基于实例迁移学习的跨工况刀具剩余寿命预测

doi:10.7527/S1000-6893.2023.29038

Abstract

Abstract:

Accurate and reliable predictions of tool remaining useful life could reduce the rate of over-utilization and under-utilization of tools during machining， thereby maximizing the machining reliability and reducing production costs. Traditional machine learning methods for tool remaining useful life prediction rely heavily on the assumption that training and test data follow the same distribution， as well as extensive offline measurement data. However， in actual machining process， prediction accuracy of the traditional methods is reduced due to the variation in machining conditions and limited tool wear data. To address this problem， an Instance-based Transfer Learning framework is proposed to accurately predict the tool remaining useful life cross different working conditions. Firstly， a transfer learning algorithm is used to dynamically adjust the weights of all instances in multiple source domains， which aims to make full use of the source domain information that is highly correlated with the target data. Thus， the generalization ability of the model is improved， and the remaining tool life of the target working conditions could be well predicted with only a small amount of target domain data. Secondly， recurrent Gaussian process regression model is further developed as the base learner to improve the time series prediction capability of the transfer learning algorithm. The model limits the tool remaining useful life at adjacent moments through delayed feedback， while reducing the feature preparation time and the model complexity are reduced. The results indicate that the proposed framework can effectively improve the prediction accuracy of the tool remaining useful life cross different working conditions， and the prediction effectiveness also confirms the stability and reliability of the framework.

Key words: cross working conditions, instance-based transfer learning, Recurrent Gaussian process regression, time series prediction, tool remaining useful life

CLC Number:

V262.3⁺5

Biyao QIANG, Kaining SHI, Junxue REN, Yaoyao SHI. Instance transfer for tool remaining useful life prediction cross working conditions[J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(13): 629038-629038.

Figures/Tables 14

Fig.1

Fig.2

Table 1

Fig.3

Fig.4

Fig.5

Table 2

Tool wear curve slope analysis of variance

差异源	离差平方和	自由度	均方	$F$ 值	$P$ 值
总计	122.83	8
主轴转速	48.64	2	24.32	10.05	0.090
进给速度	46.27	2	23.14	9.56	0.094
切削深度	23.08	2	11.54	4.77	0.173
误差	4.84	2	2.42

Table 2

Table 3

Fig.6

Fig.7

Fig.8

Fig.9

Table 4

Description of variables and functions

符号	描述
$u (v q)$	优化模型函数
$v q$	$v q ∈ v$ 为参数空间中第 $q$ 个点
$v$	$v = v q q = 1 J$ 为 $J$ 个点组成的优化参数空间
$P I$	（Probability of Improvement，PI），指定的采集函数
$T$	优化模型循环次数
$D$	$D = v q, k q q = 1 J$ 构成的矩阵
$k q$	$k q = u (v q)$ 为 $v q$ 对应的优化模型函数值
$k$	$k = k q q = 1 J$ 为 $J$ 个函数值构成的向量
$G P = p (k v, D)$	$k$ 遵循的高斯先验分布
$v q = a r g m a x v ∈ v P I (v, p (k v, D))$	求 $v$ 在高斯模型中的后验分布，并返回使其最大化的 $v q$

