back to SP2

on to SP4

1 Introduction

The general objective of the SP3 toolset is to serve as a link between product design, process design, system design and feedback collection from the actual line. Specific objectives of the SP3 toolset:

• to develop methods and tools based on concurrent engineering approaches, which enable the use of virtual assembly for process design and system planning.
• to integrate geometrical and non-geometrical features and knowledge from the product and assembly process.
• to include models of human-machine collaboration in the planning phase, as well as the operation of assembly lines.

Toolset architecture will be developed based on open technologies and communication standards. Thereby the novel collaborative human-machine approaches play a dominant role. The research work will cover physical feature models of assembly process related components, tolerance and tolerance chain models. This includes the specification of a vendor-neutral structure and the development of a repository of these models. Regarding the human-robot cooperation skill-oriented assembly, planning tools will be developed, considering, sequencing and scheduling based on manual and automatic assembly. Furthermore, an open knowledge base (KB) will be developed in order to ensure the future access to assembly structures (product-, process- and system knowledge). Finally a toolset prototype will be implemented, tested and evaluated using the IAS prototype environment.

2 SP3 Toolset

2.1 Knowledge Base and Middleware

The knowledge base has three main goals; the first is to act in the simple case as a PDM (Product Data Management) vault. The second goal is to provide a standardized interface between different design support system clients. The clients are able to share the information more efficiently since the native formats are mapped into the CoreOntology. The third goal is to provide knowledge reasoning and parsing based on the defined rules, mating conditions and earlier solutions saved on the KB.

The three knowledge domains, defined by ontologies, product, process and system domain, are</p> implemented using semantic mapping with concept relations to connect the knowledge from one domain to the. The ontologies were developed by using Protégé 3.3.1 and exported straight to PostgreSQL, which is a relational database. The connection to the relational database is done through the Jena Framework. All the connections from the clients to the knowledge base are done trough Web Services. This increases the security since clients don’t have direct access to the database nor can make changes to restricted areas of the knowledge. When a request is done to a Web Service, this service uses the Jena Framework to complete the request and handles an answer back to the client. The storage of information for each concept or entity is implemented with RDF files for data management and X3D (eXtensive 3D, a successor of VRML) for visualization.

2.2 Connections between Clients

2.2.1 Connection through the Knowledge Base

The KB interface for each client provides connectivity between the knowledge base and the surrounding clients. The interface provides seamless translation from specific domain scope to knowledge model scope and vice versa. This is required for the knowledge base not to build up an uncontrollable mass of redundant data, but instead to maintain a semantic database containing rules for translating terminology and semantics. An example of this is a client reading XML-files and sends specific data through SOAP to KB for storage or a client providing an output in a native binary format and taking as inputs data obtained using Web Service-requests. Currently a client has been developed to request and send information to the KB. This was possible by taking advantage of the COM interface of the client. In addition to this client, several other clients are being developed at the moment to use, export and import data from the KB.

2.2.2 Verification of the Stored Knowledge and Relations between Ontological Representation

The interface for each client provides access to the knowledge model mapped with the client-specific semantics and functionalities. The clients will connect through middleware. The consistency of the ontology was validated through KB Web Client. Currently the possibilities of the client’s queries are based on the main classes of the ontology.

2.2.3 Connection between CAMeLEAN and the KB

The connection between CAMeLEA and KB is done via mapping the ontological representations of the both tools. The CAMeLEAN ontology is very general, in order to allow the usage of the tool in different companies and for various industries. The CoreOntology in the KB goes a bit deeper in defining the relationships between the classes. The Knowledge Base ontology acts in a type of network with nodes in a type of n-dimensional space. The ontology of CAMeLEAN is limited by the nature of the XML-file trees and forms a type of 2-dimensional network. Fortunately in CAMeLEAN the products, the resources and the operations can be nested almost infinitely. This is also true for the related templates. Templates on different nesting levels can have different parameters, which via heritage can be given to the lower levels of the same tree. The combination of object levels and object templates allows for mapping of the product, resources and operations objects from the Knowledge Base into CAMeLEAN and backwards. The actual mapping of between the CAMeLEAN and KB is done via parsing the XML-files into the RDF representation and is saved to the KB.

2.2.4 Connection between 3DCreate and CAMeLEAN

3DCreate and CAMeLEAN close the loop between process plan creation and simulation of that plan. Information can be shared through a semiautomatic feedback loop that updates both the process plan in CAMeLEAN and the simulation model in 3DCreate. This feedback loop can be achieved through the knowledge base or through a simple spread sheet. Currently the connection between 3Dcreate and CAMeLEAN is done by utilizing the spread sheets and macros. The following figure shows the potential connection between CAMeLEAN and 3Dcreate through the Knowledge Base.

2.2.5 Connection between 3DCreate and 4DIAC

A reconfigurable robotic system is used at PROFACTOR to demonstrate some advantages of the
4DIAC initiative for the PISA project. A modular 6-DOF robot that is composed out of
mechatronic components is used, as depicted in Figure 5. The robot consists out of six separate
joints, which are PowerCube Modules provided from Schunk company. They are connected with
each other via special connector elements. The hardware (embedded control) setup of the
reconfigurable robotic system is also shown in this figure.

