A Sensor Fault Detection Scheme as a Functional Safety Feature for DC-DC Converters

Area of Science:

• Electrical Engineering
• Control Systems
• Power Electronics

Background:

• DC-DC converters are critical in power conversion, requiring robust functional safety measures.
• Sensor faults pose a risk to the safe operation of these converters.
• Traditional hardware redundancy for fault detection has limitations.

Purpose of the Study:

• To propose and validate a fault detection and isolation scheme for sensor faults in DC-DC converters.
• To enhance functional safety and ensure service continuity.
• To leverage observer-based techniques as an alternative to hardware redundancy.

Main Methods:

• Utilized Extended Kalman Filters (EKFs) for observer-based fault detection.
• Implemented a cross-EKF structure operating in parallel with real sensors.
• Developed a virtual sensor concept to replace faulty sensors.

Main Results:

• Demonstrated effective fault detection and isolation of sensor faults.
• Showcased the virtual sensor's ability to ensure service continuity.
• Validated the approach through experimental measurements on a buck converter prototype.

Conclusions:

• Observer-based virtual sensors offer a promising alternative to hardware redundancy for sensor fault management.
• The proposed cross-EKF scheme effectively guarantees functional safety in DC-DC converters.
• This approach contributes to improved reliability and performance in power electronic systems.