Title:
SDAE-DBN ALGORITHM-BASED ONLINE PREDICTION METHOD FOR SURFACE ROUGHNESS OF PART
Document Type and Number:
WIPO Patent Application WO/2021/128577
Kind Code:
A1
Abstract:
An SDAE-DBN algorithm-based online prediction method for surface roughness of a part. The method comprises: firstly, attaching a three-way acceleration sensor to a rear bearing of a machine tool main shaft by means of a magnetic base, placing a microphone at the front left of a machined part, to acquire vibration and noise signals during a cutting process of the machine tool; eliminating the trend of dynamic signals, and then smoothing the signals; secondly, intercepting and normalizing the data within the machining process; then constructing a stacked denoising self-encoder, training a network by adopting a greedy layer-by-layer algorithm, and taking extracted features as an input training network structure of a deep belief network; finally, performing data processing on the real-time vibration and noise signals during machining and then inputting same into a deep network, and the network outputting the quality situation of roughness of the current surface, thereby implementing real-time prediction of roughness of the surface. The method can reduce the participation of human power and expert experience, reduce the difficulty in acquiring labeled data, and can improve the accuracy of surface roughness prediction.
Inventors:
LIU KUO (CN)
SHEN MINGRUI (CN)
QIN BO (CN)
HUANG RENJIE (CN)
NIU MENGMENG (CN)
WANG YONGQING (CN)
SHEN MINGRUI (CN)
QIN BO (CN)
HUANG RENJIE (CN)
NIU MENGMENG (CN)
WANG YONGQING (CN)
Application Number:
PCT/CN2020/077096
Publication Date:
July 01, 2021
Filing Date:
February 28, 2020
Export Citation:
Assignee:
UNIV DALIAN TECH (CN)
International Classes:
G06F30/20; G06F111/10
Foreign References:
| CN110059442A | 2019-07-26 | |||
| CN108596158A | 2018-09-28 | |||
| CN108920812A | 2018-11-30 | |||
| CN103761429A | 2014-04-30 |
Other References:
YU, KUN: "Fault Diagnosis of Rolling Element Bearings Based on Deep Belief Network and Multiple Sensors Information Fusion", CHINA MASTER’S THESES FULL-TEXT DATABASE, 1 December 2016 (2016-12-01), pages 1 - 101, XP055824390, ISSN: 1674-0246
Download PDF: