Design and Development of Architecture-based Information System for an Automated Manufacturing System
M. I. Jambak, A. M. Shaharoun, A. A. Razak
Faculty of Mechanical Engineering
Universiti Teknologi Malaysia
81310 Skudai, Johor Bahru, Malaysia
Phone: +60 - 7 - 5504566
Fax: +60 - 7 – 5566159
, ,
Abstract. Automation has dramatically increased productivity in manufacturing. A complete set of information is needed by various components in order to operate a manufacturing system efficiently. Manufacturing information for such systems is complex. The complexity arises from data sharing, time-critical, high volume of data, heterogeneous systems, uncertainty and the dynamic nature of change of manufacturing system. As the centre of an information system, the database management system (DBMS) should be able to support the information system requirements. The development of the database should be based on a suitable architectural design in order to cope with changes, minimise data redundancy, reduce repetition, and coping with uncertainty. Although knowledge of the information architecture is well known in computer-integrated manufacturing (CIM) area, the application of it is still limited. One of the main reasons is that existing information architectures are too general. A major effort is required in developing the information systems. On the other hand, the development of an information system that is not based on the information architecture will create rigidity in its structure. Nevertheless, the development of an information system from scratch requires more effort, budget and time for the manufacturing companies. This paper contributes to bridging the gap between the general information of the existing information architectures and the detailed information needed by manufacturing companies in order to develop their own extendible information system.
1 Introduction
Industry today is characterised by automation. Engineering processes such as machining, fabrication etc are mainly performed by computer-controlled machines. These perform the operations automatically and are often integrated into to form an automated manufacturing system. The computer control system is the brain that controls the production system components. Computer-controlled manufacturing system is an information-driven system to support various machine controllers, which require information to perform their tasks on the shop floor. The information system that is incorporated into the control systems, play its role by co-ordinating the production activities and tracking the system status to manage the production operations. The essence of such information system is its database (11). One of the challenges in developing the database for any manufacturing application is to organise all the data into a useful, accessible and logical form (5). Two strategic approaches for tackling this problem are, firstly the creation of the information model that maintains its characteristic of wide applicability and flexibility, and secondly the use of an existing CIM architecture (5). In the next sections, the generic requirements of an information system for automated manufacturing system and the method to develop the information system based on the proposed information architecture are described.
2 The Generic Requirements of Information Systems for Automated Manufacturing Systems
The generic requirements of information systems in automated manufacturing systems can be found in the literature by (12), (17), (6), (21), (3) and (10) and can be summarised as below:
ability to handle a large quantity of information,
ability to handle complex and heterogeneous information,
ability to handle frequent updating of data,
ability to deal with uncertainty within the manufacturing system,
ability to inter-operate amongst the functions of the manufacturing system,
ability to establish communication links between different types of control systems,
consistency between the local and global information,
ability to adapt to either the sort-term change or long-term changes of a manufacturing system.
(15) noted that, “a complete catalogue of information about all the items in a factory could take in the neighbourhood of 1015 bytes…”. The volume of information increases exponentially in order to handle the whole enterprise and therefore databases need an efficient storage mechanism for fast data access and retrieval (9).
The complexity of information system is not only caused by the heterogeneous nature of manufacturing information system (because of the compartmentalisation of manufacturing system functions) but also due to the components of manufacturing system being supplied by different proprietary information-technology products vendors (12) (17). For instance, the design data from computer-aided design (CAD) applications may have different data formats according to the suppliers and this data has to be accessible to other manufacturing applications.
The speed of operation of the information system will have a significant impact on some aspects of shop floor control, such as cutting tool monitoring during the machining process (17). The information system should reduce the operation time and the information system must allow very frequent updates to obtain the required speed of operation (14).
The behaviour of a manufacturing system is plagued with uncertainty. Unpredictable interruptions, and consequent changes in the system status, occur due to breakdowns of machines, breakage of cutting tools, improper functioning of materials handling systems, sudden introduction of ‘urgent’ jobs, etc (17). The manufacturing information system should cope with these uncertainties and be able to adapt quickly to the unscheduled change by improving its information management efficiency. The information system should also be supported by a DBMS that is robust enough to withstand improper shutdown, allowing an automatic backup and ease of change.
Besides uncertainty, the manufacturing system is also prone to change. By the time a new factory is designed, its initial configuration may have changed its initial design (12).
The requirements of an information system that are described above show the necessity for proper design of the architecture of a manufacturing information system. The concept of architecture basically proposes a development framework to develop the system where a methodology (or group of methodologies) and tool (or set of tools) is used to model and implement the system.
3 Information Architecture
The information architecture expresses the subject area of information that is needed and used by an enterprise, and may include the entities and the relationships of those entities within the subject areas considering the architecture characteristics (i.e., extendibility, expandability, and adaptability) (13). The information architecture provides a guide for developing the information system and for its subsequent modification. The information architecture for an automated manufacturing system should contain a generic and complete set of information model as a means of communication during a system (re-) design process. Based on the information architecture, the users can easily develop their own databases. Figure 1 shows how the information architecture can guide the users to develop their own data structure model. The information architecture reduces the user efforts in developing their own information system compared to if they were to develop it from scratch.
4 Case Study
A research project at Universiti Teknologi Malaysia will be used to illustrate the issues involved in practically implementing the development of an information system.
The objective of the project is to develop a database for a flexible manufacturing system (FMS) that is able to cope with change and based upon an information architecture. The scope of the project is limited to the small batch FMS and focus on the machining of prismatic components. The point of view adopted in this research starts from the system level to equipment level of the manufacturing system.
