Semantically Enriched Industry Data & Information Modelling: A feasibility study on Shop-floor Incident Recognition

Dr. Τhanasis Vafeiadis Information Technologies Institute Center for Research and Technology Hellas [email protected]

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Outline  Introduction  CIDEM  Semantic Enrichment  Feasibility Study  Conclusion

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Introduction

Issues  Modelling of industry knowledge is a cumbersome task due to: o Large amount of data; o Lack of automated information-modelling tools; o Continuous emersion of innovative and complex industrial standards; o Fast and continuously changing industrial environment; and o Various heterogeneous sources.

Solution Proposed  Shop-floor implicit and explicit o Robust integration of multiple existing industrial standards into an open source common information model; o Semantic enrichment of provided information offering machine understandable data to industrial systems; and o Facilitated information flows necessary for the recognition of accidents and path optimization, during working time within high risk shop-floor areas.

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Industry Knowledge Modelling

 Building Information Modelling (BIM) o Still under active and intensive research o Potential for  reducing project cost  reducing delivery time  increasing productivity  increasing quality o Average Return-on-Investment (ROI) more than 600% in studied use cases.

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Industry Knowledge Modelling

What’s the issue?  Design and implementation of such models are time consuming,  Legal pitfalls  Limited range of applications  Data Interoperability issues  Highly diverse parties  Targeted specialization of industrial standards to specific sectors or equipment

The real world, in terms of resources, ideas, events, etc., has to be symbolically defined within physical data stores, thus capturing more meaning of the data through semantic models.

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Common Information Data Exchange Model (CIDEM)

 Translation of the information to a common understandable format

 Model information elements o Concepts o Relations o Interfaces

 Shared and common vocabulary enabling to address the information needs of various and heterogeneous environments

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Common Information Data Exchange Model (CIDEM)

 Physical storage of the information

 Handling (storing/retrieving) heterogeneous information via Restful services

 Interoperability with Commercial/Industrial standards (e.g. B2MML, gbXML, etc.)

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 CIDEM Data Structure

 XML Schemas o Broad acceptance facilitate information exchange

o Semantics compatible straightforward transition to various ontology formats (RDF/XML)

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 CIDEM Data Structure

 Static Information

Dynamic Information

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Interoperability

 Commercial & Industrial Standards o Import complete or partial XML schemas

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 CIDEM API  The client communicates with MongoDB database via RESTful

 HTTP requests call the web-service server

 Methods in order to o insert o retrieve o update o delete data in/from the CIDEM Repository

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Repository - MongoDB  An open-source NoSQL database  The information is stored in JSON-like documents  Related information is stored together for fast query access through the MongoDB query language.  Allows to build applications faster, handle highly diverse data types, and manage applications more efficiently at scale  Significantly reduced development and operational complexity  Ensures data quality with document validation

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Semantic Enrichment Ontology Design

NeOn Methodology

“A scenario-based methodology that supports a knowledge reuse approach, as well as collaborative aspects of ontology development and dynamic evolution of ontology networks in distributed environments”

Key Aspects  A set of 9 scenarios for the construction of ontologies and ontology networks  The NeOn Glossary of Processes and Activities  Methodological guidelines for a variety of processes and activities of the ontology network development process

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Semantic Enrichment Ontology Design

 Scenario 1: From specification to implementation;  Scenario 2: Reusing and re-engineering non-ontological resources (NORs);  Scenario 3: Reusing ontological resources.

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Semantic Enrichment Challenges

 Ontologies from pre-defined schemas ( CIDEM )  XML covers the syntactic level, but lacks support for reasoning.  Transformation of XML into RDF – RDFised data

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Semantic Enrichment Approach

 XSD-based Ontologies that completely adhere with the schemas  Upper Ontology – Hierarchical Structure o Reuse Concepts & Relations o Introduce Domain Concepts

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Open Semantic Framework (OSF)

 An Open Source integrated stack that through the use of semantic technologies provides: o across all content and data types o Knowledge management o across the enterprise o Distributed, differential data access and permissions, and o Publishing and managing your information.  Accessible via the Web that o Available under the Apache 2 license

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Feasibility Study – Incident Recognition

Safety of the workers in a plant with chemical processes

 Facilitate information flows for: o Recognition of accidents o Path optimization for workers’ movement

 Towards o Reducing the number of accidents o Giving optimal paths to workers o Providing meaningful information regarding shop floor safety

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Upper Ontology Shop Floor Ontology Model

Partial Model Shop Floor Workers

Concept Description Worker Operators & Technicians Pilot Plant Small Dedicated Industrial System Area Accessible or Forbidden Area

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Data Exchange Flow

Thermal Camera Depth Camera Manager Manager

CIDEM

RESTful Web Service

Forbidden Areas

RDFizing & Semantic SPARQL Workers Triple Store Loading Triple Store Manager Engine Processs

Incident Statistics

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 SPARQL Engine

 RDF Query Language  Retrieving Semantic Results

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 Conclusions

 Semantics Enriched CIDEM: o Offers a shared vocabulary for shop floor knowledge; o Ensures easy access and interoperability; o Introduces a multi-layered, accessible and comprehensive information ecosystem to all involved stakeholders;

 Through the feasibility study (preliminary results) o Enhanced worker safety o Reduced accidents

14th International Conference on Industrial Informatics Poitiers - 19/07/2016 The End…

Thank you very much for your attention

Questions / Comments ?

This work was partially supported by the EU funded H2020 – IA – 636302 – SatisFactory project

14th International Conference on Industrial Informatics Poitiers - 19/07/2016