A human-centered framework to update digital twins.

Area of Science:

• Engineering
• Computer Science
• Industrial Systems

Background:

• Digital twins (DTs) integrate predictive models with real-time data from physical counterparts.
• Challenges exist in confirming DT steady-state operation and distinguishing sensor data from sensor failure.
• Current DT development lacks clear termination points for design-time activities like uncertainty assessment.

Purpose of the Study:

• To propose a human-centered framework for establishing, refining, and updating digital twins.
• To address the persistent challenge of differentiating sensed changes from sensor malfunctions.
• To improve the reliability and maintainability of digital twin systems.

Main Methods:

• Development of a human-centered framework for digital twin lifecycle management.
• Utilizing condition-based maintenance and gear backlash in production equipment as case studies.
• Focusing on refining predictive models and ensuring data relevance over time.

Main Results:

• Early work suggests a framework can guide the establishment and refinement of digital twins.
• Identified key challenges in maintaining DT accuracy and distinguishing sensor issues.
• Demonstrated the framework's applicability using industrial examples.

Conclusions:

• A human-centered approach is crucial for robust digital twin development and maintenance.
• Addressing uncertainty and data relevance are ongoing processes in DT lifecycle.
• The proposed framework offers a structured method for managing digital twin evolution.