Using Semantic Technologies for More Intelligent Steel

Manufacturing

Nikolaos Matskanis

, Stephane Mouton

, Alexander Ebel

and Francesca Marchiori

Centre of Excellence in Information & Communication Technologies (CETIC), Rue Freres Wright 29, Gosselies, Belgium

VDEh-Betriebsforschungsinstitut (BFI), Düsseldorf, Germany

Centro Sviluppo Materiali (CSM), Rome, Italy

Keywords: Data Integration, Data Interoperability, Domain Ontology, Semantic Services, Semantic Software Agents.

Abstract: In recent years, the steel industry has significantly raised its demands regarding product quality,

optimization of production cost, environmental issues and lead-time. The demand for improved production

performance has in turn increased the demand on information systems, in particular highlighting the need

for improved factory- and company-wide collaboration and information exchange. The heterogeneity in

structure, technology and architecture of the information systems deployed in manufacturing plants presents

further challenges to the design and implementation of a data exchange system for process optimization.

1 INTRODUCTION

This article proposes a system that can be used to

increase collaboration of heterogeneous information

systems as part of a platform that aims to optimize

production and product delivery of steel

manufacturing plants. The proposed architecture

uses a distributed paradigm that: i) enables

simultaneous and easy exchange of mutual data; ii)

provides interoperability in the data structure, data

semantics and message exchanging; and iii) applies

rules and data reasoning techniques to optimize

resource allocation.

This service oriented, semantic interoperability

solution is being implemented in the context of an

systems integration project in the Steel

Manufacturing domain. It is a project funded by an

EC RFCS program and its consortium consists of

both large industrial partners and small or medium

enterprises focusing in technological research and

development. The project represents a new approach

to controlling and supervising steel manufacturing

processes that uses Agent technology (Haag and

Cummings, 2006) to infuse intelligence into the

control system. Each agent is a software component

that is characterized by a large degree of autonomy,

and the resulting multi agent system will be capable

of integrating semantics and will be easily

deployable in distributed environments. One use

case that we are considering for the semantic

interoperability services is product re-allocation: i.e.,

steel products that are either not compliant with their

initial order or are over-produced are re-allocated to

a new order. The re-allocation may include further

product processing at any of the plants. Additionally

takes into consideration transportation costs, costs of

trimming or other processing of the product in order

to comply with the order specifications.

2 SEMANTIC

INTEROPERABILITY

SERVICES

One of the platform’s main challenges is to enable

software components, which are agnostic of the

local and remote information systems, to request

data from data stores and sensors and seamlessly

perform optimization operations despite the

syntactic/structural and semantic heterogeneity of

the different data models of steel plants. The

structural heterogeneity issues are related to the way

information is represented at each data source. The

semantic heterogeneity relates to the use of different

terms, languages for referring to the same concept or

different definitions of the same entity.

424

Matskanis, N., Mouton, S., Ebel, A. and Marchiori, F..

Using Semantic Technologies for More Intelligent Steel Manufacturing.

In Proceedings of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2015) - Volume 2: KEOD, pages 424-428

ISBN: 978-989-758-158-8

2.1 Related Work

The idea of using Web Services and Semantic Web

technologies for establishing data integration and

interoperability is an area of scientific research that

is being explored in several application domains and

several approaches already exist and some solutions

have been developed. In (Yang et al., 2005) paper it

is described a Web Service Oriented Architecture

(SOA) and platform that uses semantics and

specifically OWL-S for automatic integration of

manufacturing systems. It focuses on the dynamic

discovery, selection of web services and other

business processes that can be described using

OWL-S. Another approach is presented in (Uddin et

al., 2011), which focuses on the ontology and formal

description of data sources using a domain ontology.

Finally the (Chondrogiannis et al., 2011) and

(Martin et al., 2008) approaches are accomplishing

interoperability of heterogeneous medical

information systems by transforming SPARQL

queries to the different clinical ontologies and code

systems that they have linked as well as different

database schemas that are mapped to domain

ontologies. In their approaches the linking between

the ontologies is done manually or the

transformation is performed by using term

transformation services.

In our approach we have designed and built the

domain ontology and the plant specific extensions

with concept inter-linking in in mind. This linking

between the concepts is encoded in the ontologies

either with the form of inheritance or by using

semantic relationship links between them. By having

the concepts of the ontologies inter-linked, building

services that offer query and schema transformation

and mapping across different manufacturing plants

becomes easier since the mapping information is

already encoded in the ontologies. This enables

software agents to be domain agnostic and query the

plant data using the platform’s upper terminology.

2.2 Service Architecture

In order to achieve the goal of enabling the

platform’s software components to automatically

resolve both heterogeneity types, we have

introduced into the architecture and deployed on the

platform semantic interoperability services. These

services use ontologies (Berners-Lee et al., 2001)

that describe the steel-manufacturing domain and

provide a formal definition of the types, properties

and interrelationships of the entities in this domain

model. The ontologies are structured into two layers,

the high-level domain ontology and the lower level

plant ontologies.

Interoperability is achieved by linking the

concepts/terms of the high-level (core) steel domain

ontology, which is being developed in the project

(Zillner et al., 2014) to provide a common

terminology for all participating plants, and the plant

specific ontology, which provide the terminology,

description of methods and model used in each

plant. These links are implemented either by an

inheritance relation between entities or by explicitly

defined interrelationships between entities of the

high–level and plant ontologies, for example the

Web Ontology Language (OWL) property

owl:sameAs. We are also developing a 3rd level

which maps the database schema with the plant

ontology and effectively adds a SPARQL (Query

Language for RDF) interface for the database. This

way the plant databases can also be accessed and

exposed as RDF (Resource Description Framework)

graphs and effectively queried the same way as

ontologies. The proposed architecture design and

proof of concept implementation of the semantic

service using the domain ontology, plant ontologies

and plant data sources is shown in Figure 1.

Figure 1: Architecture of the proof of concept platform.

The semantic and other services are presented.

The semantic mapping between the ontology

layers creates associations of related concepts of the

domain ontology, the plant ontology and the actual

data in the data stores. This approach allows

translation of the client query from the domain

ontology that all - participating in the project

platform - plants are able to understand and use, to

the local one used in each of the plants, and

Using Semantic Technologies for More Intelligent Steel Manufacturing

425

eventually – if requested – to the database native

query language. This approach is generic enough to

be reused in other applications, for example a similar

solution was applied in the clinical research domain

(Chondrogiannis et al., 2011). The reallocation use

case of the project, which involves multiple plants

with an agent platform each, benefits from this

approach when agents of different plants i) request

for product data values, ii) query a plant’s model to

identify manufacturing processes, logistical or other

plant information.

2.3 Proof of Concept Implementation

2.3.1 Technologies and Tools

Several tools and technologies exist that can be used

for the implementation of a semantically enabled

service for data integration and interoperability. For

the development of the ontology we have considered

the Protégé ontology editor but as our requirements

were for an easily accessible and demonstrable

environment for the plant engineers, the Semantic

Media Wiki (SMW) (Vrandečić and Krötzsch, 2009)

appeared to be more appropriate for this purpose.

The backend of the Media Wiki is a MySQL

database server and for storing the OWL RDF triples

of the SMW extension we have used the 4store

triplestore. 4store provides a SPARQL endpoint and

is used for storing and querying both the Domain

Ontology and the plant specific ones. The choice of

the triplestore was based on its compatibility with

the SMW and because the set of requirements for the

triplestore technology was rather small, a

lightweight server was preferred from the feature-

rich OpenLink Virtuoso server that we were also

considering.

Another key part of the architecture is the

mapping of the plant ontology and the data itself.

Several approaches were investigated. From those

we have tested the Virtuoso server, the Semantika

server (Hardi, 2014) and the D2RQ server (Bizer

and Seaborne, 2004). An additional complication

and challenge was the nature of the Databases we

have used in the project for the plant products and

customer orders. These two datasets are stored in a

NoSQL document-oriented database (Sadalage and

Fowler, 2012), the MongoDB, which dynamically

adjusts its schema of the database to the data that is

stored in each document, which in our case it is each

entry in the database. The absence of concrete

schema introduced problems to the mapping server

of the ontology with the database, which was solved

using the unityJDBC (www.unityjdbc.com) driver

with the D2RQ server. The D2RQ server does not

require validating the schema, unlike Semantika, and

the unityJDBC driver provides the core ODBC

functionality and API. This enabled the plant

product and orders databases to be accessed as RDF

Triplestores using a SPARQL endpoint and be

browse-able by both humans and machines as

Linked Data (Heath et al., 2010), though we have

not openly published them.

Finally the Semantic agents implement the

recommendation of FIPA Agent Management

Specification, which suggests integrating the

ontologies by providing so called “Ontology

Services”. In this approach the access to the external

ontology server by internal agents is organised

through a specialised agent who is providing the

ontology services. These services include forming

the SPARQL queries, utilising the semantic links

and mappings and of course accessing the semantic

services described above. The internal agents can

use the same communication mechanism to access

the ontology as for the intercommunication between

them. The semantic agent can implement additional

functionalities such as transformation of data

returned by the ontology or provide knowledge

about mappings between the different ontologies.

Figure 2 shows a general scheme of the semantic

agent interaction with the ontology services. Here

the sematic agent provides the functionality to

provide a term mapping.

Figure 2: Retrieving local term based on the domain

ontology.

2.3.2 Data Querying

The first use case for the semantic services is to

directly retrieve data from the databases using

SPARQL queries. The mapping of the D2RQ server

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

426

between the plant ontology and the database schema

allows translation of the SPARQL queries to SQL

and retrieval of the requested data from the database.

The semantic agent can either include the terms

from both the domain and plant ontologies in the

query or translate the query by replacing the domain

terms with the plant ones. In our experiments so far

we have used template queries that use the later

approach.

2.3.3 Schema Translation and Mapping

In this use case of the semantic service, the semantic

agent queries the ontology service for a term or

concept that is used in the plant data schema. In

order to produce the result the service requires a

term translation from the core domain ontology to a

term used in the database schema of the plant. The

translation is achieved using the links between the

ontologies that are structured into two layers, the

high-level domain ontology and the lower level plant

ones. The semantic links between concepts that were

previously explained associate the related concepts

of the two ontologies and with the addition of the

URI of the database field provided by the D2RQ

mapping, the complete path to the correct document

and field of the database is produced.

This path is then cached to a JSON file in order

to avoid having the semantic agent calling the

ontology service every time there is a need to access

the database. With this caching of mappings the

agents of the platform can be agnostic of the schema

of the plant database and only need to the domain

ontology terms for the data they need to access. The

format of the cached mapping is provided in the

JSON code bellow:

[{"definition":"http://steel.eu/product.coil

"label":"Coil", ... ,

"dataAttributes":[{

"definition":"http://steel.eu/product.co

il_width",

"label":"Width", ...}],

"equivalentTo":{

"definition":"http://steelcompany.com/Coil",

"label":"Coil", ...,

"dataAttributes":[{

"definition":"http://steelcompany.com/Wi

dth",

"label":"Width", ...}],

"equivalentTo":{

"definition":"https://steelcompany.com

/d2rq/resource/PRODUCTS/Sagunto/PRODUCTS_Sag

unto", ...,

"dataAttributes":[{

"definition":"https://steelcompany.com/d

2rq/resource/PRODUCTS/Sagunto/Width",

"label":"Width", ...}]}

...}]

3 CONCLUSIONS AND FUTURE

WORK

In this paper we describe the semantic services

architecture and prototype that we have produced

around the domain ontology of steel manufacturing

that we have developed and its plant ontology

extensions. The services use the semantic links

between the ontologies and provide mappings to the

data sources of the plants. The semantic agent

connects the puzzle by creating the SPARQL queries

based on templates.

The prototype that was developed was deployed

in one of the partner plants in order to be part of the

reallocation use case experiment of the project.

Although the reallocation agent responsible for

performing allocations of products had no

information on the Products, Orders and other plant

databases and their schemata, it had been successful

in querying these data sources by using the project’s

core domain ontology and by performing schema

term translation operations, which are being cached

for performance reasons. This allows the project to

extend its experiments by easily adding new plants

that are distributed in many locations in Europe and

although belong to the same organisation they have

very different infrastructure and systems. One of the

goals of the project is to allow dynamic adding or

removing plants into the product reallocation service

at runtime and we believe that this architecture and

proof of concept implementation is a step towards

this goal.

With the ontology mapping and semantic

services in place we can further explore the

reasoning capabilities on the ontology and data for

deriving interrelationships, thus expanding the

mapping between ontology and data, which can lead

to identifying new product reallocations using the

order and product data stores. This will further

empower the resource allocation optimization

processes of the platform.

REFERENCES

Berners-Lee, T., Hendler, J., Lassila, O., 2001, The

Semantic Web, Scientific American, 284(5):34-43.

Zillner, S., et al., 2014, A Semantic Modeling Approach

for the Steel Production Domain, 1th European Steel

Technology & Application Days & 31th Journées

Sidérurgiques Internationales (JSI), Paris.

Chondrogiannis, E., Matskanis, N., et al., 2011, Enabling

semantic interlinking of medical data sources and

EHRs for clinical research purposes, eChallenges

Using Semantic Technologies for More Intelligent Steel Manufacturing

427

conference.

Haag, S., Cummings, M., 2006, Management Information

Systems for the Information Age, pp. 224–228.

Uddin, M. K., et al., 2011, An ontology-based semantic

foundation for flexible manufacturing systems, 37th

Annual Conference of the IEEE Industrial Electronics

Society, IECON 2011, pp. 340–345.

Yang, Z., et al, 2005, Automating integration of

manufacturing systems and services: a semantic Web

services approach, 31st Annual Conference of IEEE

Industrial Electronics Society, IECON 2005.

Martin, L., Tsiknakis, M., et al. 2008, Ontology based

integration of distributed and heterogeneous data

sources in ACGT, 1st International Conference on

Health Informatics, HEALTHINF 2008, (vol. 1, pp.

301-306).

Bizer, C., Seaborne, A. 2004, D2RQ – Treating Non-RDF

Databases as Virtual RDF Graphs, International

Semantic Web Conference (ISWC2004).

Hardi, J., 2014, Introduction to Semantika DRM,

http://www.codeproject.com/Articles/787371/Introduc

tion-to-Semantika-DRM.

Vrandečić, D., Krötzsch, M., 2009, Semantic MediaWiki.

In John D. et al., eds.: Semantic Knowledge

Management. Springer.

Heath, T. Bizer C., Berners-Lee, T., 2010. Linked data-the

story so far, International Journal on Semantic Web

and Information Systems, Vol. 5(3), Pages 1-22. DOI:

10.4018/jswis.2009081901

Sadalage, P., Fowler, M., 2012, NoSQL Distilled: A Brief

Guide to the Emerging World of Polyglot Persistence,

Addison-Wesley, ISBN 0-321-82662-0.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

428