Intelligent Turning Tool Monitoring with Neural Network Adaptive Learning

Maohua Du, Peixin Wang, Junhua Wang, Zheng Cheng and Shensong Wang

Complexity, 2019, vol. 2019, 1-21

Abstract:

Tool state monitoring is a key technology in intelligent manufacturing. But it is still in a research stage and lacks general adaptability for different machining conditions. To overcome this limitation, this work systematically investigates an intelligent, real-time, and visible tool state monitoring through adopting integrated theories and technologies, i.e., (a) through distinctively designed experimental technique with comprehensive consideration of cutting parameters and tool wear values as variables, (b) through bisensor fusion for simultaneous measurements of low and high frequency signals, (c) through multitheory fusion of wavelet decomposition and the Relief-F algorithm for performing dual feature extraction and feature dimension reduction to achieve more accurate state identification using neural network, and (d) through an innovative programming technique of MATLAB-nested labVIEW. This tool monitoring system has neural network adaptive learning ability with the change of the experimental sample signals which are collected simultaneously by selected vibration and acoustic emission (AE) sensors in all factors turning experiments. Based on LabVIEW and MATLAB hybrid programming, the waveforms of signals are observed and the significant characteristics of signals are extracted through the time-frequency domain analysis and then the calculation of the energy proportion of each frequency band obtained using 4 levels of wavelet packet decompositions of the vibration signal as well as 8 levels of wavelet multiresolution decompositions of the AE signal; the ensuing Relief-F algorithm is used to further reextract the features that are most relevant to the tool state as input of neural network pattern recognition. Through the BP neural network adaptive learning, tool state recognition model is finally established. The results show that the correct recognition rate of BP network model after samples training is 92.59%, which can more accurately and intelligently monitor the tool wear state.

Date: 2019
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
http://downloads.hindawi.com/journals/8503/2019/8431784.pdf (application/pdf)
http://downloads.hindawi.com/journals/8503/2019/8431784.xml (text/xml)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:hin:complx:8431784

DOI: 10.1155/2019/8431784

Access Statistics for this article

More articles in Complexity from Hindawi
Bibliographic data for series maintained by Mohamed Abdelhakeem (mohamed.abdelhakeem@hindawi.com).