JOHN ATANASOFF SOCIETY OF AUTOMATICS AND INFORMATICSJOHN ATANASOFF SOCIETY OF AUTOMATICS AND INFORMATICS

International Conference

АUTOMATICS AND INFORMATICS’2014

PROCEEDINGS

Edited by Michail Petrov

Published by JOHN ATANASOFF SOCIETY

OF AUTOMATICS AND INFORMATICS

Bulgaria, Sofia, October 1-3, 2014

JOHN ATANASOFFCELEBRATION DAYS

JOHN ATANASOFF SOCIETY OF AUTOMATICS AND INFORMATICS

Secretariat Adress Bulgaria

1000 Sofia

108 Rakovsky str.

Tel. (+359 2) 987 61 69

Fax (+359 2) 987 61 69

e-mail: sai@infotel.bg

www.sai.infotel.bg

www.sai.bg

PROCEEDINGS: ISSN 1313-1850CD: ISSN 1313-1869

JOHN ATANASOFF SOCIETY OF AUTOMATICS AND INFORMATICS

International ConferenceAUTOMATICS AND INFORMATICS’2014

1-3 October 2014, Sofia, Bulgaria

©© I - 273

MODEL DRIVEN DEVELOPMENT OF MANUFACTURING EXECUTION SYSTEMS I. Antonova, I. Batchkova Dept. of Industrial Automation, University of Chemical Technology and Metallurgy, Bul. Kl. Ohridski 8, Sofia, Phone: +359 2 8163326 , -mail: {iskra.antonova, idilia}@uctm.edu Abstract: The main aim of the presented paper is to analyse the current approaches in the field of manufacturing execution systems and to suggest an approach for development of unified modeling framework and software process model (SPM) for their development. The proposed approach is based on the combined use of UML profile for system engineering SysML, ANSI/ISA-S95 standard and MES-ML modeling language. A case study applying the proposed approach is presented. Finally some conclusions are made. Keywords: MES, UML, SysML, ANSI/ISA-S95, MES-ML

INTRODUCTION

One of the trends in industry today is connected with using of Manufacturing Execution Systems (MES), which main advantages are associated with increasing of productivity, reducing the total time to perform customer orders and the time of product delivery, improving of product quality and customer services, reacting to unforeseeable events and facilitating the activity of the operating personnel. MES are combined, information, process-oriented, event and time activated software systems which represent the interface between the automation layer and ERP system and their functionalities [1]. An essential requirement to execute manufacturing processes is the requirement for horizontal integration of services in the MES layer and meanwhile, MES require the vertical integration to the systems and devices at the Control layer.

The rapid progress of computer technologies leads the possibilities for development of new methodologies, methods and approaches suitable for development of software and hardware in industrial systems. Some of the most promising and challenging approaches are these of Model Driven Development (MDD) and Model Driven Engineering (MDE) [2], where the systems are presented as models that conform to meta-models, and the model transformations are used to manipulate the various representations. Model Driven Architecture (MDA) [3] is a remarkable MDD initiative of Object Management Group (OMG), consisting in transformation of different platform independent models towards executable applications. In the core of MDA are the open standards - UML, MOF, XMI, etc. UML [4] does not specify a methodology for model driven software or system development but aims to provide an integrated modeling framework, covering structural, functional and behavior descriptions.

The currently available MES, based on traditional approaches, have monolithic architecture, are not flexible enough and have a limited scope. Due to the difficulties in their configuration and adaptation to specific needs, they do not meet the requirements of modern industry. The main trend to overcome these shortcomings is the use of new, advanced software process model (SPM) for MES development.

The main aim of the proposed paper is to suggest an approach for development of unified modeling framework and SPM for development of the important views of MES components and coupling their established modeling notations, based on the

combined use of UML profile - SysML and speziMES modeling technique.

The paper is organized in 5 parts. After the introduction, in part 2 the current trends in the field of manufacturing execution systems are discussed. Part 3 of the paper presents the main modeling approaches for MES. In part 4 the proposed SPM for MES development is described. A case study concerning an application of the suggested SPM for an industrial unit is presented in part 5. Finally some conclusions are made.

CURRENT TRENDS IN MES

MES engineering is an interdisciplinary challenge and requires cooperation between plant engineers, MES engineers, plant manager, production manager, IT engineers, solution developer, etc. The development and implementation of MES is a difficult task and the main challenges facing the field of MES are associated with the need of unified modeling framework and SPM that can be used to represent and execute such type of systems. To meet the challenges of modern enterprises, MES have to deal with certain difficulties which can be summarized as follows [5, 6]: a lot of knowledge about the underlying process is needed; more complex user interface; complex interfaces to upper and lower levels;

more complex rules for transformation of MES models into code than for embedded systems;

difficulty to cooperate co-workers from different domains; different requirements for technical and business processes from different co-workers;

need from different type of information coming from different kinds of processes and their interactions and time relations;

using of different modeling languages and standards model for plant structure;

using formal based modeling languages for MES modeling in order to check automatically the model;

using of different and limited notations which are not intuitively understandable, mostly complex and suited for interdisciplinary discussion.

