Research on failure identification of mechanical parts of automobile transmission based on deep reinforcement learning
by Heyang Guo; Qiong Li
International Journal of Computational Intelligence Studies (IJCISTUDIES), Vol. 12, No. 1/2, 2023

Abstract: Aiming at the problems of low accuracy and slow speed of traditional method, this paper proposes failure identification method of mechanical parts of automobile transmission based on deep reinforcement learning. Firstly, the failure factors of mechanical gears and bearings in automobile transmissions are analysed, and the failure data are extracted. Then, the mean value, skewness value and kurtosis value of the dimensional parameters of the mechanical part failure fault are calculated, and the pre-processing of the part failure data is realised. Finally, the weight and offset value of the failure data are calculated, the failure data tag is constructed, the loss function through in-depth reinforcement learning is determined, and the failure point of mechanical parts is determined by searching for the failure reward value of mechanical parts to complete the failure identification. Experimental results show that the proposed method can effectively improve the accuracy of failure identification of mechanical parts, and the speed is fast.

Online publication date: Mon, 24-Jul-2023

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Computational Intelligence Studies (IJCISTUDIES):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com