Search

In this paper we introduce an architecture for a cognitive digital twin that uses an ontology-based knowledge graph to improve the analysis of manufacturing systems. The architecture is evaluated in a use case at Atlas Copco Airpower, revolving around an adaptive measurement strategy for quality control. We report on the requirements, outcomes, and the resulting challenges of using a cognitive digital twin, and provide a roadmap for future research.

To develop AI-based models capable of governing or providing decision support to complex manufacturing environments, abstractions and mechanisms for unified management of data storage and processing capabilities are needed. Specifically, as such models tend to include and rely on detailed representations of systems, components, and tools with complex interactions, mechanisms for simplifying, integrating, and scaling management capabilities in the presence of complex data requirements (e.g., high volume, velocity, and diversity of data) are of particular interest. A data fabric is a system that provides a unified architecture for management and provisioning of data. In this work we present the background, design requirements, and high-level outline of the ASSISTANT data fabric - a flexible data management tool designed for use in adaptive manufacturing contexts. The paper outlines the implementation of the system with specific focus on the use of domain models and the data modeling approach used, as well as provides a generic use case structure reusable in many industrial contexts.

The European project ASSISTANT will provide a set of AI-based digital twins that helps process engineers and production planners to operate collaborative mixed-model assembly lines based on the data collected from IoT devices and external data sources. Such a tool will help planners to design the assembly line, plan the production, operate the line, and improve process tuning. In addition, the system monitors the line in real-time, ensures that all required resources are available, and allows fast re-planning when necessary. ASSISTANT aims to make cost-effective decisions while ensuring product quality, safety and wellbeing of the workers, and managing the various sources of uncertainties. The resulting digital twin systems will be data-driven, agile, autonomous, collaborative and explainable, safe but reactive.

In Domain-Specific Modelling (DSM) the general goal is to provide Domain-Specific Modelling Languages (DSMLs) for domain users to model systems using concepts and notations they are familiar with, in their problem domain. Verifying whether a model satisfies a set of requirements is considered to be an important challenge in DSM, but is nevertheless mostly neglected. We present a solution in the form of ProMoBox, a framework that integrates the definition and verification of temporal properties in discrete-time behavioural DSMLs, whose semantics can be described as a schedule of graph rewrite rules. Thanks to the expressiveness of graph rewriting, this covers a very large class of problems. With ProMoBox, the domain user models not only the system with a DSML, but also its properties, input model, run-time state and output trace. A DSML is thus comprised of five sublanguages, which share domain-specific syntax, and are generated from a single metamodel. Generic transformations to and from a verification backbone ensure that both the language engineer and the domain user are shielded from underlying notations and techniques. We explicitly model the ProMoBox framework's process in the paper. Furthermore, we evaluate ProMoBox to assert that it supports the specification and verification of properties in a highly flexible and automated way.

bibliog