Högskolan i Skövde

Virtual engineering increases the rate of and diversity of models being created; hence requires maintenance in a product lifecycle management (PLM) system. This also induces the need to understand their creation contexts, known as historical or provenance information, to reuse the models in other engineering projects. PLM systems are specifically designed to manage product- and production-related data. However, they are less capable of handling the knowledge about the contexts of the models without an appropriate extension. Therefore, this research proposes an extension to PLM systems by designing a new information model to contain virtual models, their related data and knowledge generated from them through various engineering activities so that they can be effectively used to manage historical information related to all these virtual factory artifacts. Such an information model is designed to support a new Virtual Engineering ontology for capturing and representing virtual models and engineering activities, tightly integrated with an extended provenance model based on the W7 model. In addition, this paper presents how an application prototype, called Manage-Links, has been implemented with these extended PLM concepts and then used in several virtual manufacturing activities in an automotive company.

Manufacturing data and information are produced and used during the lifecycle of product development. Product lifecycle management (PLM) systems provide a suitable platform for managing them. For appropriate management of manufacturing data, it needs to be identified, classified, and stored based on the structure of PLM systems. In this paper, the results of an industrial manufacturing data collection study are interpreted, and their relation to the main structures in PLM systems is specified. Subsequently, a new information model for assigning this data and information to the PLM data model is presented. The main contribution of this information model is the definition of property and change objects and integrating them with the structure of PLM systems; changes and revisions of those data are formally defined and hence traceable.

Reusing previously designed virtual models and the knowledge extracted from running them can reduce time and costs. Since these models are representations of physical artifacts, they have been built based on some criteria, assumptions, and limitations. Being aware of these criteria, assumptions, and limitations, as well as having information regarding things like the purpose of the virtual model, can help a user evaluate the model for reuse in another study. The knowledge gained by generating and using virtual models also offers a significant advantage when making decisions about whether to reuse them and how to reuse them in new studies and experiments. This historical information and knowledge can be derived from the engineering activities performed when creating and using those virtual models. The research presented in this dissertation deals with the management of virtual factory artifacts, including virtual models, their related data, and historical information and knowledge in product lifecycle management (PLM) platforms. Since PLM systems are developed to manage product- and production-related data, they are suitable for managing virtual models as well, but they are incapable of handling the knowledge generated without appropriate extension. Therefore this research focuses on extending PLM systems so that they can manage historical information related to virtual models, in addition to managing the data and knowledge generated, together with related engineering activities. This management will be based on various requirements, including saving, searching, and retrieving virtual factory artifacts in different projects, studies, and engineering activities. A new information model was developed taking into account four aspects of management, namely: (1) hierarchical structures in PLM systems and manufacturing data; (2) the lifecycle of manufacturing systems; (3) projects, studies, and experiments; and (4) engineering activities, provenance data, and knowledge management. In addition to managing knowledge extracted from virtual models and their utilization, ontology-based knowledge management for virtual models and engineering activities is also provided to build up an ontology in this domain. Ultimately, the developed information model was implemented in several application studies in the industry. These studies cover different types of virtual models from various levels of manufacturing systems. The applicability of the invented information model in these application studies has confirmed its capability to manage and link virtual factory artifacts in a PLM context.

Using virtual models instead of physical models can help industries reduce the time and cost of developments, despite the time consuming process of building virtual models. Therefore, reusing previously built virtual models instead of starting from scratch can eliminate a large amount of work from users. Is having a virtual model enough to reuse it in another study or task? In most cases, not. Information about the history of that model makes it clear for the users to decide if they can reuse this model or to what extent the model is needed to be modified. A provenance management system (PMS) has been designed to manage provenance information, and it has been used with product lifecycle management system (PLM) and computer-aided technologies (CAx) to save and present historical information about a virtual model. This chapter presents a sequence-based framework of the CAx-PLM-PMS chain and two application case studies considering the implementation of this framework.

Discrete event simulation (DES) models imitates the behavior of a production system. Models can be developed to reflect different levels of the production system, e.g supply chain level or manufacturing line level. Product Lifecycle Management (PLM) systems have been developed in order to manage product and manufacturing related data. DES models is one kind of product lifecycle’s data which can be managed by a PLM system. This paper presents a method and its implementation for management of interacting multi-level models utilizing a PLM system.

Saving and managing virtual models’ provenance information (models’ history) can increase the level of reusability of those models. This paper describes a provenance management system (PMS) that has been developed based on an industrial case study.

