T13.6 - Fault Diagnosis and Tolerant Strategy for Triple-Port Hydrogen Converter using SSA-Optimized Random Forest Algorithm

Wednesday, March 19, 2025

10:40 AM – 11:00 AM ET

Location: Level Three, A316

Technical Lecture Presenter(s)

Shiqi Zhang

Yanshan University

Presentation Paper

Co-Author(s)

YT

Yiyina Teng

Department of Electrical Engineering, Yanshan University
ND

Naizhe Diao

Department of Electrical Engineering, Yanshan University
XG

Xiaoqiang Guo

Yanshan University
VT

Vladimir Terzija

Newcastle University
LW

Lichong Wang

Shenke Electronic Co., Ltd.

Multi-port hydrogen converters are crucial for hydrogen production power sources, requiring high reliability. However, current fault diagnosis and fault-tolerant solutions for such converters are inadequate, relying on numerous sensors, offering limited diagnostic accuracy, and struggling with multiple switch fault detection. Additionally, existing fault-tolerant strategies cannot ensure stable power for electrolyzers or enable power-sharing between ports. To address these challenges, this paper presents a method for diagnosing open-circuit faults and implementing fault tolerance in triple-active bridge (TAB) hydrogen converters. The method uses a sparrow search algorithm (SSA) optimized random forest (RF) algorithm to achieve high accuracy, diagnosing faults in multiple switches using fewer sensors. For fault-tolerant operation, auxiliary switches maintain stable system output and enable power-sharing between ports after a fault. This ensures the hydrogen electrolyzer operates efficiently and reliably. Simulations validate the solution's effectiveness, demonstrating its economic viability and reliability, providing a foundation for safe and efficient green hydrogen production.