Currently there are a lot of attempts to model MES systems based on different modeling techniques and standards. There are different approaches which may be summarized as follow: Temporal descriptions; Synchronization processes; Description of the behavior of processes;

I - 274

Structure processes; Representation; Graphical representation of rules and queries; Support for multiple points of view.

OVERVIEW OF MODELING APPROACHES FOR MES

The modeling approaches and techniques in the field of MES can be categorized in three basic groups – approaches using of standard terminology and reference models, provided by ANSI/ISA-S95 standard [1, 7, 8]; approaches using of MDE and MDA based approaches [2, 3, 4, 11, 12] and a set of approaches based on MES process modeling [13, 14, 15, 16].

Approaches based on ANSI/ISA-S95. ANSI/ISA-S95 [1] provides standard models, terminology and consistent set of concepts for defining the interfaces between enterprise business systems and corresponding manufacturing control systems. The ANSI/ISA-S95 series of standards separate the functions/activities of business process from production processes. The information is structured in UML models, which are the basis for the development of standard interfaces, MES systems and databases. An extended approach based on ANSI/ISA-S95 series of standards integrates ontologies in order to achieve interoperability of the internal and external systems. A core meta-ontology, based on the widely accepted standards for data integration (ANSI/ISA-S95 - ISO/IEC 62264) is used as a global data model. Domain knowledge is captured by domain ontologies, the meta-ontology is used as a guideline of building of domain models. Domain ontology is created using the data in the legacy data bases. Relational databases are rendered as ontologies dynamically, effectively creating a proxy file to the data in the database [9, 10].

MDE and MDA based approaches. The UML [4] as a general purpose modeling language and an open standard supports the MDE and MDA. It does not specify a methodology for software or system design but aims to provide an integrated modeling framework, covering structural, functional and behavior descriptions. The UML notations support the development of various diagrams that reflect different aspects of the system in order to capture the full complexity in the phases of detailed analysis and system design. The first attempts in this direction are connected with using of colored state chart diagrams, simulation activity diagrams and integration between UML block diagrams, state chart and collaboration diagrams with SDL (Specification Description Language) [15, 16] In the last years, there has been an increasing striving to use UML in control, automation, and industrial enterprise engineering. Different object-oriented approaches based on UML and its profile for system engineering SysML [17] for development of control systems, for different industry areas, are investigated. Detailed description and analysis of the different suggestions may be found in [19, 20]. The approach for MES modeling, based on SysML provides either an interaction between software components of MES and software and hardware components of the manufacturing system or possibilities for abstract and detailed modeling of MES activities.

Approaches using MES process modeling. The modeling techniques in this group can be categorized into three sub-groups: approaches for MES process modeling using traditional process modeling languages, approaches for MES process modeling using process decomposition and approaches for MES process modeling using hybrid techniques [15, 20].

In the group of approaches for MES process modeling using traditional process modeling languages, Petri nets and IDEF family and their extensions and transformation are included. The decomposition of MES processes is another promising

approach – agent-based decomposition, role-based decomposition and task-based decomposition. Various modeling languages and notations are used for modeling of interdisciplinary processes in MES, such as Business Process Modeling Notation (BPMN), Business Process Execution Language (BPEL), and SysML. The BPMN standard offers a number of structural elements, which are missing in other technologies and the information for MES specification is spared among multiple co-workers who have different view for technical and business process, know different details about the processes and have different requirements. BPMN is insufficient for modeling of complex systems but enables to present MES system as a collection of separate processes (which represent the MES functions), and so it is possible to model these processes separately under one large model [15, 21].

In [6] M. Witsch and B. Vogel-Heuser suggest a modeling framework that integrates the different important views for MES specification by coupling their established modeling notations. The main contribution of this framework is the integration of a technical system model, a production process model and an MES functional model with their interconnections and dependencies by a formal link model. MES-ML proposes a graphical MES modeling language. The technical system model represents the static technical systems that perform MES functions. This model either could be on the abstraction level of an entire plant, or detailed to atomic function units depending on the needs. The second model is the production process model, which represents the manufacturing systems in MES. The authors propose to use UML activity diagram or flow chart as the modeling notation.

