An Extreme Gradient Boosting Aided Fault Diagnosis Approach: A Case Study of Fuse Test Bench
Muhammad Gibran Alfarizi, Jørn Vatn, Shen Yin
Abstract
The health status of a fuse test bench is essential to monitor to ensure quality control of the fuse. A system failure during operation will lead to significant impacts on the final quality of fuses. Thus, it is important to have a fault diagnosis system to detect, classify, and identify the root causes of faults to prevent operation failure. An effective fault diagnosis system should have high accuracy, fast diagnosis time, and interpretable root cause analysis. This article proposes an integrated fault diagnosis system based on extreme gradient boosting for an automated fuse test bench to solve those challenges. The proposed diagnosis system is then validated using the dataset from PHM 2021 Data Challenge. Performance comparison of the fault diagnosis system with other standard approaches in practice is also carried out. Experimental results show that the diagnostic accuracy of the proposed system outperforms several standard fault diagnostic approaches.