An ISO/IEC 81346-inspired domain specific language to extract time series data for analytics
Magnus Bakken · 2022 IEEE/SICE International Symposium on System Integration (SII) · 2022
OPC UA information models support standardized models of industrial assets, but have problematic support for queries. SPARQL is the query language of Semantic Web Technologies (SWT). Translation of OPC UA information models to SWT and using SPARQL has been proposed as a solution to querying OPC UA information models and to extracting time series data from them. We argue that users of SPARQL based OPC UA queries may be overloaded by the amount of background knowledge required. In addition to knowing SPARQL, users must understand the industrial standards that are used for representing their industrial asset, how they are encoded using OPC UA information models and represented in SWT after translation. For the case of ISO/IEC 81346 (RDS) providing contextualization and IEC 61850 describing sensor data from equipment we create a domain specific query language for extracting time series data for analytics purposes which produces much shorter queries with lower requirements for technical knowledge than their SPARQL equivalents. We describe the expressiveness of the language, and argue that it is well suited to the needs of data scientists and data engineers. A grammar together with a translator to SPARQL is open sourced under the Apache 2.0 license.