A Dynamic Software Product Line Architecture for Prepackaged Expert Analytics: Enabling Efficient Capture, Reuse and Adaptation of Operational Knowledge

Abstract

Advanced asset health management solutions blend business intelligence with analytics that incorporate expert operational knowledge of industrial equipment and systems. Key challenges in developing these solutions include: streamlining the capture and prepackaging of operational experts' knowledge as analytic modules, efficiently evolving the modules as knowledge grows, adapting the analytics in the field for diverse operating circumstances and industries, and executing the analytics with high performance in industrial and enterprise software systems. A Quality Attribute Workshop (QAW) was used to elicit and analyze variability at development time and runtime for creating, integrating, evolving, and tailoring reusable analytic modules for ABB/Ventyx asset health solution offerings. Dynamic software product line (DSPL) architecture approaches were then applied in designing an analytics plug in architecture for asset health solutions. This paper describes our approach and experiences in designing the analytics product line architecture and its SME Workbench toolset, and how we achieved significant improvements in speed and flexibility of deploying industrial analytics.