Fault diagnosis method for complex feeding and ramming mechanisms based on SAE-ACGANs with unbalanced limited training data
YAN Xiaojia1,LIANG Weige1,ZHANG Gang2,SHE Bo1,TIAN Fuqing1
1.College of Weaponry Engineering, Naval University of Engineering, Wuhan 430033, China;
2.Missile and Gun Department, Dalian Naval Academy, Dalian 116000, China
Abstract:The problem of imbalance and limited training data is the key factor that restricts the application effect of deep learning technology in the field of fault diagnosis of complex feeding and ramming mechanism. In order to overcome the shortcomings of traditional deep learning methods that are difficult to obtain the internal distribution of limited data and the defect that traditional unbalanced data processing methods do not consider the equalization of category information, a fault diagnosis method for complex feeding and ramming mechanism based on wavelet time-frequency diagram and SAE-ACGANs is proposed. Firstly, the continuous wavelet transform (CWT) is performed on the vibration signal of the feeding and ramming mechanism to obtain a two-dimensional time-frequency diagram reflecting the time-frequency characteristics of the signal. Then, the sparse encoder in the model is used to extract image features and merge them with category information into hidden variables. It is used to strengthen the ability of latent variables to represent the characteristics related to the category of the image. After that, the generator maps the fused latent variables to generated samples similar to the real sample distribution to expand the training data set. The discriminator mines effective depth features from the extended data set and realizes the judgment of the authenticity and category of the sample. Finally, through adversarial learning and training mechanism, the optimized generator and discriminator alternately optimize each other to achieve the Nash balance. The method improves the sample generation quality and fault judgment ability with a unbalanced limited training data. The research results of the test bench for complex feeding and ramming mechanism show that: The SAE-ACGANs framework can fully learn the internal distribution and depth characteristics of the input samples. Compared with the original ACGANs framework, the method improves the performance of the discriminator, and realizes the improvement of model convergence speed, training accuracy and stability. Compared with traditional unbalanced data processing algorithms, the model's ability to identify minority fault samples is greatly improved.
闫啸家1,梁伟阁1,张钢2,佘博1,田福庆1. 非均衡小样本条件下基于SAE-ACGANs的复杂供输机构故障诊断方法[J]. 振动与冲击, 2023, 42(2): 89-99.
YAN Xiaojia1,LIANG Weige1,ZHANG Gang2,SHE Bo1,TIAN Fuqing1. Fault diagnosis method for complex feeding and ramming mechanisms based on SAE-ACGANs with unbalanced limited training data. JOURNAL OF VIBRATION AND SHOCK, 2023, 42(2): 89-99.
[1] 李 伟,马吉胜,狄长春,等. 考虑参数随机性的供输弹系统动力学及动作可靠性仿真研究[J]. 兵工学报,2012, 33(6): 747-752.
LI Wei, MA Ji-sheng, DI Chang-chun, et al. Simulation research on dynamics of ramming system and action reliability considering the randomness of the parameters[J]. Acta Armamentarii, 2012, 33(6): 747-752.
[2] 潘宏侠,张玉学. 基于SST时频图纹理特征的供输弹系统故障诊断[J]. 振动与冲击,2020, 39(6): 132-137,175.
PAN Hong-xia, ZHANG Yu-xue. Fault diagnosis of the ammunition supply system based on the texture features of SST time-frequency distribution image. [J]. Journal of Vibration and Shock, 2020, 39(6): 132-137,175.
[3] 袁建虎,韩 涛,唐 建,等. 基于小波时频图和CNN的滚动轴承智能故障诊断方法[J]. 机械设计与研究,2017, 33(2): 93-97.
YUAN Jian-hu, HAN Tao, TANG Jian, et al. An approach to intelligent fault diagnosis of rolling bearing using wavelet time-frequency representations and CNN[J]. Machine Design & Research, 2017, 33(2): 93-97.
[4] 李 恒,张 氢,秦仙蓉,等. 基于短时傅里叶变换和卷积神经网络的轴承故障诊断方法[J]. 振动与冲击,2018, 37(19): 124-131.
LI Heng, ZHANG Qin, QIN Xian-rong, et al. Fault diagnosis method for rolling bearings based on short-time Fourier transform and convolution neural network[J]. Journal of Vibration and Shock, 2018, 37(19): 124-131.
[5] 王晓龙,唐贵基. 一种基于连续小波变换的滚动轴承早期故障诊断新方法[J]. 推进技术,2016, 37(8): 1431-1437.
WANG Xiao-long, TANG Gui-ji. A New Diagnosis Method Based on ContinuousWavelet Transform for Incipient Fault of Rolling Bearing[J]. Journal of Propulsion Technology, 2016, 37(8): 1431-1437.
[6] CHENG Y, LIN M, WU J, et al. Intelligent fault diagnosis of rotating machinery based on continuous wavelet transform-local binary convolutional neural network[J]. Knowledge-Based Systems, 2021, 216(1): 106796.
[7] 刘洋,刘洋,许立雄,等. 计及数据类别不平衡的海量用户负荷典型特征高性能提取方法[J]. 中国电机工程学报,2019, 39(14): 4093–4104.
