APPLICATION OF ISA95 DATA MODELS IN MANUFACTURING ...€¦ · Manufacturing Execution Systems (MES) are service-oriented interfaces that join the word of business transactions with

STUDIA INFORMATICA 2016

Volume 37 Number 4A (127)

Udo WOZAR, Huseyin ERDOGAN,

Conti Temic microelectronic GmbH, Ingolstadt, Germany

Rafal CUPEK, Szymon ZIEMEK

Silesian University of Technology, Institute of Informatics, Gliwice, Poland

APPLICATION OF ISA95 DATA MODELS IN MANUFACTURING

EXECUTION SYSTEMS FOR LEAN PRODUCTION

Summary. Manufacturing Execution Systems (MES) are service-oriented

interfaces that join the word of business transactions with the world of production

systems. Nowadays IT systems have to provide very detailed information that is

related to an underlying production process and also to actual product. There are a few

emerging business models that require accurate and timely production data. This

document examines two approaches to database architecture that can be used in

Manufacturing Execution Systems (MES). It focuses on the support of the LEAN

business model. The main research goal is to support the flexible access to production

data, but the efficiency of database is also very important factor. Authors compare the

classical relational database model with the object-oriented one. Considered use cases

include the Oracle DB and Oracle Objects applications for MES. Presented object

oriented approach follows the ISA95 standard. The practical use cases are based on

the production of electronic devices carried out by the company Continental

Ingolstadt. Although object oriented databases are not well accepted by the industry

due to their low efficiency, the authors show that in the case of LEAN production, the

database system based object-oriented models can be far more convenient than a

classical relational database. The main benefits are more flexible data model and

highly adjustable MES that can follow changes in the underlying production system.

By the case of LEAN manufacturing, authors show that the flexible object oriented

database is more efficient solution comparing to the relational database. Moreover

such an approach can help to avoid well known big data problems that are common in

classical MES.

Keywords: MES; Manufacturing Execution System; LEAN production; ISA95

34 U. Wozar, H. Erdogan, R. Cupek, Sz. Ziemek

ELASTYCZNE MODELE DANYCH ZGODNE ZE STANDARDEM

ISA95 DLA SYSTEMÓW REALIZACJI PRODUKCJI KLASY MES

Streszczenie. Systemy realizacji produkcji klasy MES (Manufacturing Execution

Systems) dostarczają zorientowanych na usługi interfejsów informatycznych, które

łączą świat transakcji biznesowych ze światem systemów produkcyjnych.

Współczesne systemy informatyczne stosowane w przemyśle przekazują bardzo

szczegółowych danych, które opisują zarówno samą realizację procesu

produkcyjnego, jak i właściwości wytwarzanych produktów. Jednocześnie pojawiają

się nowe modele biznesowe, które wymagają dokładnych i aktualnych danych

produkcyjnych. Niniejszy dokument porównuje dwa podejścia do architektury baz

danych, które mają być wykorzystane w systemach realizacji produkcji klasy MES

współpracujących z systemami biznesowymi bazującymi na modelu LEAN. Głów-

nym problemem jest zapewnienie elastycznego dostępu do danych produkcyjnych,

tym niemniej wydajność systemu i szybki dostęp do gromadzonych danych są także

istotnymi czynnikami wpływającymi na wybór architektury systemu. Rozpatrywane

przypadki użycia obejmują zastosowanie rozwiązań Oracle DB i Oracle Objects

w systemach MES. Podejście obiektowe zilustrowano na rozwiązaniach opartych na

standardzie ISA 95. Przedstawiona analiza została zilustrowana przez prezentację

przypadków użycia zaczerpniętych z obszaru produkcji zaawansowanej elektroniki

samochodowej realizowanej przez firmę Continental Ingolstadt. Zastosowanie

obiektowego modelu bazy danych w przemyśle jest w chwili obecnej bardzo

ograniczone ze względu na niską wydajność dostępnych rozwiązań. Autorzy

wykazują, że w przypadku systemów MES wspierających model biznesowy „LEAN

production” zastosowanie podejścia obiektowego pozwala na elastyczne dopasowanie

interfejsu systemu bazodanowego zarówno do modelu biznesowego, jak i do formatu

źródeł danych na poziomie systemu produkcyjnego. Ponadto, podejście obiektowe

umożliwia ograniczenie nadmiernego rozrostu wolumenu danych (Big Data problem)