SHORTLY DESCRIPTION OF SUGGESTED APPROACH

The modeling approaches in the area of MES described above are not sufficient to satisfy the users and system requirements and for development of the important views of MES components. MES are not flexible enough, they have a limited scope and monolithic architecture and may not be reused in new manufacturing applications.

The success of UML and SysML in unifying many different object-oriented approaches and the graphical modeling language support of MES-ML to the required engineering process of MES led to the idea of their combined use for development of unified modeling framework and software process model for MES components. Supplementary a separation of functions/activities of business processes from production processes by using of ANSI/ISA-S95 standard models and terminology is undertaken.

The proposed model-driven approach is shown in fig.1. The basic cycles for task flow development process included in the proposed methodology, such as – requirements analysis, system functional analysis, architectural design and hardware/software specification are the stages from Harmony SE based methodology for software development processes, described in details in [18, 19]. The suggested approach uses different kinds of ANSI/ISA S95 standard models, modeled through SysML modeling constructs.

The first step in the development of MES using the proposed approach is creating on UML/SysML profile extended with new predefined stereotypes based on MES-ML modeling notations. The development process of the MES requirements and performance analysis function is divided into two main steps – define the concrete customer functional requirements and selection of MES functions represented through the Requirements Diagram and Requirements table; and MES functionality definition using Use Case Diagram and Block Definition Diagram.

I - 275

Fig.1: Approach for development of MES

The Architecture design is the next stage in the proposed methodology and includes the design of ANSI/ISA S95 models using Block Definition Diagram which are customized through the MES-ML stereotypes. The different views of MES system - technical system model, a production process model and MES functional model are modeled with Block Definition Diagram, Internal Block Diagram and Activity Diagram customized through the MES-ML stereotypes.

CASE STUDY

To show some of the advantages of the suggested approach, a real industrial Pick and Place Unit (PPU) constructed by Technical University Munich [22] is used (fig.2). The PPU consists of a stack (1), working as a work piece input storage, a conveyor (2) working as a work piece output storage, a stamp (3) for stamping work pieces and a crane (4) for transporting work pieces by picking and placing them between three working positions – for black, white and metallic work pieces. The PPU may work in various different manifestations called Scenarios.

Fig.2: Pick and Place unit

Fig.3 shows the MES Specification Model of Pick and Place unit according the suggested approach. MES-ML provides three models in order to present three different views of the modeled system - technical system model, production process model and MES functional model with their interconnections. The MES-ML model of the PPU is presented through BDD (fig.3-1). The internal structure of the each model is presented respectively as BDD for technical system model (fig.3-4), activity diagrams for production process model (fig.3-2) and MES functional model (fig.3-3).

The technical system models for PPU present the physical structure of the plant in detailed level, which is composed by the following equipment: crane, stack, stamp, ramp and conveyor, different types of sensors, valves, motors, micro

switches, reed switches etc. The Unit’s signals are characterized with attributes. Fig.3-2 shows the production process model modeled as activity diagram. The process of picking and placing can been described shortly with the following activities – initialization of crane and stack; choosing of scenario; turn crane to stack, provide WP, wait, pick up WP, retract stack, turn crane to ramp and put down WP. According MES-ML, special modeling notations are used to present this model. In order to specify these modeling notations new stereotypes are defined. <<Timer Event>> MES-ML modeling notation is assigned as a stereotype of activity named “Wait”, <<Automation Subprocess>> of “Initialization of crane and stack” activity and <<Production Process Activity>> for the rest of used activities.

A new customer order is displayed with MES Functional Model in fig.3-3. The diagram is separated in three main parts with swim lines representing respectively the activities for ERP system, MES system and Process Control System. The activities in this diagram can be decomposed in sub-process activities in separated diagrams. For depicting of activities for this model, stereotypes named <<Sub Process>>, <<Manual Task>>, <<Parallel Instances>>, <<MES Activity>>, <<Data Store>>, <<Start>> and <<End>> are used.

CONCLUSIONS The analysis of the proposed object-oriented approach and UML in the development of control systems indicates that it is applicable to the modeling of systems in various industrial fields. The possibilities to expand object-oriented approaches and UML allows to use advantages of an approach and adds properties and advantages from other languages, frameworks and approaches to achieve specific custom solutions.

The paper presents an approach for development of unified modeling framework and SPM for modeling of important views of MES and their components. The UML/SysML profile is extended and at the same time restricted with different stereotypes from MES-ML and ANSI/ISA S95 standard models. The benefits of using this approach are summarized as follow: development of open, flexible and reusable components, faster specification of MES functionality, improvement of capabilities for requirements definition, etc.

The future research activities are concentrated mainly on applying the suggested approach to continuous and batch systems as well as application of the suggested approach in industry.

