A new health prediction model for a sensor network based on belief rule base with attribute reliability

Area of Science:

• Sensor networks
• Artificial Intelligence
• Reliability Engineering

Background:

• Sensor network reliability is crucial for effective maintenance and health prediction.
• Environmental interference and data limitations degrade monitoring information accuracy.
• Network complexity and data uncertainty pose significant challenges to health prediction.

Purpose of the Study:

• To propose a novel sensor network health prediction model.
• To enhance the accuracy and reliability of health prediction in complex environments.
• To address issues of unreliable monitoring data and expert knowledge uncertainty.

Main Methods:

• Development of a belief rule base model with attribute reliability (BRB-r).
• Incorporation of an attribute reliability mechanism to handle unreliable monitoring data.
• Calculation of sensor reliability using the average distance method.
• Establishment of an optimization model to refine the health prediction model.

Main Results:

• The BRB-r model effectively handles qualitative knowledge and quantitative data.
• The attribute reliability mechanism successfully addresses unreliable sensor network information.
• The optimized model demonstrates improved accuracy in health status estimation.
• A case study on oil storage tank sensor networks validates the proposed method's effectiveness.

Conclusions:

• The proposed BRB-r model offers a robust solution for sensor network health prediction.
• The method enhances network reliability and maintenance guidance.
• The approach effectively manages data uncertainty and network complexity.