Słowa kluczowe: MES; Manufacturing Execution System; LEAN; ISA95

1. Introduction

Nowadays control systems provide very detailed information that is related to an

underlying production process [1]. This information is further used by Business Intelligence

systems, which are localised on the Enterprise level. Manufacturing Execution Systems

(MES) are service-oriented interfaces that joins these two worlds. MES offer sets of

interfaces that connect the physical world of production with the virtual world of business

transactions. The classical MES are defined by their static hierarchy, which makes them very

Application of ISA95 Data Models in Manufacturing… 35

difficult to modify. This issue forces the research on a new approach to the architecture of an

MES that will support the huge stream of information that is exchanged between business and

production systems [2].

The commonly accepted definitions of MES functions and standards can be found in the

sets of documents that are managed by MESA International (Manufacturing Enterprise

Solutions Association), which is a worldwide not-for-profit community of manufacturing

companies, information technology hardware and software suppliers, system integrators,

consulting service providers, analysts, editors, academics and students [3]. According to

MESA’s definition, MES support production in the following activities: job scheduling,

launching the orders, responding to random events, adjusting production plans, tracing

product genealogy, managing production quality and managing maintenance activities. The

above-mentioned areas are systematised into conceptual and functional models which are

described the in ANSI/ISA-95 (IEC/ISO 62264) standard that is the international standard for

the integration of enterprise and control systems [4].

ISA-95 defines the MES data structure and MES services that are related to

manufacturing operations: defining the product, forecasting production, managing production

capability and evaluating production performances. ISA-95 consists of models and

terminology and describes the information that is exchanged between the systems for sales,

finance and logistics and the systems for production, maintenance and quality. This

information is structured in the form of UML (Unified Modelling Language) models, which

are the basis for the development of standard interfaces between ERP (Enterprise Resource

Planning) and MES systems.

Nowadays production series become shorter and MES has to operate with mass-

customised production. Moreover production systems become more autonomous and MES

architectures must follow these changes [5]. The underlying database system efficiency is not

such a problem like data structure flexibility and ability to adjust changes in production

system environment[6]. To ensure enough flexibility ISA-95 models are built on the object-

oriented paradigm used to define the interface between control systems and business

application. ISA-95 defines the services that are required for the manufacturing support that is

designed according to the object-oriented model [4]. These services are not only based on

information exchange but also on the aggregated data or the history of the realisation of the

process that has to be managed by database systems.

Another issue is communication with database in MES. The desktop clients are due their

low mobility obsolete. New Object oriented standards and services has to be implemented. A

good example is OData protocol that can reduce the volume of information exchanged

between client and servers [7]. Instead of processing big data sets[8] by client the information


processing can be more effectively done by servers but new client server paradigm has to be

established – interactive services replace exhausting data queries. OData becomes that foe

web applications what SQL for Database systems is. As is shown in [9], the growing amount

of data that is processed by industry has created the necessity to create the right approach and

tools to convert the data into useful, actionable information. As a result, a huge amount of

data has to be stored in the database systems that support ISA95 standard.

The main goal of this paper is to present the comparative research study on relational

database and object database architectures used for ISA95 based MES. The use case analysis

was based on the example related to managing materials that are used in the production of

electronic devices. The authors compare two different approaches including the classical data

representation in a relational database and the object-oriented data representation that directly

reflects the ISA-95 models and is realized as an object-oriented database. The authors applied

for MES database system the classical Oracle and Oracle Objects. The advantages and

disadvantages of both solutions including comparison in efficiency, size of data that is stored

and the time that is required to perform operations on the data are presented in the

experimental part of this paper.

The paper is organised as follows: the discussion on applying the LEAN manufacturing

approach not only to save physical resources but also to make their representation in database

more close to its object based nature is presented in chapter one. Because the costs of

database application are also related to cost of its maintenance and time necessary for changes

adoption, the dependency between Lean manufacturing and Lean Database, is considered

during selection of data base model as presented in the chapter two. Chapter three presents a

practical use cases of MES databases and the comparison of the efficiency. The conclusions

are presented in chapter four.