REFERENCES

1. ANSI/ISA-S95.00.03-2005, Enterprise-Control System Integration Part 3: Models of Manufacturing Operations Management. ISA, Raleigh, North Carolina, USA. 2. Kent, S., Model Driven Engineering. In Proceedings of IFM, 2002, LNCS 2335, Springer. 3. OMG-MDA. MDA Guide version 1.0.1. OMG document omg/2003-06-01, 2003. 4. OMG-UML, 2010. OMG. http://www.omg.org/ 5. Witsch M., Vogel-Heuser B., Towards a Formal Specification Framework for Manufacturing Execution Systems, IEEE Transactions on industrial informatics, VOL. 8, NO. 2, MAY 2012, pp. 311 – 320. 6. Witsch M., Vogel-Heuser B., Modeling of Manufacturing Execution Systems: an Interdisciplinary Challenge, 15th IEEE International Conference on Emerging Technologies and Factory Automation, 2010.

JOHN ATANASOFF SOCIETY OF AUTOMATICS AND INFORMATICS

International ConferenceAUTOMATICS AND INFORMATICS’2014

1-3 October 2014, Sofia, Bulgaria

©©I - 276

7. ANSI/ISA-S95.00.01-2000, Enterprise-Control System Integration Part 1: Models and Terminology, ISA, Raleigh, North Carolina, USA. 8. ANSI/ISA-S95.00.02-2001, Enterprise-Control System Integration Part 2: Object model attributes. ISA, Raleigh, North Carolina, USA. 9. Dobrev M., Gocheva D., Batchkova I., An ontological approach for planning and scheduling in primary steel production, In Proceedings of the 4-th International IEEE Conference on Intelligent Systems, Vol.1, pp.6-14: 6-19, Varna, Bulgaria, September 6-8, 2008. 10. Gocheva D., I. Batchkova, Ontology Based Approach for Achieving Interoperability of Manufacturing Execution Systems, Proceeding of the Anniversary Scientific Conference with International Participation “40 Years Department of Industrial Automation”, 18 March 2011, UCTM, Sofia, pp.153 – 158. 11. Pietrac L., Leleve A., Henry S., On the use of SysML for Manufacturing Execution System design, IEEE 16th Conference on Emerging Technologies & Factory Automation (ETFA), 2011, 5-9 Sept., Toulouse, pp.1 – 8. 12. Kohler H.J., Nickel U., Niere J., Zundorf A., Integrating UML diagrams for production control systems. In Software Engineering, 2000. Proceedings of the 2000 International Conference on, pp. 241÷251. 13. Witsch M., Vogel-Heuser B., Formal MES Modeling Framework – Integration of Different Views, Preprints of the 18th IFAC World Congress Milano (Italy) August 28 - September 2, 2011, pp. 14109 – 14114.

14. Vogel-Heuser B., Erhöhte Verfügbarkeit und transparente Produktion, 2011, ISBN:978-3-86219-178-9. 15. Mili H., Tremblay G., Jaoude G.B., Lefebvre E., Elabed L., Boussaidi G.E., Business process modeling languages: Sorting through the alphabet soup. ACM Computing Surveys (CSUR), 43(1):4, 2010. 16. Fengler O., Fengler W., Duridanova V., Modeling of complex automation systems using colored state charts. In Robotics and Automation, 2002. Proceedings. ICRA'02. IEEE International Conference on, volume 2, pp.1901÷1906, 2002. 17. OMG-SysML, 2006. The OMG Systems Modeling Language, http://omgsysml.org/index.htm, May. 18. Batchkova I., Antonova I., Improving the Software Development Life Cycle in Process Control using UML/SysML, Preprints of the 18th IFAC World Congress Milano (Italy) 28.08 – 02.09, 2011, pp. 14133-14138. 19. Antonova I., Batchkova I., Development of Multi-Agent Control Systems using UML/SysML, book capter, pp.67-90, InTech, 2011, ISBN: 978-953-307-174-9. 20. Ferrarini L., Fogliazza G., Mirandola G., Veber C., Metamodeling techniques applied to the design of recongurable control applications. EURASIP Journal on Embedded Systems, 2008:8, 2008. 21. Michalik P., Štofa J., Zolotová I., The use of BPMN for modelling the MES level in information and control systems, Quality innovation prosperity XVII/1 – 2013, ISSN 1335-1745, pp.39-47. 22. Institute of Automation and Information Systems. The Pick and Place Unit – Demonstrator for Evolution in Industrial Plant Automation, 2013. http://www.ais.mw.tum.de/ppu/.

1

2

3

4

Fig.3: Illustration of speziMES model for Pick and Place unit