基于逆传感网络模型辨识的激波管动态压力重构

摘要
图/表
参考文献(17)
相关文章 (15)

全文: PDF (3064 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要提出一种基于逆传感网络模型辨识的激波管动态压力重构方法。首先，基于经验模态分解，将压力传感器动态响应信号分解为一系列不同频带的分量；然后，采用相关系数和振铃幅值占比指标，实现振铃分量和趋势分量的识别，构建逆传感网络模型训练集和测试集；最后，基于双向长短期记忆神经网络训练及测试，建立压力传感器逆传感网络模型，实现激波管动态压力重构。分别通过仿真实验和激波管测量实验验证本文方法对于动态压力重构的性能。仿真实验结果显示，重构的动态压力信号的RMSE和MAPE远小于传统趋势估计法，其值比LSTM方法得到的结果分别减小了2倍和5倍，并且在不同阶数的压力传感器仿真实验中验证了本文方法的鲁棒性，通过对比不同压力传感器系统下该方法的动态压力重构精度，验证了本文方法的适用性；激波管测量实验结果显示，模型训练和测试输出的RMSE和MAPE分别为0.0016 V、0.0036%和0.0025 V、0.062%，重构得到的激波管动态压力在平稳区间内的平均相对误差约为2.14%。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	李永生
	姚贞建
	刘臣
	丁义凡

关键词 ：激波管, 压力传感器, 经验模态分解, 双向长短期记忆神经网络, 动态压力

Abstract：A reconstruction method of shock tube dynamic pressure is proposed based on inverse sensing network model identification. First, the dynamic response signal of pressure sensor is decomposed into a series of components with different frequency bands based on empirical mode decomposition. Then, the ringing component and trend component are identified by correlation coefficient and ring amplitude ratio indexes to construct the training set and test set of the inverse sensing network model. Finally, the inverse sensing network model of pressure sensor is established based on the training and testing of Bi-directional long short-term memory neural network, and the shock tube dynamic pressure is reconstructed. Simulation experiments and shock tube measurement experiments are conducted to verify the performance of the proposed method. The simulation results show that the RMSE and MAPE of the reconstructed dynamic pressure signal are far less than those obtained by the traditional trend estimation method, and the values are reduced by 2 and 5 times respectively compared with those obtained by the LSTM method. The robustness of the proposed method is verified in the simulation experiments of pressure sensors with different orders. The applicability of this method is verified by comparing the accuracy of dynamic pressure reconstruction for different pressure sensors. The experimental results show that the RMSE and MAPE of the model training and test output are 0.016V and 0.0036% as well as 0.0025V and 0.062%, respectively. The average relative error of the reconstructed dynamic pressure is about 2.14% in the stable interval.

Key words： shock tube pressure sensor Empirical Mode Decomposition Bi-directional Long Short-Term Memory neural network dynamic pressure

收稿日期: 2022-11-15 出版日期: 2023-12-15

引用本文:

李永生，姚贞建，刘臣，丁义凡. 基于逆传感网络模型辨识的激波管动态压力重构[J]. 振动与冲击, 2023, 42(23): 223-230.
LI Yongsheng,YAO Zhenjian,LIU Chen,DING Yifan. Dynamic pressure reconstruction of shock tube based on inverse sensing network model identification. JOURNAL OF VIBRATION AND SHOCK, 2023, 42(23): 223-230.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2023/V42/I23/223