2. Database model

“A model is a useful representation of a specific situation or thing. Models are useful

because they describe or mimic reality without dealing with every detail of it. They typically

help people analyse a situation by combining a framework’s ideas with information about the

specific situation being studied.” [10]. The biggest challenge today is not only to make the

data "LEAN", but also to allow enterprises and customers to speak the same language, which

means reducing the information exchange time and complexity for this process. It is very

important that each side can quickly understand the information that is received. One


component that can complement the architecture of a system with a complete, comprehensive

and easy to exchange data model is the ANSI/ISA-95 standard.

In general, ISA-95 defines the MES data structure and MES activities that are related to

manufacturing operations: defining a product, forecasting production, managing production

capability and evaluating production performances. ISA-95 consists of models and

terminology and describes the information that is exchanged between systems for sales,

finance and logistics with the systems for production, maintenance and quality. This

information is structured in the form of UML (Unified Modelling Language) models, which

are the basis for the development of standard interfaces between ERP (Enterprise Resource

Planning) and MES systems [11].

What is interesting is that the second part of ISA-95 introduces one common data model

that includes all of the possible functionality that may be needed in enterprises. Enterprises

can follow this model to ensure data consistency not only within the enterprise but even in a

network of companies. This model is very easy to introduce. It consists of smaller models that

are defined as UML diagrams. Each class from the diagram is widely described as the general

idea for a given class and all of its attributes. We can assume that ISA-95 is an object-

oriented model, so the conclusion is that maybe we should try to use an object-oriented

database. Because the structure of ISA-95 models is defined in an object-oriented way, it

seems that the most convenient physical representation should also be a database system that

is organised in an object-oriented way.

An Oracle Database, which is commonly referred to as Oracle RDBMS or simply as

Oracle, is an object-relational database management system that is produced and marketed by

the Oracle Corporation. PL/SQL, which is the Oracle procedural extension of SQL, is a

portable, high-performance transaction-processing language[12]. With objects, users can

break complex problems down into easily understandable object methods within object types.

Using object types in a PL/SQL block subprogram or package is a two-step process that

involves:

defining the object types using the SQL statement CREATE TYPE. After an object type

is defined, users can use it in any PL/SQL block, subprogram or package;

declaring the variables for the user-defined object types in PL/SQL. After the variables are

defined, users can refer to the proper methods that were declared for a given variable.

New object types can be created from any built-in database types and any previously

created object types, object references and collection types. Object types can work with

complex data such as images, audio and video. An Oracle Database stores the metadata for

user-defined types in a schema that is available to SQL, PL/SQ, Java and other languages.


This is one big advantage of the object model – users have the data and its description in one

location [7]. Using objects has three very important advantages:

encapsulation – only the necessary data and methods are available outside an object and

the rest is hidden in the object structure. Users have a plain interface without any

redundant data or methods and can easily use it;

efficiency – all object members and methods are defined inside the database. This ensures

that all of the structures are available and ready to use at any time. Developers do not need

to create object definitions in their applications because they can use the objects that are

already defined;

simplified logical structure – it is more convenient and natural to organise a data structure

using objects (e.g., a car object that includes the objects of its components) that represent

real-life things.

The final point is about the architecture of a database. In general, the two most popular

architectures are central architecture and distributed architecture. An illustration of the

concept of a centralised MES architecture is presented in Fig. 1. A central database governs

all production data. Data access can be supported by stored procedures that are executed on

the database server, which allows all of the business logic to be placed inside the database.

Such a solution has the following advantages: better performance – all queries can be

executed faster; a programmer does not need to implement functionality inside applications

and the data is always consistent.

Fig. 1. An example of centralized MES database architecture

Rys. 1. Przykład scentralizowanej architektury systemu klasy MES


The main drawback of a centralised database architecture is the difficulty and high risks

that are connected with any change in its structure. Production process modifications are

relatively rare in mass production systems and classical architectures are used because of their

efficiency. Nowadays, manufacturing is extremely competitive in most industrial sectors and

the financial margins that differentiate between success and failure are very tight [13,14].

This means that it is very difficult to fill volatile market demands without an effective and

flexible Manufacturing Execution System.

A flexible MES should meet objectives that are connected with reconfigurable machinery

and robots and should be capable of easy adaptation to production requirement changes,

which are very characteristic in contemporary production models. Specific components of

flexible MES systems should be decentralised, autonomous and supported by service-oriented

communication protocols. A database should provide the interfaces that are necessary to

communicate with the different workstations that realise the individualised paths of

production flow. This requirement should be extended to production lines and should reflect

flexible configurations as well as changes in production scenarios. Flexibility at the MES

level has to be supported by a flexible information structure and by an extensible and object-

oriented underlying database system.