The product lifecycle management (PLM) system, as a main data management system, is responsible for receiving virtual models and their related data from Computer-Aided technologies (CAx) and providing this information for the PMS. In this paper, the management of discrete event simulation data with the PLM system will be demonstrated as the first link of provenance data management chain (CAx-PLM-PMS).

Anything that is produced and offered to the markets can be a product. Each product has a lifecycle, whether the product is a car or a software program or a service. Product lifecycle has different phases from the time that it is raised as business idea at the first phase to disposal as the last stage. Product lifecycle management is a business approach for management and the use of product, process and resource related data, information and knowledge. With this management, enterprises try to use their company’s intellectual capitals. The initial PLM systems had been developed to store and manage Computer Aided Design (CAD) files and giving access to these data in different stage of the product lifecycle for users. Afterward, PLM systems, more developed to cover the management of process and resource data, and later on, managing of product related data, information and knowledge on all phases of product lifecycle. Each enterprise, according to its needs and competencies, implementing and using different capabilities of PLM systems, but still Bill of Material (BoM), Bill of Process (BoP) and Bill of resource (BoR) are forming the core of PLM systems. PLM systems try to manage data by integration with other engineering software programs to import data from them and manage those data for exporting to other software programs or makes them accessible for users. These integrations cause the managing of virtual data and information by the PLM systems, which are generated by different engineers such as designers or manufacturing engineers. CAD files and simulations are two types of virtual data. These data consist of some knowledges that had been generated by different engineers, which can be called virtual knowledge.Real World Knowledge is another type of knowledge that are exist in the enterprises. This knowledge can be captured from the happenings in the real environment such as failure reports, quality and audit reports, product performance, production data and operator’s experiences. Sometimes capturing these knowledges is very easy for example production throughput, but sometimes it is very hard, because they are unwritten and uncodified.Capturing and managing these real world knowledges, can help manufacturers to reduce their costs by making a better decisions and reusing of virtual models.Firstly these knowledges can clarify consequences and of previous decisions. They can also clarify some hidden and unconsidered issues about decision cases. The Real World Knowledge covers different types of knowledge, such as production reports, maintenance reports or operator experience.Secondly, the real world knowledge, can support to determine the level of virtual confidence (Oscarsson et al., 2015). Virtual models as one kind of virtual knowledge which had been explained before, have been designed to reduce costs by simulating the reality. The correctness and accuracy of a virtual model, clarify the level of confidence for that model and its results, for reusing that model to solve another problem. With comparing of the real world knowledge and virtual models expectations, the accuracy of the model can be evaluated, and the reliability of the model can be measured.There are lots of knowledge management systems have been developed, but most of them are trying to manage the organizational knowledge. The focus of this research is collecting the real world knowledge, in an automotive industry and converting them to the usable and classified format. Afterward, those knowledges should be stored and managed in the extended PLM platform.

Shortening of the product development process time is one of the main approaches for all enterprises to offer their products to the market. Virtual manufacturing tools can help companies to reduce their time to market, by reduction of the engineering lead time. Extensive use of virtual engineering models results in a need for verification of the model’s accuracy. This virtual engineering usability and assessment have been named virtual confidence. The two main factors of the achievement of this confidence are the accuracy of the virtual models and the virtual engineering results.

For controlling of both above factors, a complete virtual model and related virtual model knowledge are needed. These knowledges can be tacit or explicit. For exploring explicit knowledge, a data and information collection from different disciplines in the organization is needed.

In this paper, a data map with focus on the manufacturing engineering scope will be presented. This data map is generated from different data sources at a manufacturing plant, and gives an overview of different data that exist at different data sources, in the area of manufacturing. Combining real world data from different sources with virtual engineering model data supports, amongst others, establishment of virtual confidence.

This report is about the product data classification methodology that is useable for the Volvo Cars Engine (VCE) factory's production data, and can be implemented in the Teamcenter software. There are many data generated during the life cycle of each product, and companies try to manage these data with some product data management software. Data classification is a part of data management for most effective and efficient use of data.

With surveys that were done in this project, items affecting the data classification have been found. Data, attributes, classification method, Volvo Cars Engine factory and Teamcenter as the product data management software, are items that are affected data classification. In this report, all of these items will be explained separately.

With the knowledge obtained about the above items, in the Volvo Cars Engine factory, the suitable hierarchical classification method is described. After defining the classification method, this method has been implemented in the software at the last part of the report to show that this method is executable.