Sensor Selection Framework for Designing Fault Diagnostics System

Area of Science:

• Engineering
• Systems Science
• Data Science

Background:

• Complex engineered systems and their manufacturing processes are increasingly intricate, posing risks of cascading failures.
• Effective health monitoring and prognostic maintenance are vital for ensuring the reliability and quality of these systems.
• Current sensor selection methods often overlook critical sensor specifications, limiting cost-effectiveness and accuracy.

Purpose of the Study:

• To address limitations in existing sensor selection approaches for complex engineered systems.
• To propose a novel method, OFCCaTS (Optimal Framework for Cost-Conscious sensor Configuration and Targeting), for sensor selection.
• To enhance system health monitoring accuracy while minimizing associated costs.

Main Methods:

• Developed a scalable multi-objective framework for sensor selection.
• Integrated sensor specifications and cost considerations into the selection process.
• Utilized a wind turbine gearbox as a case study to demonstrate the framework's efficacy.

Main Results:

• The OFCCaTS framework effectively balances maximizing fault detection rates with minimizing total sensor costs.
• Demonstrated the ability to select an optimal subset of informative sensors, avoiding redundancy.
• Successfully applied the framework to a real-world engineering system (wind turbine gearbox).

Conclusions:

• The proposed OFCCaTS method offers a significant improvement over existing sensor selection techniques.
• Optimizing sensor selection is key to cost-effective and accurate prognostic health management in complex systems.
• This framework provides a practical approach for designers to enhance system reliability and reduce operational expenses.