Another database problem is availability. Data should be stored in one place only and this

place should be available at all times. To ensure that data is safe and will always be available,

cyclical backups must be made. The main issue in distributed database architecture is

synchronising the data between databases. The next important feature is the ability to store

the metadata that describes the location and localisation of the data that is processed by an

MES system. Metadata can be stored in an additional separate place and managed by

additional agent components that support the MES system in database access. The idea of the

distributed MES database that is described above is presented in Fig. 2.

The biggest advantage of a distributed approach is that data can be distributed and backed

up in different locations. Distributed and redundant data are safer (the main part of data will

still be available even when parts of the system fail) but with an additional cost in system

performance and the need for additional data management functionalities.

Communication between databases and other stations can be done using the same

protocols that are used in centralised architecture. In this case, part of functionality can be

implemented as stored procedures in databases. The rest of the system logic can be

implemented by MES agents. This can cause distributed architectures to work more slowly

than centralised ones. Another problem can occur with data consistency since functionality

and data structures are defined in different locations. To avoid this problem, an MES should

follow a clear and unequivocal data structure and unified data representation. According to


the ISA-95 standard, an application data model seems one of the more promising solutions to

fulfil this requirement in a distributed MES.

Fig. 2. An example of distributed MES database architecture

Rys. 2. Przykład rozproszonej architektury systemu klasy MES

3. Models comparison of use cases with electonic devices production

Two similar applications were prepared. The first one was prepared using a usual

approach (using a relational database and an often-used approach to model the data) and the

second one was prepared using the ISA-95 standard implemented as an Oracle Objects model.

The application was prepared as a database- centralized system, which means that all of the

functionality is included in a centralised database as storage procedures or as member

functions in objects. It was done this way for two reasons. It allows programmers to change

the technologies that are being used (in GUI applications) without the need for changes in

functionality and also it allows some database tests to be performed without any influence of

the network or connections. A short comparison between both applications and their data

models done is presented in this chapter. For transparency and to provide a better

understanding, it will be prepared using only a small part of the "Production operations

management information" model, which is defined in the first part of the ISA-95 standard

(Models and Terminology) [15] – Product definition.


To make this comparison more understandable, we can describe a real example of how

the Product definition functionality works. The Short Range Radar 200 is used as the example

device. The Short Range Radar 200 is a device that is part of the safety features in a car. As

we can see on the Fig. 3, the SRR200 is simply a small black box with mounting wings and a

connector.

Fig. 3. Exploded Short Range Radar, model SRR 200

Rys. 3. Schemat budowy radaru krótkiego zasięgu, model SRR 200

During the production process, it is necessary to know the structure of the device and all

of the sub-materials. In the example that is given, we can see that the SRR 200 consists of

five sub-materials: Housing, RF Antenna board, EMC Shield, PCB and Back Cover.

Precisely that information (of course with more details) should be stored in the product

definition application. In addition, we can assume that the SRR200 consists of other

materials, and that it can also be defined as a sub-material for other materials (e.g., a car). If

necessary, this information can also be stored in the product definition. Now the question is:

How to store and manage this data?

Relational databases are well known and deeply explored mechanisms. In this approach,

programmers must think in terms of tables, relations, etc. The authors want to show the bad

habits of programmers, which cause the project to become huge and inconvenient for

providing any additional functionality. This approach leads us to "big data" that is not

expected and that is difficult to apply to data mining methods with a good performance. A

typical data model in this case looks like the model in Figure 4 (only the most important part

of the whole model shown) here.

The materials table is the most important table. As we can see, material has its own id,

description and some other static properties such as product type, location or additional

properties. Relations between materials are made by using a material_use table in which a

parent and its child can be defined. This model should be surrounded with some functionality,


which is done using storage procedures that are grouped within a package. After such a model

was defined, some tests were performed, which will be described later.

Fig. 4. Database structure for the first approach

Rys. 4. Struktura bazy danych dla pierwszego podejścia

An Object-oriented database using ISA-95 approach is slightly different. Users do not

have to think in the term of tables, but rather in term of "real objects". This means that users

don't need to worry about the database structure, but only about what kinds of properties he

wants to assign to a given material. Tables are only created to store the list of objects.

