상호운용적 제조지능화를 위한 OPC UA 기반 데이터 애널리틱스 모델 표현 방법

Abstract

Manufacturing intelligence necessitates interoperable data and model exchange so that machines seamlessly collect and share their data to create and use data-driven models for predictive planning and control. Open Platform Communications Unified Architecture (OPC UA) has become a powerful standard for data interoperability, and has been extended to various industrial sectors through publishing OPC UA companion specifications with collaboration of industrial consortiums. However, companion specifications for the interoperability of data analytics models have not been released yet. The present work proposes an OPC UA-based representation method for the data analytics models formalized by Predictive Model Markup Language (PMML), an XML-based language for interchanging data analytics models. This article includes: definition of mapping rules between PMML and OPC UA, design of OPC UA information models compliant with PMML, and implementation of a server-client prototype that publishes and subscribes two OPC UA model instances transformed from PMML-based regression and neural network models.