Title:
METHOD FOR MONITORING TOOL WEAR IN DEEP HOLE MACHINING BASED ON SSAE-LSTM MODEL
Document Type and Number:
WIPO Patent Application WO/2021/046737
Kind Code:
A1
Abstract:
A method for monitoring tool wear in deep hole machining based on SSAE-LSTM model, comprising: first, respectively installing two three-way acceleration sensors (5, 7) on the outside of the bushings of stands of two tool bars (6) of a deep hole machine tool, and installing a sound collector (3) at the entrance of a deep hole workpiece (1) to collect the tool bar vibration and cutting sound data during the machining; next, performing greedy training on a stacked autoencoder layer by layer using the collected data, and performing feature selection on the data using the trained stacked autoencoder to obtain simplified data; and then, training a long-term and short-term memory network using the simplified data. If the training prediction error is lower than a set value, δ, the model can be used for tool wear prediction; and in real-time monitoring, the real-time vibration and sound data are input into the trained stacked autoencoder and the long-term and short-term memory network, so that the network outputs the tool wear. The present method can realize the monitoring of tool wear during deep hole machining.
Inventors:
LIU KUO (CN)
LI DAWEI (CN)
SHEN MINGRUI (CN)
LIU QINGJUN (CN)
REN HUIMIN (CN)
WANG YONGQING (CN)
LI DAWEI (CN)
SHEN MINGRUI (CN)
LIU QINGJUN (CN)
REN HUIMIN (CN)
WANG YONGQING (CN)
Application Number:
PCT/CN2019/105282
Publication Date:
March 18, 2021
Filing Date:
September 11, 2019
Export Citation:
Assignee:
UNIV DALIAN TECH (CN)
International Classes:
B23Q17/09; B23Q17/12; G06N3/04
Foreign References:
| CN110153802A | 2019-08-23 | |||
| CN109048492A | 2018-12-21 | |||
| CN108319962A | 2018-07-24 | |||
| JP2000233369A | 2000-08-29 | |||
| DE3627796C1 | 1987-10-22 | |||
| JPS59211856A | 1984-11-30 |
Download PDF: