Department of Mechanical and Industrial Engineering, Norwegian University of Science and Technology-NTNU, 7491 Trondheim, Norway.
Water Linked AS, 7041 Trondheim, Norway.
Sensors (Basel). 2022 Jul 11;22(14):5187. doi: 10.3390/s22145187.
Aiming at early detection of subsurface cracks induced by contact fatigue in rotating machinery, the knowledge-based data analysis algorithm is proposed for health condition monitoring through the analysis of acoustic emission (AE) time series. A robust fault detector is proposed, and its effectiveness was demonstrated for the long-term durability test of a roller made of case-hardened steel. The reliability of subsurface crack detection was proven using independent ultrasonic inspections carried out periodically during the test. Subsurface cracks as small as 0.5 mm were identified, and their steady growth was tracked by the proposed AE technique. Challenges and perspectives of the proposed methodology are unveiled and discussed.
针对旋转机械中因接触疲劳导致的表面下裂纹的早期检测，提出了一种基于知识的数据分析算法，通过对声发射（AE）时间序列进行分析来实现健康状况监测。提出了一种鲁棒的故障检测器，并通过对经过渗碳硬化处理的钢制成的滚子进行的长期耐久性测试证明了其有效性。在测试过程中定期进行独立的超声波检查，证明了表面下裂纹检测的可靠性。该 AE 技术能够识别出小至 0.5 毫米的表面下裂纹，并跟踪其稳定增长。揭示并讨论了所提出方法的挑战和展望。
