A bio-inspired memory model embedded with a causality reasoning function for structural fault location.

Area of Science:

• Engineering
• Computer Science
• Data Science

Background:

• Structural health monitoring (SHM) faces significant challenges in managing large datasets and accurately locating structural faults.
• Existing methods often struggle with data storage demands and pinpointing the root cause of failures.

Purpose of the Study:

• To propose a bio-inspired memory model with causality reasoning for effective fault location in SHM systems.
• To address the limitations of massive data storage and improve the accuracy of fault identification.

Main Methods:

• Dividing SHM data into three temporal memory areas to manage data volume.
• Mining causal relationships within structural state monitoring data.
• Developing causality and dependence indices for quantitative description of events.
• Implementing a causality reasoning mechanism for fault localization.

Main Results:

• The proposed model demonstrated effective real-time data acquisition and compact storage.
• Successful application in system fault location was shown through a deformation experiment on a steel spring plate.
• The model's performance was validated against typical methods using an experimental benchmark dataset.

Conclusions:

• The bio-inspired memory model with causality reasoning offers a viable solution for SHM challenges.
• The model enhances efficiency in data management and fault localization in structural monitoring.
• This approach shows promise for real-time structural health assessment and maintenance.