Liu Yang, Liu Yang, Xu Lixiong, et al. A high performance extraction method for massive user load typical characteristics considering data class imbalance[J]. Proceedings of the CSEE, 2019, 39(14): 4093-4104(in Chinese).
[8] GOODFELLOW I, POUGET-ABADIE J, MIRZA M, et al. Generative Adversarial Nets[J]. MIT Press, 2014.
[9] ODENA A, OLAH C, SHLENS J. Conditional Image Synthesis With Auxiliary Classifier GANs[J]. Proceedings of the 34th International Conference on Machine Learning, 2017, PMLR 70: 2642-2651.
[10] 孙灿飞,王友仁,夏裕彬. 基于SCAE-ACGAN的直升机行星齿轮裂纹故障诊断[J]. 振动、测试与诊断,2021, 41(3): 495-502, 620-621.
SUN Can-fei, WANG You-ren, XIA Yu-bin. Fault diagnosis of helicopter planetary gear crack based on SCAE-ACGAN[J]. Journal of Vibration,Measurement & Diagnosis, 2021, 41(3): 495-502, 620-621.
[11] 牛伟宇,许 华,刘英辉,等. 基于PACGAN与差分星座轨迹图的辐射源个体识别[J]. 信号处理,2021, 37(8): 1559-1567.
NIU Wei-yu, XU Hua, LIU Ying-hui, et al. Individual identification method based on PACGAN and differential constellation trace figure[J]. Journal of Signal Processing, 2021, 37(8): 1559-1567.
[12] 朱克凡,王杰贵,刘有军. 小样本条件下基于数据增强和WACGAN的雷达目标识别算法[J]. 电子学报,2020, 48(6): 1124-1131.
ZHUKe-fan, WANG Jie-gui, LIU You-jun. Radar target recognition algorithm based on data augmentation and WACGAN with a limited training data[J]. Acta Electronica Sinica, 2020, 48(6): 1124-1131.
[13] 李恒辉,郭 交,韩文霆,等. 栈式稀疏自编码网络的多时相全极化SAR散射特征降维[J]. 遥感学报,2020, 24(11): 1379-1391.
Li Heng-hua, Guo Jiao, Han Wen-ting, et al. Scattering feature dimension reduction of multitemporal fully PolSAR image based on Stacked Sparse AutoEncoder[J]. Journal of Remote Sensing, 2020, 24(11): 1379-1391.
[14] ZHANG W, SHAN S, CHEN X, et al. Local Gabor Binary Patterns Based on Kullback-Leibler Divergence for Partially Occluded Face Recognition[J]. IEEE Signal Processing Letters, 2007, 14(11): 875-878.
[15] 赵 川,冯志鹏. 时变工况下行星轮轴承特征分布拟合与智能故障诊断[J]. 振动与冲击,2021, 40(14): 252-260.
ZHAO Chuan, FENG Zhi-peng. Features distribution fitting and intelligent fault diagnosis of planet bearings under time-varying condition[J]. Journal of Vibration and Shock, 2021, 40(14): 252-260.
[16] 刘云鹏,许自强,和家慧,等. 基于条件式Wasserstein生成对抗网络的电力变压器故障样本增强技术[J]. 电网技术,2020, 44(4): 1505-1513.
LIU Yun-peng, XU Zi-qiang, HE Jia-hui, et al. Data augmentation method for power transformer fault diagnosis based on conditional wasserstein generative adversarial network[J]. Power System Technology, 2020, 44(4): 1505-1513.
[17] 朱 敏,刘 奇,刘 星,等. 基于LMKL和OC-ELM的航空电子部件故障检测方法[J]. 系统工程与电子技术,2020, 42(6): 1424-1432.
ZHU Min, LIU Qi, LIU Xing, et al. Fault detection method for avionics based on LMKL and OC-ELM[J]. Systems Engineering and Electronics, 2020, 42(6): 1424-1432.
[18] ZHANG J, CHEN L. Clustering-based undersampling with random over sampling examples and support vector machine for imbalanced classification of breast cancer diagnosis[J]. Computer Assisted Surgery, 2019, 24(sup2): 62-72.
[19] YU D-J, HU J, TANG Z-M, et al. Improving protein-ATP binding residues prediction by boosting SVMs with random under-sampling[J]. Neurocomputing, 2013, 104: 180-190.
[20] 李睿峰,许爱强,孙伟超,等. 基于样本重采样的电路非平衡数据预处理方法[J]. 系统工程与电子技术,2020, 42(11): 2654-2660.
LI Rui-feng, XU Ai-qiang, SUN Wei-chao, et al. Preprocessing method based on sample resampling for imbalanced data of electronic circuits[J]. Systems Engineering and Electronics, 2020, 42(11): 2654-2660.
[21] 陈 晓,李 新,陈 鹏,等. 基于ADASYN-随机森林的安卓恶意应用检测研究[J]. 信息技术与信息化,2020(12): 89-92.
CHEN Xiao, LI Xin, CHEN Peng, et al. Research on android malicious application detection based on ADASYN-Random forest[J]. Information Technology and Informatization, 2020(12): 89-92.