[1] 姚贞建, 王中宇, 王辰辰, 等. 压力传感器动态特性参数不确定度评定[J]. 北京航空航天大学学报, 2018, 44(8): 1672-1681.
YAO ZJ, WANG ZY, WANG CC, et al. Uncertainty evaluation for dynamic characteristic parameters of pressure sensors[J]. Journal of Beijing University of Aeronautics and Astronautics, 2018, 44(8): 1672-1681.
[2] Yao ZJ, Li YS, Ding YF, et al. Improved shock tube method for dynamic calibration of the sensitivity characteristic of piezoresistive pressure sensors[J]. Measurement, 2022, 196: 111271.
[3] 杨军, 薛斌. 激波管管长对阶跃压力波形的影响分析[J]. 振动与冲击, 2019, 38(03), 252-257.
YANG J, XUE B. Effects of shock tube length on step pressure waveform[J]. Journal of Vibration and Shock, 2019, 38(03), 252-257.
[4] 杨凡, 孔德仁, 姜波, 等. 基于激波管校准的冲击波压力传感器动态特性研究[J]. 南京理工大学学报, 2017, 41(3): 330-336.
YANG F, KONG DR, JIANG B, et al. Dynamic characteristic of shock wave pressure sensor based on shock tube calibration[J]. Journal of Nanjing University of Science and Technology, 2017, 41(3): 330-336.
[5] YAO ZJ, DING YF, LI YS, et al. Experimental investigation of pressure characteristics behind the reflected shock waves in a metrological shock tube[J]. Metrologia,2022, 59: 055008.
[6] KNOTT J, ROBINSON A. Dynamic characterization of pressure transducers using shock tube methods[J]. Transactions of the Institute of Measurement and Control, 2020, 42(4): 743-748.
[7] Yao ZJ, Liu XJ, Wang CC, et al. Improve traceable measurement of the reflected step pressure in shock tube with the compensation of shock wave attenuation[J]. Aerospace Science and Technology, 2020, 107: 106302.
[8] 赵博, 李鹤. 结合EMD和LSF的振动信号降噪方法的研究[J]. 振动、测试与诊断, 2022, 42(3): 606-610+624.
ZHAO B, LI H. Noise Reduction Method of Vibration Signal Combining EMD and LSF[J]. Journal of Vibration, Measurement and Diagnosis, 2022, 42(3): 606-610+624.
[9] 余腾, 胡伍生, 吴杰, 等. 基于小波阈值去噪与EMD分解方法提取润扬大桥振动信息[J]. 振动与冲击, 2019, 38(12): 264-270.
YU T, HU WS, WU J, et al. Exteaction of Runyang bridge vibration information based on a fusion method of wavelet threshold denoising and EMD decomposition[J]. Journal of Vibration and Shock, 2019, 38(12): 264-270.
[10] SHEN SL, NJOCK PGA, ZHOU AN, et al. Dynamic prediction of jet grouted column diameter in soft soil using Bi-LSTM deep learning[J]. Acta Geotechnica, 2020, 16(1): 303-315.
[11] ZHENG X, CHEN WZ. An Attention-based Bi-LSTM Metho0d for Visual Object Classification via EEG[J]. Biomedical Signal Processing and Control, 2021, 63: 102174.
[12] 赵志宏, 赵敬娇, 魏子洋. 基于BiLSTM的滚动轴承故障诊断研究[J]. 振动与冲击, 2021, 40(1): 95-101.
ZHAO ZH, ZHAO JJ, WEI ZY. Rolling bearing fault diagnosis based on BiLSTM network[J]. Journal of Vibration and Shock, 2021, 40(1): 95-101.
[13] DU SD, LI TR, YANG Y, et al. Multivariate time series forecasting via attention-based encoder-decoder framework[J]. Neurocomputing, 2020, 388: 296-279.
[14] YU Y, SI XS, HU CG, et al. A Review of Recurrent Networks: LSTM Cells and Network Architectures[J]. Neural Computation, 2019, 31(7): 1235-1270.
[15] ZHANG N, SHEN SL, ZHOU A, et al. Application of LSTM approach for modelling stress-strain behaviour of soil[J]. Applied Soft Computing, 2021, 100: 106959.
[16] 李永生, 姚贞建, 丁义凡. 激波管反射阶跃压力上升时间高精度估计[J]. 仪器仪表学报, 2022, 43(9): 39-47.
LI YS, YAO ZJ, DING YF. High-accuracy estimation of the rise time of shock tube reflection step pressure[J]. Chinese Journal of Scientific Instrument, 2022, 43(9): 39-47.
[17] YAO ZJ, LIU XJ, YANG WJ, et al. A coarse-to fine denoising
method for dynamic calibration signals of pressure sensor based on adaptive mode decompositions[J]. Measure, 2020, 163: 107935.