A Framework for an Indoor Safety Management System Based on Digital Twin

Area of Science:

• Smart Buildings
• Building Information Management
• Artificial Intelligence in Safety

Background:

• Current indoor safety management lacks comprehensive data analysis due to poor integration of safety data and building information.
• Safety risk assessment heavily relies on the experience of safety management personnel, leading to potential inconsistencies.
• The rise of smart residential spaces necessitates advanced solutions for building indoor safety intelligence.

Purpose of the Study:

• To propose a novel framework for intelligent indoor safety management in buildings using digital twins (DTs).
• To enhance the comprehensiveness of safety data analysis and reduce reliance on subjective experience for danger assessment.
• To improve the overall intelligence level of building indoor safety management systems.

Main Methods:

• Development of a digital twin model (DTM) by integrating Building Information Modeling (BIM) with operational data from Internet of Things (IoT) sensors.
• Implementation of a Support Vector Machine (SVM) model for automated classification and assessment of danger types and levels using DTM data.
• Utilization of the Internet as a communication medium for system interaction and information dissemination.

Main Results:

• The proposed DT-based system effectively integrates BIM and IoT data for real-time safety monitoring.
• The SVM model accurately identifies and quantifies danger levels, providing objective risk assessments.
• The system demonstrated practical application in a Beijing Winter Olympics venue, showcasing scene display, danger warning, classification, and handling suggestions.

Conclusions:

• Digital twins offer a robust framework for advancing building indoor safety management.
• The integration of IoT, BIM, and AI (SVM) significantly enhances the intelligence and efficiency of safety systems.
• The proposed system provides a scalable and effective solution for smart building safety, applicable to various real-world scenarios.