On June 21, 9:15-10:45, at the Scienitfic Council Hall, 1st floor, the Director of the IPPT PAN invites all the faculty and students to the special IPPT PAN Seminar:
Professor Stephen Ekwaro-Osire
Interim Chair, Department of Mechanical Engineering, Texas Tech University, USA
Title of part 1:
Research at the Department of Mechanical Engineering
Abstract of part 1:
Texas Tech University (TTU) is a comprehensive public research university and one of four component institutions, under the umbrella of the TTU System. In fall 2018, more than 38,000 students were enrolled in the university. It comprises 12 colleges and schools. TTU is one of 120 US universities and 86 public institutions to be listed in the Carnegie Very High Research Activity category. The department of mechanical engineering is one of the largest departments at the university. The enrollment in fall 2017 was 1,762. It awarded a total of 389 degrees in December 2017, May 2018, and August 2018. In the past year, department faculty have served as PI/Co-PI on nearly $10 million in externally sponsored research, of which nearly 75% were from federal sources. The department has five research themes, namely, multi-scale biomechanics and mechano-biology, sustainable and resilient energy systems, micromechanics of deformable materials, energetic nanomaterials, and advanced technologies in manufacturing.
Title of part 2:
Condition Monitoring of Gearbox Components Using Deep Learning with Simulated Vibration Data
Abstract of part 2:
Transmission components are prone to fatigue damage due to high and intermittent loading cycles, which often results in premature failure of gearbox components. Early fault diagnostics of these components is essential to avoid sudden failures during operation. Recently, several vibration-based diagnostics approaches using Machine Learning (ML) and Deep Learning (DL) algorithms have been proposed to identify gearboxes faults. However, most of them rely on a large amount of training data collection from physical experiments, which is often associated with high costs. This study offers an ML and DL based condition monitoring techniques for a gear pair, a planetary gear set, and a journal bearing with simulated vibration datasets. Dynamic models of a gear pair and planetary gear set were developed with several crack fault conditions and a journal bearing with several levels of wear and ovalization faults. Using the dynamic models, all the datasets were produced and fed to ML and DL methods and accuracy results were compared. Results revealed the superiority of deep learning algorithms and their great accuracies in fault predictions. This research contributes to the prevention of catastrophic failures in gearbox components by early fault detection and maintenance schedule optimization.
The seminar will be followed by presentations of the IPPT PAN research projects, June 21, 10:45-12:15, the same place.