Respectively, two robot axes are controlled by PC/104 embedded PCs equipped with a Debian Linux and an OSADL real-time kernel patch. Furthermore, the FORTE runtime environment is installed on all the PC/104 controllers. The data exchange between the PC/104 and thePowerCubes is done via CAN. The engineering and visualization PC is connected via standard Ethernet to the PC/104 controllers.

The engineering of the IEC 61499 robot control program is carried out via the 4DIAC-IDE. The visualization and HMI are also modeled with IEC 61499 FBs. Therefore, a visualization extension based on the wxWidgets to the FORTE runtime environment is used for the execution of the visualization application. The hardware structure of the modular 6-DOF robot described above has lead to the implementation of a Robot Control Application (RCA) and a Robot HMI Application (RHA). The RCA contains the implemented control concept as an IEC 61499 application. Each component of the control concept has its own IEC 61499 FB representative. For visualization and parameterization issues, a specific RHA was developed. The RHA is also built out of special IEC 61499 HMI Service Interface Function Blocks (SIFBs). The two applications communicate with each other via Publish/Subscribe SIFBs.

The IEC 61499 system model for the robot controller consists out of four different remote devices. One remote device is executed on the Visualization PC, the other three devices are
executed on the PC/104 embedded hardware (all equipped with the FORTE). The communication among all devices is done via Ethernet. The whole RHA is mapped to the visualization device while the trajectory calculation and the feed-forward control function blocks are mapped to the PowerCube 1 & 2 Device. A High-Gain Observer FB is mapped to the PowerCube 5 & 6 Device. The other control modules (CAN SIFB and controller FBs) are represented as sub-applications in the RCA and are distributed to the three PowerCube devices. For the interaction of the PC/104 embedded controllers and the PowerCube modules the special
CAN SIFBs have been developed.

The 4DIAC toolset has been integrated to 3DCreate, to validate the control logic in a simulated device prior to the deployment of the device, or the actual building or construction of the machine, reducing time-to-production of factory floor equipment by having the control logic debugged and ready for production the moment the machine is ready. The advantage of this approach is that the control logic can be downloaded and tested on the simulation, and afterwards the same control logic can be deployed to the hardware, as depicted in the preceding figure.

2.3 Client Specific Development

2.3.1 Resource Manager in 3DCreate

3DCreate includes a new and innovative Resource Modeling Infrastructure as part of the core simulation engine that allows system integrators and end users to simulate, analyze and optimize the resources in the factory floor prior to deployment. Resources can be anything in a factory environment, such as cranes, machines, transport systems or humans.

2.3.2 Life-cycle Engine in CAMeLEAN

The latest version V1R8.3 of CAMeLEAN supports the usage of the life-cycle engine for all objects within the CAMeLEAN process plan. In the current research activities it was proposed to define life-cycle states, which reflect the various levels of potential re-usage of a production resource. The resources can get new life-cycle states, which are related to calendar periods after the SOP date. The process planning engineer, who has the intention to upgrade a production line from an existing one to a future one, which should deliver a new product, has different needs for life-cycle states than the production engineer, who runs a current production line. Once the life-cycle states are defined and linked to the individual resources, the process planning engineer can use the life-cycle states for filtering the available resources with respect to these life-cycle states.

3 Conclusion

The previously accomplished work forms a solid basis for the integration of the required tools and methodologies needed to fulfill the requirements from the industry. The integration of the toolset is proceeding steadily and is expected to be finalized during the third year.

List of Related Publications

M.Lanz, F.Garcia, T.Kallela, R.Tuokko, <i>Product-Process ontology For Managing Assembly Specific Knowledge between Product Design and Assembly System Simulation, IPAS'08 10.-12.2.2008 Chamonix, France, p.10.

M.Lanz, T.Kallela, G.Velez, R.Tuokko, Product, Process and System Ontologies and Knowledge Base for Managing Knowledge between Different Clients Proceedings of the 2008 IEEE International Conference on Distributed Human-Machine Systems, March 9-12, 2008 Athens, Greece, pp. 608-513.

M. Lanz, T.Kallela, E.Järvenpää, R.Tuokko, Ontologies as an Interface between Different Design Support Systems, WSEAS Neural Networks 2008, 2.-4.5.2008 Sofia, Bulgaria, p.6.

M. Rooker, T. Strasser, G. Ebenhofer, M. Hoffmann, R. Velez Osuna, Modelling Flexible Mechatronical based Assembly Systems through Simulation Support, Proceedings of the 13th IEEE International Conference on Emerging Technologies and Factory Automation, September 15-18, 2008, Hamburg, Germany.

T. Strasser, M. Rooker, G. Ebenhofer, Distributed Control Concept for a 6-DOF Reconfigurable Robot Arm, Proceedings of the 4th I*PROMS Virtual International Conference, July 1-14, 2008. R. Velez, Infrastructure for the Design, Planning and Optimization of Flexible Hybrid Assembly Systems, Proceedings of the 2008 IEEE International Conference on Distributed Human-Machine Systems, March 9-12, 2008, Athens, Greece, pp. 514-518.