Title:
TOOL CONDITION MONITORING DATASET ENHANCEMENT METHOD BASED ON GENERATIVE ADVERSARIAL NETWORK
Document Type and Number:
WIPO Patent Application WO/2021/128576
Kind Code:
A1
Abstract:
The present invention relates to the technical field of machining condition monitoring, and provides a tool condition monitoring dataset enhancement method based on a generative adversarial network. The method comprises: firstly, obtaining vibration and sound signals in a tool cutting process by using a sensor acquisition system; secondly, inputting noise data complying with prior distribution into a generator to generate data, and inputting the generated data and the acquired real sample data into a discriminator for discrimination, adversarial training being performed between the generator and the discriminator until the training is completed; then, generating sample data by using the trained generator, and determining whether the distribution of the generated sample data and the distribution of real tool condition sample data are similar; and finally, predicting the accuracy of tool conditions in combination with a deep learning network model to test the availability of the generated data. The method has the greatest advantage of enhancing the tool condition dataset, thereby improving the accuracy of predicting tool conditions by the deep learning network model.
Inventors:
WANG YONGQING (CN)
NIU MENGMENG (CN)
LIU KUO (CN)
QIN BO (CN)
SHEN MINGRUI (CN)
LI DAWEI (CN)
NIU MENGMENG (CN)
LIU KUO (CN)
QIN BO (CN)
SHEN MINGRUI (CN)
LI DAWEI (CN)
Application Number:
PCT/CN2020/077095
Publication Date:
July 01, 2021
Filing Date:
February 28, 2020
Export Citation:
Assignee:
UNIV DALIAN TECH (CN)
International Classes:
B23Q17/09
Foreign References:
| CN109158953A | 2019-01-08 | |||
| CN104741638A | 2015-07-01 | |||
| CN108107838A | 2018-06-01 | |||
| CN109977464A | 2019-07-05 |
Download PDF: