An OPC UA-Compliant Interface of Data Analytics Models for Interoperable Manufacturing Intelligence

Research output: Contribution to journalArticlepeer-review

3 Scopus citations


The open platform communications unified architecture (OPC UA) has received attention as a standard for data interoperability in industries. In particular, OPC UA extends its applicability across various industrial sectors by publishing OPC UA companion specifications created in collaboration with other industrial consortiums. However, OPC UA is limited to ensure the interoperability of data analytics models because the relevant companion specifications have not been developed yet. OPC UA should be extended to implement such model interoperability so that machines seamlessly use and share data analytics models across the layers of manufacturing systems to predict and optimize their performance autonomously and collaboratively in terms of interoperable manufacturing intelligence. This article proposes an OPC UA-compliant interface for the exchange of predictive model markup language (PMML), a domain-independent standard for representing XML-based data analytics models. This article includes the design of mapping rules and OPC UA information models for the exchange between PMML and OPC UA, as well as the implementation of an OPC UA server-client prototype to publish and subscribe to OPC UA-compliant regression and neural network models which have been transformed from PMML-based models.

Original languageEnglish
Article number9200708
Pages (from-to)3588-3598
Number of pages11
JournalIEEE Transactions on Industrial Informatics
Issue number5
StatePublished - 2021 May


  • Cyber-physical production systems
  • data analytics
  • manufacturing intelligence
  • model interoperability
  • open platform communications unified architecture (OPC UA)
  • predictive model markup language (PMML)


Dive into the research topics of 'An OPC UA-Compliant Interface of Data Analytics Models for Interoperable Manufacturing Intelligence'. Together they form a unique fingerprint.

Cite this