Moreover, Oracle objects themselves are nothing more than a kind of "interface" with a

relational engine behind them. So why have the authors decided to focus on objects? The first

point is that objects are very convenient and much easier to understand, even for

programmers who are working with them for the first time. The second point is that objects

can be implemented using the ISA-95 standard. The ANSI/ISA-95 is the standard that leads

us to less data. Why? Because the data that is required in the models are minimised in most

cases to just two fields – Id and description. If some other fields are needed, they are added as

dynamically assigned properties. The attributes of the model are described in detail in the

second part of the standard [16]. We can say that the idea of using Oracle objects together

with the ISA-95 should lead us along a common path with the LEAN methodology [17,18].

In order to check the accuracy of the sentence above, an object-oriented model was

prepared. The base for the model was a part of the Material model, which is defined in the

second part of the ISA-95 standard (Fig. 5). Four classes are defined: Material class, Material

definition, Material class property and Material definition property.


CREATE OR REPLACE TYPE material_t AS OBJECT

(

mat_NO VARCHAR2(100),

description CLOB,

CONSTRUCTOR FUNCTION material_t(I_mat_NO VARCHAR2, I_description

VARCHAR2) RETURN SELF AS RESULT,

MEMBER FUNCTION get_property_value(I_prop_name VARCHAR2)

RETURN VARCHAR2,

MEMBER PROCEDURE set_property_value(I_prop_name VARCHAR2, I_prop_value

VARCHAR2),

MEMBER PROCEDURE add_to_class_list,

MEMBER PROCEDURE add_to_definition_list

MEMBER PROCEDURE remove_from_list );

Fig. 5. Material definition part of the Material model of ISA-95

Rys. 5. Diagram definicji materiału według modelu w standardzie ISA-95

Using the standard attributes and connections that are defined for each of the four classes

[16], a user can define the structures of the objects in the database (Fig. 6).

Fig. 6. Definition of material object type

Rys. 6. Implementacja struktur reprezentujących informacje o materiale


Three tables created to store the list of the material class, material definition and its

material properties:

MATERIAL_CLASS_TAB

MATERIAL_DEFINITION_TAB

MATERIAL_PROPERTIES_TAB.

With such a defined model, the functionality and data are quite similar to the application

from the first approach and a comparison between them can be done. Working with this

model is just like working with objects in C++ – users simply execute the object’s member

functions. The comparison was done based on several fields: data size and data redundancy,

model flexibility (possibility to extend the model) and performance. The results are presented

in Table 1. The performance results were obtained as a time measurement for the following

operations: insert, update and delete. Fifty of the following tests were done for each of those

operations:

− time for the operation for 1 record

− time for the operation for 100 consecutive records

− time for the operation for 1000 consecutive records

− time for the operation for 10000 consecutive records.

Each total time was divided by the number of records in the operation and the average

value for the operation for one record was taken. The following figures present the time

results for each of the tests as the average time needed for one operation. In the figures above

we can see that the times for the insert and update operations are much better for relational

approach. On the other hand, the times for the delete operations are slightly better for the

object approach. Only a part of the ISA-95 model was tested. The whole model and all its

properties would certainly take more physical disk space, but they would also provide more

knowledge about the model and the metadata, which could be helpful when the data would be

analysed (data mining).


Table 1

Results of the comparison

Approach 1:

Relational database

Approach 2:

Oracle Objects with ISA-95

Ph

ysi

cal

data

siz

e

Tables’ sizes after inserting 10000 records:

3 211 264 Bytes

Huge size of the MATERIALS table. A

large amount of data redundancy

because each material has own list of

properties, even within the same

material class.

1 835 008 Bytes

Less data, the main properties are

assigned to the material class. Materials

contain only unique properties and

property with class id.

Mate

rial

pro

per

ties

Statically defined

Dynamically defined

(user can simply add a new property to

the list)

Exte

nd

ing t

he

mod

el

Model is static. To extend it, the

database structure needs to be changed.

ISA-95 provides a very flexible model.

To extend it, a user can just add a new

property (which consists of Id,

description, value and measure unit). In

the event of a strong need, a user can also

add a new field to the previously defined

object types.

Per

form

an

ce

test

Results are the average values for one operation taken from the results for 1, 100,

1000 and 10000 operations:

INSERT: 0.0002258 s

UPDATE: 0.0024325 s

DELETE: 0.0017900 s

INSERT: 0.0010298 s

UPDATE: 0.0056245 s

DELETE: 0.0012303 s

Com

men

ts This approach is good if the company

sells only one product during its whole

life, but is not good for corporations in

which products evolve. However,

unfortunately, it is still often used.