Table 4

Table 5

References 32

1	彭振龙，张翔宇，张德远. 航空航天难加工材料高速超声波动式切削方法［J］. 航空学报， 2022， 43（4）： 525587.
	PENG Z L， ZHANG X Y， ZHANG D Y. High-speed ultrasonic vibration cutting for difficult-to-machine materials in aerospace field［J］. Acta Aeronautica et Astronautica Sinica， 2022， 43（4）： 525587 （in Chinese）.
2	万鹏，李迎光，刘长青，等. 基于域对抗门控网络的变工况刀具磨损精确预测方法［J］. 航空学报， 2021， 42（10）： 524879.
	WAN P， LI Y G， LIU C Q， et al. Method for accurate prediction of tool wear under varying cutting conditions based on domain adversarial gating neural network［J］. Acta Aeronautica et Astronautica Sinica， 2021， 42（10）： 524879 （in Chinese）.
3	张晟斐，李天梅，胡昌华，等. 基于深度卷积生成对抗网络的缺失数据生成方法及其在剩余寿命预测中的应用［J］. 航空学报， 2022， 43（8）： 225708.
	ZHANG S F， LI T M， HU C H， et al. Missing data generation method and its application in remaining useful life prediction based on deep convolutional generative adversarial network［J］. Acta Aeronautica et Astronautica Sinica， 2022， 43（8）： 225708 （in Chinese）.
4	刘斌，许靖，霍美玲，等. 基于多尺度自适应注意力网络的剩余寿命预测［J］. 航空学报， 2023， 44（5）： 226918.
	LIU B， XU J， HUO M L， et al. Remaining useful life prediction based on multi-scale adaptive attention network［J］. Acta Aeronautica et Astronautica Sinica， 2023， 44（5）： 226918 （in Chinese）.
5	BAGRI S， MANWAR A， VARGHESE A， et al. Tool wear and remaining useful life prediction in micro-milling along complex tool paths using neural networks［J］. Journal of Manufacturing Processes， 2021， 71： 679-698.
6	常钰，魏乐，王瑞祥. 数据驱动的刀具剩余寿命预测方法综述［J］. 电子技术与软件工程， 2020 （19）： 193-197.
	CHANG Y， WEI L， WANG R X. Summary of data-driven prediction methods for tool remaining life［J］. Electronic Technology & Software Engineering， 2020（19）： 193-197 （in Chinese）.
7	ZHU K P， ZHANG Y. A generic tool wear model and its application to force modeling and wear monitoring in high speed milling［J］. Mechanical Systems and Signal Processing， 2019， 115： 147-161.
8	GOMATHI K， BALAJI A. Tool condition monitoring of PCB milling machine based on vibration analysis［J］. Materials Today： Proceedings， 2021， 45： 3386-3397.
9	COOPER C， WANG P， ZHANG J J， et al. Convolutional neural network-based tool condition monitoring in vertical milling operations using acoustic signals［J］. Procedia Manufacturing， 2020， 49： 105-111.
10	QIANG B Y， SHI K N， LIU N， et al. Application of cutting power consumption in tool condition monitoring and wear prediction based on Gaussian process regression under variable cutting parameters［J］. The International Journal of Advanced Manufacturing Technology， 2023， 124（1）： 37-50.
11	刘胜辉，张人敬，张淑丽，等. 基于深度神经网络的切削刀具剩余寿命预测［J］. 哈尔滨理工大学学报， 2019， 24（3）： 1-8.
	LIU S H， ZHANG R J， ZHANG S L， et al. Prediction of remaining life of cutting tool based on DNN［J］. Journal of Harbin University of Science and Technology， 2019， 24（3）： 1-8 （in Chinese）.
12	ZHU T， WANG W B， YU M. A novel hybrid scheme for remaining useful life prognostic based on secondary decomposition，BiGRU and error correction［J］. Energy， 2023， 276： 127565.
13	SHEN B， GUI Y F， CHEN B， et al. Application of spindle power signals in tool condition monitoring based on HHT algorithm［J］. The International Journal of Advanced Manufacturing Technology， 2020， 106（3）： 1385-1395.
14	ZHU Y M， WU J C， WU J， et al. Dimensionality reduce-based for remaining useful life prediction of machining tools with multisensor fusion［J］. Reliability Engineering & System Safety， 2022， 218： 108179.
15	WU J， SU Y H， CHENG Y W， et al. Multi-sensor information fusion for remaining useful life prediction of machining tools by adaptive network based fuzzy inference system［J］. Applied Soft Computing， 2018， 68（C）： 13-23.
16	KUMAR A， CHINNAM R B， TSENG F. An HMM and polynomial regression based approach for remaining useful life and health state estimation of cutting tools［J］. Computers and Industrial Engineering， 2019， 128（C）： 1008-1014.
17	LI H， WANG W， LI Z W， et al. A novel approach for predicting tool remaining useful life using limited data［J］. Mechanical Systems and Signal Processing， 2020， 143： 106832
18	LI D H， LI Y G， LIU C Q. Gaussian process regression model incorporated with tool wear mechanism［J］. Chinese Journal of Aeronautics， 2022， 35（10）： 393-400.
19	LIU C F， ZHU L D. A two-stage approach for predicting the remaining useful life of tools using bidirectional long short-term memory［J］. Measurement， 2020， 164： 108029.
20	AN Q L， TAO Z R， XU X W，et al. A data-driven model for milling tool remaining useful life prediction with convolutional and stacked LSTM network［J］. Measurement， 2020， 154： 107461.
21	ZHU Y M， ZI Y Y， XU J， et al. An unsupervised dual-regression domain adversarial adaption network for tool wear prediction in multi-working conditions［J］. Measurement， 2022， 200： 111644.
22	ZHOU J T， ZHAO X， GAO J. Tool remaining useful life prediction method based on LSTM under variable working conditions［J］. The International Journal of Advanced Manufacturing Technology， 2019， 104（9）： 4715-4726.
23	蔡伟立，胡小锋，刘梦湘. 基于迁移学习的刀具剩余寿命预测方法［J］. 计算机集成制造系统， 2021， 27（6）： 1541-1549.
	CAI W L， HU X F， LIU M X. Prediction method of tool remaining useful life based on transfer learning［J］. Computer Integrated Manufacturing Systems， 2021， 27（6）： 1541-1549 （in Chinese）.
24	XIAO Y Q， WANG J Q， HE Z M， et al. Deep transfer learning with metric structure for fault diagnosis［J］. Knowledge-Based Systems， 2022， 256： 109826.
25	王妍，胡小锋. 基于深度迁移学习的刀具剩余寿命预测［J］. 组合机床与自动化加工技术， 2022（8）： 133-136.
	WANG Y， HU X F. Prediction of tool remaining useful life based on deep transfer learning［J］. Modular Machine Tool & Automatic Manufacturing Technique， 2022（8）： 133-136 （in Chinese）.
26	张倩，李海港，李明，等. 基于多源动态TrAdaBoost的实例迁移学习方法［J］. 中国矿业大学学报， 2014， 43（4）： 713-720.
	ZHANG Q， LI H G， LI M，et al. Instance-based transfer learning method using multi-source dynamic TrAdaBoost［J］. Journal of China University of Mining & Technology， 2014， 43（4）： 713-720 （in Chinese）.
27	DAI W Y， YANG Q， XUE G R， et al. Boosting for transfer learning［C］∥ Proceedings of the 24th International Conference on Machine Learning. New York： ACM， 2007： 193-200.
28	PARDOE D， STONE P. Boosting for regression transfer［C］∥ Proceedings of the 27th International Conference on International Conference on Machine Learning. New York： ACM， 2010： 863-870.
29	QIANG B Y， SHI K N， LIU N， et al. Integrating physics-informed recurrent Gaussian process regression into instance transfer for predicting tool wear in milling process［J］. Journal of Manufacturing Systems， 2023， 68： 42-55.
30	International Organization for Standardization. Tool life testing in end milling：［S］. 1989.
31	LIU Z Y， GUO Y B， SEALY M P， et al. Energy consumption and process sustainability of hard milling with tool wear progression［J］. Journal of Materials Processing Technology， 2016， 229： 305-312.
32	SHI K N， ZHANG D H， LIU N， et al. A novel energy consumption model for milling process considering tool wear progression［J］. Journal of Cleaner Production， 2018， 184： 152-159.