5 The Proposed Methodology and Tool for Development of Information System
There is no best method for the design and implementation of information systems, just as there is no best method for the design and implementation of manufacturing system (15) (16). Some of the existing information architectures such as “Purdue Enterprise Reference Architecture” (PERA), “CIM-Open System Architecture” (CIMOSA), “GRAI Integrated Methodology” (GIM), and “Generalised Enterprise Reference Architecture and Methodology” (GERAM) in the literature survey by (8), (9), (19), (20), (22), (2), (18), and (7) have been studied for this work. A number of traditional database development methods have also been studied before the following method is proposed and is shown in figure 2.
Four major stages to develop an information system for such system were adopted. These are system requirement specification, conceptual design (or information architecture development), detail design, and implementation. All the concepts and information from the literature survey are presented as the models at the information architecture development step. This information architecture consists of functional model and information model and is developed by using two of IDEF[1] family modelling language provided by IDEF™ integrated software. Information architecture is designed to be independent from database conceptual schema types such as relational or object-oriented type, which the user can decide upon at the detail design and implementation stages.
6 The Proposed IDEF0 Functional Model
There are four major functions to perform the manufacturing activities, where two of them are at the planning (system) level namely process planning and resources planning. The others are at the execution (cell, workstation, and equipment) level namely shop floor control and shop floor activities as shown in figure 3. However, the functions of a manufacturing system are not limited to the ones listed above. Depending on the designer viewpoint, other functions might also be incorporated.
7 The Proposed IDEF1x Information Model
In determining the information flow, the IDEF0 functional model was analysed. The components in the IDEF0 blocks, such as input, control, output and mechanism (symbolised by an arrow) gives an understanding of the information that needs to be captured. Figure 4 shows the common information entities and attributes captured for establishing the manufacturing activities.
8 The Implementation of the Proposed Information Architecture
The conceptual data model is mapped onto the logical data model in the detail design stage. Since the logical data model is database schema type-dependent, the database conceptual schema type must be chosen for mapping the conceptual data model to the particular logical data model. The chosen database conceptual schema type is relational because of two reasons. Firstly the relational and object-oriented are the most commonly used and widely available in the market. Indeed, (6) found that relational and object-oriented databases are more flexible than other types of databases. Secondly the IDEF1x modelling language is closer to the relational type and has a facility to export the model to the relational database manipulation language called structured query language (SQL). The relational DBMS (RDBMS) software’s that been chosen for the work are MS Access and Oracle DBMS. MS Access was used in the pilot study, which was intended to develop a PC-based information system. It is later intended to migrate the system onto Oracle, DBMS that has a larger capability as the final target.
9 Discussions
The methodology that was used to develop the information system provides a formal and structured framework for the system development effort. Although it requires a lot of discipline, prior planning and discussions before execution, the direct benefits of the structured method could be observed through out the rest of the design and implementation process. The system designer was able to reduce the number of possible errors and mistakes on ad-hoc method would have caused. It provides clear milestones providing easier tracking of the project status. It was also possible to relate each piece of information with respect to its origin, destination and functions within the overall system. Also, as further changes were made to the functional and information models, it was possible to establish the impact of these changes upon the whole system. The use of the methodology was fully indicated when it was implemented in the case study. The tools that were used in this case are IDEF0 and IDEF1x have limitation that can fully support the methodology if as in the conceptual design it is required for the behaviour or dynamics system should be modelled too. The strength of IDEF1x lies in its roots and the weakness of IDEF1x is that it has a very limited space in every page and therefore the modeller must be experienced in order to create good models. There are two ways of creating the tables. The first option is to directly create the table using table wizard facility in the DBMS. The second option is export the IDEF1x model to the SQL file and execute the commands to create the tables.
10 Conclusions
This paper has described the generic requirements of an information system in an automated manufacturing system. This research work contributes to bridging the gap between general information of the existing information architecture and the detail information needed by manufacturers in order to develop their own information system. The information architecture was implemented onto the database tables and preliminary assessment on the information architecture is currently being conducted.
11 Acknowledgement
This research is conducted under an IRPA project vote number 72091 at the Faculty of Mechanical Engineering, Universiti Teknologi Malaysia.
12 References
1. Ang, C. L., Automatic generation of IDEF0 models", in Journal of Intelligent Manufacturing, Chapman and Hall, 1994, pp. 79-92, No. 5.
2. Domeingts, G. et al., GIM (GRAI Integrated Methodology) and its evolutions: A methodology to design and specify Advanced Manufacturing Systems, Information infrastructure for manufacturing (B-14), Yoshikawa, H. and Goosenaerts, J. (Eds.), Netherlands: Elsevier Science B.V. (North-Holland), 1993, pp. 101-117.
3. Gong, Dah-Chuan and Hsieh, Yueh-Wen, Conceptual design of a shop floor control information system, in International Journal Computer Integrated Manufacturing, Taylor and Francis Ltd., 1997, pp. 4-16, Vol. 10, No. 1-4.
4. Hars, A., Heib, R., Kruse, Chr., Michely, J., Scheer, A., -W., Reference Models for Data Engineering in CIM, Procs. Eight CIM-Europe Annual Conference, Birmingham, UK, 1992, pp 249-260
5. Hannam, R.G., Computer Integrated Manufacturing: from concepts to realisation, 1st Ed, England: Addison Wesley Longman, 1996.
6. Hsu, Cheng, Manufacturing Information System, in Handbook of Design, Manufacturing, and Automation, Dorf, R.C. and Kusiak, A. (Eds.), Canada: John Wiley & Sons, 1994, pp. 737-766.
7. IFAC/IFIC Task Force on Architecture for Enterprise Integration, GERAM: Generalised Enterprise Reference Architecture and Methodology (http://www.cit.gu.edu.au/~bernus/taskforce/geram/versions/geram1-6-3/v1.6.3.html)