A novel method for tool condition monitoring based on long short-term memory and hidden Markov model hybrid framework in high-speed milling Ti-6Al-4V

Tao, Zhengrui; An, Qinglong; Liu, Gongyu; Chen, Ming

doi:10.1007/s00170-019-04464-w

Cited by 43 publications

(19 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A total of 50,000 model parameters of the posterior distribution of samples were achieved, in which previous 10,000 samples were used to ensure the convergence of the MCMC. The estimated values of h ¼ 0.0826 and l ¼ 7.347 were obtained from the posterior samples and substituted into equation (10) to obtain the trend of tool reliability over time, as shown in Figure 8.…”

Section: Tool Reliability Analysis Without Considering Individual Difmentioning

confidence: 99%

“…According to the literature, 16 the prior distribution of the two parameters is d*Ga (10,64), g*Ga(64,100), and h*U(0,50). The same OpenBUGS software was used to generate tool model parameter samples.…”

Section: Tool Reliability Analysis With Considering Individual Differmentioning

confidence: 99%

“…Yu et al 9 proposed a tool life prediction method based on weighted hidden Markov model, which was verified the higher accuracy than the traditional Markov model by collecting data from high-speed CNC milling experiment. Tao et al 10 proposed a prediction model for the tool residual life based on the shorttime memory network and hidden Markov combination. Benkedjouh et al 11 put forward a tool residual life prediction model based on nonlinear feature reduction and support vector regression.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Tool reliability of sintered diamond drill bit for processing silicon carbide ceramics based on Bayesian theory

Guo

Shi

et al. 2020

Advanced Composites Letters

View full text Add to dashboard Cite

Silicon carbide (SiC) ceramics, as a superhard nonmetallic material, would induce severe tool wear during drilling. To ensure the quality of processing, the drill bits must be replaced in advance before failure, which leads to serious waste without effective usage. Aiming at the tool reliability problem determined by the tool wear during drilling of the high hardness nonmetallic materials, this article proposed a reliability analysis method of sintered diamond drill bit for machining SiC ceramics based on the distribution of tool degradation increment and the modeling of gamma process, and a tool reliability model of degradation increment was established based on the Bayesian theory. The drilling experiment of sintered SiC ceramics was carried out by sintering diamond drill bit to obtain the degradation increment data. The accuracy of the model was tested by the reliability model deviation rate, which indicated that the reliability model with considering individual differences was more accurate. This research can provide a theoretical model and analytical method for improving the reliability of the nonmetallic brittle material processing tools.

show abstract

Section: Tool Reliability Analysis Without Considering Individual Difmentioning

confidence: 99%

Section: Tool Reliability Analysis With Considering Individual Differmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Tool reliability of sintered diamond drill bit for processing silicon carbide ceramics based on Bayesian theory

Guo

Shi

et al. 2020

Advanced Composites Letters

View full text Add to dashboard Cite

show abstract

“…Second, the simple architecture of ANNs lacks sufficient breadth and depth to map complex nonlinear relationship. The development of deep learning have relieved the above problems to a certain extent [19,20]. Deep learning can adaptively learn hierarchical representation without extracting the fault features manually [21,22], which is beneficial to improve adaptability.…”

Section: Introductionmentioning

confidence: 99%

Tool remaining useful life prediction using deep transfer reinforcement learning based on long short-term memory networks

Yao

Zhang

2021

Int J Adv Manuf Technol

View full text Add to dashboard Cite

show abstract

“…[14][15][16], the latter creates more and more competitive architectures in the different fields, medicine, maintenance, and Aerospace, etc. among these architectures are Probabilistic Neural Network [17], Multi-Layer Perceptron, Convolutional Neural Network and Recurrent Neural Network (RNN) [18,19], the advantage of the latter is remarkable especially with datasets of sequential nature, to capture the temporal relation between the different sequences, unfortunately, despite these advantages, it presented a major disadvantage, is a gradient vanishment, which can cause losses of long-term information, and to remedy this problem, an architecture has been proposed which is LSTM, allowing to ensure long and short term dependency [20].…”

Section: Introductionmentioning

confidence: 99%

Tool Wear Conditions Monitoring Based on Kernel Smoothing Density and Deep Learning

Bazi¹,

Benkedjouh²,

Habbouche³

et al. 2021

Preprint

View full text Add to dashboard Cite

Life prediction and health assessment of cutting tools is a challenging problem in industrial manufacturing, and plays an important role for Prognostic activities. The degradation of this tool can cause significant economic losses and risks for machine users. However, due to the random nature of system degradation and the sensitivity of various features characterizing cutter degradation, may vary considerably under different operation conditions. This paper propose a new data driven approach for tool condition monitoring, based on kernel smoothing density to model the random phenomena of degradation progression. The health assessment is insured using one directional neural network to extract useful information’s. Deep learning allows relating between extracted features and health indicator progression, and estimate remaining useful life. This methodology is evaluated using regression metrics and compared with recent frameworks, based on an experimental dataset from PHM Data Challenge 2010. The obtained results shows that the proposed approach can effectively predict the tool wear and obtain higher prediction accuracy than other comparison models.

show abstract

A novel method for tool condition monitoring based on long short-term memory and hidden Markov model hybrid framework in high-speed milling Ti-6Al-4V

Cited by 43 publications

References 39 publications

Tool reliability of sintered diamond drill bit for processing silicon carbide ceramics based on Bayesian theory

Tool reliability of sintered diamond drill bit for processing silicon carbide ceramics based on Bayesian theory

Tool remaining useful life prediction using deep transfer reinforcement learning based on long short-term memory networks

Tool Wear Conditions Monitoring Based on Kernel Smoothing Density and Deep Learning

Contact Info

Product

Resources

About