Vibration-based Fault Diagonsis of Rotating Machinery Using Intelligent Algorithms
Author :
Assistant Professor Dheivendran .R, Assistant Professor Mr. Shubham S. ShindeJourna Name:
International Journal of Science, Engineering and Technology Volume:
14 issue:6 Year:Volume-14-issue-6 Views : 3
Abstract:
Despite being the most widely used approach for condition monitoring of rotating machines, the problem of vibration signal analysis suffers from non-stationary conditions of operation, noise pollution of fault signature and multi-fault coupling. This paper introduces an intelligent diagnostic approach based on adaptive time-frequency processing, deep feature learning and attention-aware temporal modeling of fault features. In the introduced framework, the proposed dual-branch approach includes one-dimensional convolutional branch which is responsible for impulsive feature extraction from raw vibration signals and bidirectional recurrent branch to capture long-range dependencies. The attention module is applied to weight features in order to reduce influence of the noisy components of vibration signals. The effectiveness of the proposed method was proved experimentally using benchmark bearing dataset with different load and speed conditions and comparison to the baseline shallow and deep learning models. The analysis shows that the introduced approach has a high inter-class separability and operates in a stable way without feature engineering.
APA:Assistant Professor Dheivendran .R, Assistant Professor Mr. Shubham S. Shinde. (Volume-14, Issue-6 -(Year-Volume-14-issue-6)). Vibration-based Fault Diagonsis of Rotating Machinery Using Intelligent Algorithms. Retrieved from https://www.ijset.in/wp-content/uploads/IJSET_V14_issue6_101.pdf
Chicago:Assistant Professor Dheivendran .R, Assistant Professor Mr. Shubham S. Shinde. "Vibration-based Fault Diagonsis of Rotating Machinery Using Intelligent Algorithms" Example, Volume-14-issue-6-Year-Volume-14-issue-6-2348-4098. https://www.ijset.in/wp-content/uploads/IJSET_V14_issue6_101.pdf.