A sensor fault detection scheme as a functional safety feature for DC-DC converters

Research output: Journal contributionsJournal articlesResearchpeer-review


DC-DC converters are widely used in a large number of power conversion applications. As in many other systems, they are designed to automatically prevent dangerous failures or control them when they arise; this is called functional safety. Therefore, random hardware failures such as sensor faults have to be detected and handled properly. This proper handling means achieving or maintaining a safe state according to ISO 26262. However, to achieve or maintain a safe state, a fault has to be detected first. Sensor faults within DC-DC converters are generally detected with hardware-redundant sensors, despite all their drawbacks. Within this article, this redundancy is addressed using observer-based techniques utilizing Extended Kalman Filters (EKFs). Moreover, the paper proposes a fault detection and isolation scheme to guarantee functional safety. For this, a cross-EKF structure is implemented to work in cross-parallel to the real sensors and to replace the sensors in case of a fault. This ensures the continuity of the service in case of sensor faults. This idea is based on the concept of the virtual sensor which replaces the sensor in case of fault. Moreover, the concept of the virtual sensor is broader. In fact, if a system is observable, the observer offers a better performance than the sensor. In this context, this paper gives a contribution in this area. The effectiveness of this approach is tested with measurements on a buck converter prototype.

Original languageEnglish
Article number6516
Issue number19
Number of pages18
Publication statusPublished - 01.10.2021

Bibliographical note

Funding Information:
Since the Academic Year 2017/2018, Paolo Mercorelli has been a Visiting Professor at Łódź University of Technology (Poland) and he would like to thank the students of the course of "Modelling Methods of Analog Circuits" of the summer semesters of the year 2019 at the Master Course in Robotics and Automation for the common discussion during the lectures related to Kalman Filter structures.

Publisher Copyright:
© 2021 by the authors. Licensee MDPI, Basel, Switzerland.

    Research areas

  • DC-DC power converters, Fault detection, Kalman filters, Power system fault protection, Safety
  • Engineering