实验编号	主轴转速/ （r·min^-1）	进给速度/ （mm·min^-1）	切削深度/ mm
C1	1 500	200	0.6
C2	1 500	300	1.0
C3	1 500	400	0.8
C4	2 000	200	0.8
C5	2 000	300	0.6
C6	2 000	400	1.0
C7	2 500	200	1.0
C8	2 500	300	0.8
C9	2 500	400	0.6

算法		平均绝对误差/%			最大误差/%
算法		RGPR算法	LSTM算法	TL算法	RGPR算法	LSTM算法	TL算法
方案1	C1	2.80	3.41	0.45	4.22	7.42	1.83
	C6	14.27	12.56	1.24	17.11	16.59	2.62
	C8	5.85	9.72	0.75	8.88	12.69	1.80
方案2	C1	4.82	6.44	0.27	5.87	9.71	0.87
	C2	7.14	21.36	0.54	14.47	32.62	1.26
	C3	27.35	25.34	7.65	31.95	37.26	15.75
方案3	C1	5.82	7.82	1.07	17.45	16.25	2.53
	C2	9.07	17.29	2.05	26.28	36.31	4.67
	C3	32.05	34.60	11.60	43.15	40.16	23.55

[1]	ZHANG Xian, WANG Hongli. Failure Rate Time Series Prediction Based on Support Vector Empirical Mode Decomposition [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2011, 32(3): 480-487.
[2]	Zhang Xian;Wang Hongli. Condition Time Series Prediction Using Least Squares Support Vector Machinewith Adaptive Embedding Dimension [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2010, 31(12): 2309-2314.

Instance transfer for tool remaining useful life prediction cross working conditions

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 14

References 32

Related Articles 2

Recommended Articles

Metrics

Comments