Slightly slower (performance), but very

flexible. A good way to implement a data

model that is easy to extend. ISA-95

ensures consistency of the data.

4. Conclusions

In this document, the authors have presented the LEAN MES approach, which can be

realised in practice using the Oracle Objects database together with the model definitions

given by the ANSI/ISA-95 standard. As shown an curtail issue for ISA-95 implementation is


the flexible database approach that allow for LEAN implementation of ISA-95 models. An

object-oriented application is not a high performance solution especially in the case of the

time necessary for the realisation of INSERT and UPDATE operations.

Fig. 7. Time results for the insert operation

Rys. 7. Wyniki pomiarów dla operacji wstawiania nowych rekordów

Fig. 8. Time results for the update operation

Rys. 8. Wyniki pomiarów dla operacji aktualizacji rekordów

On the other hand the object oriented database is represented in more compact way and

takes up less disk space than its relational equivalent. The main advantage of the object

oriented database model is its flexibility. In the case when time for database operation is not

critical (like for example in the case of MES for short series production) the cost of big

efforts related to data model adjustment in the case relational database can be far more


unfavourable than benefits from fast data access. ISA95 and its object oriented

implementation offer ready to extend data models without the need to change the database

structure. To successfully implement IS95 model in object oriented database it is necessary to

make a fusion between an object-oriented database engine and ISA95 models. Such a fusion

will lead to the unified MES information flow model and results in a flexible Manufacturing

Execution System that supports both the business and production levels.

Fig. 9. Time results for the delete operation

Rys. 9. Wyniki pomiarów dla operacji usuwania rekordów

Presented research results show the relation between selected database type and the

efficiency of the data access. As shown by the experiments, in the classical relational database

is slightly more efficient than object oriented one. On the other hand, even in the case of

Oracle objects, the particular operation time is measured in fractions of a second. The main

benefit from object oriented architecture is its flexibility. It means that system designer will

spend less time on changing the underplaying database structure, in the case when changes in

MES are necessary. Since, MES do not work as real-time systems, authors conclude that

benefits from system flexibility are more significant than slightly longer database access time.

Acknowledgements

This work was supported by the European Union from the FP7-PEOPLE-2013-IAPP

AutoUniMo project “Automotive Production Engineering Unified Perspective based on Data

Mining Methods and Virtual Factory Model” (grant agreement no: 612207) and research

work financed from funds for science in years 2016-2017 allocated to an international co-

financed project (grant agreement no: 3491/7.PR/15/2016/2).


BIBLIOGRAPHY

1. Shipp S.S., Gupta N., Lal B., Scott J.A., Weber C.L., Finnin M.S., Blake M., Newsome

S., Thomas S.: Emerging global trends in advanced manufacturing, DTIC Document,

2012.

2. Blanc P., Demongodin I., Castagna P.: A holonic approach for manufacturing execution

system design: An industrial application. Engineering Applications of Artificial

Intelligence. Volume 21, Issue 3, April 2008, p. 315÷330.

3. http://www.mesa.org

4. http://www.isa-95.com

5. Cupek R., Erdogan H., Huczała Ł., Wozar U., Ziębiński A.: Agent Based Quality

Management in Lean Manufacturing. In Computational Collective Intelligence,

Springer 2015, p. 89÷100.

6. Do N.: Integration of engineering change objects in product data management databases

to support engineering change analysis. Comput. Ind. 73, 2015, p. 69÷81. DOI:

10.1016/j.compind.2015.08.002.

7. Cupek R., Huczała Ł.: OData for service-oriented business applications: Comparative

analysis of communication technologies for flexible Service-Oriented IT architectures.

IEEE International Conference on Industrial Technology (ICIT), 2015, p. 1538÷1543.

8. Lee J. et al: Recent advances and trends in predictive manufacturing systems in big data

environment. Manufacturing Letters 1.1 (2013), p. 38÷41.

9. Choudhary A.K., Harding J.A., Tiwari M.K.: Data mining in manufacturing: a review

based on the kind of knowledge. Journal of Intelligent Manufacturing, October 2009,

Volume 20, Issue 5, p. 501÷521.

