Fault diagnosis of rotation vector reducer for industrial robot based on a convolutional neural network

Shuai Yang (Avtor), Xing Luo (Avtor), Chuan Li (Avtor)

Povzetek

As a key component of a mechanical drive system, the failure of the reducer will usually cause huge economic losses and even lead to serious casualties in extreme cases. To solve this problem, a two-dimensional convolutional neural network (2D-CNN) is proposed for the fault diagnosis of the rotation vector (RV) reducer installed on the industrial robot (IR). The proposed method can automatically extract the features from the data and reduce the connections between neurons and the parameters that need to be trained with its local receptive field, weight sharing, and subsampling features. Due to the aforementioned characteristics, the efficiency of network training is significantly improved, and verified by the experimental simulations. Comparative experiments with other mainstream methods are carried out to further validate the fault classification accuracy of the proposed method. The results indicate that the proposed method out-performs all the selected methods.

Ključne besede

fault diagnosis;convolutional neural networks;RV reducers;

Podatki

Jezik:	Angleški jezik
Leto izida:	2021
Tipologija:	1.01 - Izvirni znanstveni članek
Organizacija:	UL FS - Fakulteta za strojništvo
UDK:	681.5:007.52
COBISS:	82633475
ISSN:	0039-2480
Št. ogledov:	143
Št. prenosov:	51
Ocena:	0 (0 glasov)
Metapodatki:

Ostali podatki

Sekundarni jezik:	Slovenski jezik
Sekundarni naslov:	Diagnosticiranje napak na reduktorjih RV za industrijske robote na osnovi konvolucijske nevronske mreže
Sekundarne ključne besede:	diagnosticiranje napak;konvolucijske nevronske mreže;reduktorji RV;
Vrsta dela (COBISS):	Članek v reviji
Strani:	str. 489-500
Letnik:	ǂVol. ǂ67
Zvezek:	ǂno. ǂ10
Čas izdaje:	Okt. 2021
DOI:	10.5545/sv-jme.2021.7284
ID:	13770059