had its CAD data translated into ontology This allows the line components and layouts to be inter rogated by Semantic Web Services that can read the on tologies The ontologies are captured using the Prot g ontology design tool 17 and are presented as an ontology in the format OWL 18 In order to support this type of engi neering the product data needs to be both managed and integrated into the overall line design This can only be achieved by the linkage of PLM with the machine design lifecycle In the system the ontologies are available to services involved in both the runtime of the line and sup port of the wider product lifecycle For these services the ontologies present an improvement on the previous method of human based interaction with line and compo nent information stored in physical CAD diagrams Using semantic web services linked to ontologies many of these processes can be automated For example if the user wants to query about all sensor elements being used on a specific workstation or characteristics of a workpiece required to carry out a successful assembly operation can be answered with the help of ontologies and services 4 3 Service Interaction In order to demonstrate the use of Semantic Web Ser vices and the enhanced use of data they bring to PLM the rig was implemented with several supporting services These supporting services provide the function to both support the line design process in PLM and also add live analysis of line execution using the data from the orches trator linked to the rig The translation of the Teamcentre data stored in mul tiple largely CAD formats is achieved into the OWL format via the use of the ontology design tool Prot g In future applications could be developed to translate spe cific CAD formats into ontology standards The OWL ontologies currently are stored in a single location In the future it is envisaged that the ontologies will be stored alongside the data and possibly linked The main PLM functionality is provided by a stand alone PLM service This service will be a Web Service that reads in the ontology of the Teamcentre data The information from the ontology will be used to present knowledge of the current line status by using line data from the Control Service 4 4 Constraint Checking Central to the use of the ontologies is the ability for rules to be conducted on them For example if a new product is added to the line with a weight of 100 kg it is imperative that the components in the line can support that weight Therefore a key aim of the services is to check these new constraints and return areas where redes ign is needed In terms of the PLM service this function will be used for production reconfiguration process described in 5 2 For example a simple assembly operation of crankshaft with block this assembly operation is stored in PLM and can be readily accessed through rules and queries as de scribed above 5 Use cases The framework implemented had two main use cases demonstrating the use of semantic web services for knowledge capture and it use with the production process 5 1 Live System The live system uses the PLM service to notify inter ested parties of the events from the line For example if there is a jam on a conveyor or the current load level of the indexing table etc This information can then be used by the PLM to aid in the diagnosis of errors in the system If there are errors the PLM service uses the line ontology to find a remedy to treat the errors This process auto mates some of the response provided by a production engineer using the knowledge from Teamcentre via the ontology in a standardised way For example using the ontology the PLM service can instruct the control mechanism to notify dependant sta tions on the line that an error has occurred and even re quest a halt in the production In the case of a multi routed system the PLM could use the ontology of the line to diagnose a possible alternative route for the workpiece Notifications to engineers can be enhanced with the PLM services knowledge of the line from the ontology and live line data Outside of the line the use of the ontology by the PLM service will enable the notification of other appropriate services in the supply chain For example the ontology may be used to order a replacement part for the line by interrogating the components affected by the error A supporting knowledge base of previous faults linked to probable cause could aid in this process The PLM could also notify other services such as CRM and ERP of issues on the line and therefore potential impacts on production output 5 2 Production Reconfiguration In this scenario the line is reconfigured to support a new product Here the line ontology in its current live configuration is used alongside an ontology of the prod uct Engine assembly line is a highly sophisticated and complex combination of sequential operations and activi ties which are mostly automatic One of the aims of this research study was to develop methodologies to facilitate Ford company to visualise model and re configure new changed assembly line fairly automatically for build ing new changed engine Key concepts introduced into the ontology are work piece work piece characteristics workstation work station characteristics assembly operation operator and end product Based upon these concepts relations among products processes and resources are established into the ontology With the help of this knowledge in ontology a quick evaluation of many potential configurations is pos sible as well as the best suited one for a changed product 5 3 Evaluation Currently in Ford the process of preparing a line for a new product is both a manual and quite an unpredictable process The main problems in the process can be seen as the large amount of time taken between line specification and launch along with the unexpected costs of the proc esses A key aim of PLM at Ford has been to improve the speed and reduce costs by improving the efficiency of the line commissioning process However the structure of the process can be seen to remain the same Using services to the processes can be enhanced using knowledge from the PLM service linked to ontologies Using ontologies the new product specifications can be compared to the line layout in a more automated fashion With the help of ontology this process is becoming smoothened and helping Ford engineers to perform para metrical relationship analysis between engine and work stations with relevant assembly processes through ontol ogy For example this approach will automatically pick out any issues with the dimensions of the product and the dimensions of certain elements in the line These