10. Anvari A., Ismail Y., Hojjati S.M.H.: A Study on Total Quality Management and Lean

Manufacturing: Through Lean Thinking Approach, 2011.

11. Alter S.: Information Systems. Pearson, Upper Saddle River 2002.

12. Oracle Database documentation Library. 12c Release 1 (12.1).

http://docs.oracle.com/database/121/index.htm. 2014.

13. Bleeker A.: Prozessoptimierung in der Produktion heute. Productivity Management,

März 2012, p. 24÷26.

14. Charantimath P. M.: Total Quality Management. Pearson, 2011.

15. ANSI/ISA-95.00.01-2010. Enterprise-Control System Integration – Part 1: Models and

Terminology. Approved 13 May.

16. ANSI/ISA-95.00.02-2010. Enterprise-Control System Integration – Part 2: Object

Model Attributes. Approved 13 May 2010.


17. Gerberich T.: Lean oder MES in der Automobilzuliefererindustrie. Gabler, Chemnitz

2010.

18. Shang Gao S.P.L.: Lean Construction Management – The Toyota Way. Springer,

Singapore 2014.

Omówienie

W niniejszym dokumencie, autorzy przedstawili przykład zastosowanie obiektowego

modelu bazy danych w połączeniu ze standardem ISA95 w systemach klasy MES.

Zaczerpnięte z produkcji elementów elektronicznych przykłady ilustrują problem efek-

tywnego wykorzystania systemów realizacji produkcji do wsparcia produkcji krótkoseryjnej

z uwzględnieniem wymagań modelu biznesowego „LEAN production”. Przeprowadzone

badania obejmowały wykorzystanie bazy danych Oracle Obiects do implementacji modeli

danych definiowanych przez standard ISA-95. Przeprowadzono analizę wydajności modelu

i porównano go z analogicznym systemem zrealizowanym w klasycznej architekturze

relacyjnej bazy danych. Wyniki praktyczne wskazują, że bezpośrednia implementacja mode-

lu ISA-95 w obiektowej bazie danych jest mniej efektywna niż analogiczna funkcjonalność

uzyskana w technologii relacyjnej. Różnice widoczne były szczególnie w przypadku czasu

niezbędnego do realizacji operacji INSERT i UPDATE.

Obiektowy model bazy danych ma bardziej zwartą strukturę i przechowywane dane zaj-

mują mniej miejsca na dysku niż w przypadku systemu opartego na bazie relacyjnej. Główną

zaletą modelu obiektowego jest jego elastyczność. W przypadku gdy czas operacji bazo-

danowych nie jest krytyczny dla produkcji (jak ma to miejsce na przykład w systemach MES

wykorzystywanych w produkcji krótkoseryjnej), czas i koszt operacji związanych z dostoso-

waniem modelu danych do profilu produkcji niezbędne dla baz relacyjnych mogą być o wiele

bardziej kosztowne niż korzyści wynikające z szybkiego dostępu do informacji. Zaletą stan-

dardu ISA 95 oraz jego implementacji w obiektowej bazie danych jest skalowalność modelu

danych bez konieczności zmiany struktury bazy. Do skutecznej implementacji standardu

ISA95 konieczna jest fuzja pomiędzy silnikiem bazy relacyjnej i modelem ISA95. Dopiero

takie rozwiązanie pozwoli na efektywną realizację zunifikowanego modelu przepływu

informacji zaproponowanego przez standard ISA95. Umożliwi to realizację systemów klasy

MES z uwzględnieniem zarówno wymagań elastyczności modelu biznesowego, jak i wyso-

kiej wydajności operacji na bazie danych.


Addresses

Udo WOZAR: Conti Temic microelectronic GmbH, Ringlerstraße 17, 85057 Ingolstadt,

Germany, [email protected]

Hueseyin ERDOGAN: Conti Temic microelectronic GmbH, Ringlerstraße 17, 85057

Ingolstadt, Germany, [email protected]

Rafal CUPEK: Silesian University of Technology, Institute of Informatics,

ul. Akademicka 16, 44 100 Gliwice, Poland, [email protected]

Szymon ZIEMEK: Silesian University of Technology, Institute of Informatics,

ul. Akademicka 16, 44 100 Gliwice, Poland, [email protected]

Documents

APPLICATION OF ISA95 DATA MODELS IN MANUFACTURING ...€¦ · Manufacturing Execution Systems (MES) are service-oriented interfaces that join the word of business transactions with