﻿ 船舶结构监测系统的应力预测研究
«上一篇
 文章快速检索 高级检索

 应用科技  2018, Vol. 45 Issue (6): 8-11, 16  DOI: 10.11991/yykj.201803009 0

### 引用本文

WU Shuang, JIAO Shuhong. Prediction of stress pattern in the ship structure monitoring system[J]. Applied Science and Technology, 2018, 45(6), 8-11, 16. DOI: 10.11991/yykj.201803009.

### 文章历史

Prediction of stress pattern in the ship structure monitoring system
WU Shuang, JIAO Shuhong
College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China
Abstract: Traditionally, the ship navigation decision-making depends too much on the decision-making experience. The aided decision-making can help decision-makers better analyze, evaluate and formulate plans in the decision-making process, the prediction research of ship structure monitoring system is a prerequisite for auxiliary decision-making in ships, so this part is the focus of research. First of all, the choice and layout scheme of the monitoring sensor for ship structure stress was discussed. Then, in view of the nonlinear and non-stationary signal characteristics of ship structure stress sequence, a structure stress prediction model was proposed by combination of the complementary ensemble empirical mode decomposition (CEEMD) and the support vector machine (SVM) that optimizes parameter through improved grid search method. Finally, a variety of test samples were used to verify the reliability and accuracy of the prediction model. Through verification, the proposed structure stress prediction model has high accuracy under different sea conditions.
Keywords: intelligent ship    monitor    sensor    non-stationary signal    CEEMD    SVM    prediction model    grid search method

1 传感器的选择及布置 1.1 传感器的选择

1.2 传感器的布置

2 基本理论 2.1 CEEMD

CEEMD分解建立在EMD分解的基础上，主要包括3个步骤：

1）向原始信号中加入n组正负成对的辅助白噪声，这样就形成了2n个合成的信号。

 $\left[ {\begin{array}{*{20}{c}} {{S_1}} \\ {{S_2}} \end{array}} \right] = \left[ {\begin{array}{*{20}{c}} 1&1 \\ 1&{ - 1} \end{array}} \right]\left[ {\begin{array}{*{20}{c}} X \\ N \end{array}} \right]$

2）对2n个信号分别进行EMD分解，每个信号得到一组IMF分量，其中第j个信号的第k个IMF分量可表示为 ${C_{jk}}$

3）通过多组分量合成方法得到原信号的第k个IMF的大小为

 ${C_k} = \frac{1}{{2n}}\sum\limits_{j = 1}^{2n} {{C_{jk}}}$

2.2 SVM的参数优化方法改进

1）K次交叉验证法

2）改进的网格搜索法

2.3 CEEMD和SVM的预测模型

1）将原始序列先经过CEEMD的分解，得到若干IMF分量，获取对原信号贡献率最大的IMF分量，然后对各IMF分量进行提取，作为训练样本。

2)对SVM的核函数、错误惩罚参数C和核函数宽度G等参数进行选择。

a）线性核函数

 ${e_{\max }} = \max \left| {\frac{{{Y_i} - {Y_i}'}}{{{Y_i}}}} \right|$

b）多项式核函数（Polynomial）

 $K\left( {{x_i},{x_j}} \right) = {\left[ {\left( {{x_i},{x_j}} \right) + 1} \right]^d}$

 $K\left( {{x_i},{x_j}} \right) = \exp \left( { - \frac{{{{\left\| {{x_i} - {x_j}} \right\|}^2}}}{{2{\delta ^2}}}} \right)$

d）双曲正切核函数（Sinmoid）

 $K\left( {{x_i},{x_j}} \right) = \tanh \left[ {a\left( {{x_i} \cdot {x_j}} \right) + b} \right]$

RBF核函数的特征空间维数为无限维，并且其搜索查找为全网络搜索，当输入样本个数有限时，其特征空间为线性可分，在没有先验信息时，比其他核函数的总体性更好。综上，SVM的核函数选择RBF对样本数据进行非线性映射。

3）利用改进的基于K次交叉验证法的网格搜索法寻找最优的CG

4）在确定核函数、CG之后，进行预测模型的训练。

5）利用测试样本建立测试模型。具体预测模型的示意图如图2所示。

 ${e_{{\rm{map}}}} = \frac{1}{m}\sum\limits_{i = 1}^m {\frac{{{Y_i} - {Y_i}'}}{{{Y_i}}}}$
 ${e_{\max }} = \max \left| {\frac{{{Y_i} - {Y_i}'}}{{{Y_i}}}} \right|$
 ${e_a} = \frac{1}{m}\sum\limits_{i = 1}^m {\left( {{Y_i} - {Y_i}'} \right)}$

3 实验仿真 3.1 结构应力信号特征提取

3.2 CEEMD和SVM预测实验

3.3 实验对比及误差分析

4 结论

1）考虑船舶监测系统的实际应用需求，完成对传感器类型及布置方法的讨论，综合考虑选用管式和埋入式封装光纤光栅传感器，通过波分复用和空分复用结合的混合复用的方法来满足监测过程中精度、传输效率及传感器间干扰的问题。

2）构建CEEMD和SVM预测模型，对SVM常用的参数寻优方法——网格搜索法，进行改进。

3）利用预测模型对高、低及混合海况情况下的船舶结构应力进行预测。

4）对于构建的预测模型，从以下两方面验证其准确性：一方面通过测试样本验证模型对高、低及混合海况的适应性，实验结果表明，该模型可以对3种情况下的结构应力进行较好的预测；另一方面，通过对SVM、EMD+SVM、EMD+SVM等模型的预测结果进行对比，结果表明，CEEMD+SVM预测模型对结构应力的预测效果及预测误差要优于其他几种预测模型。

 [1] 杨帆. 舰船破舱稳性实时计算方法研究[D]. 大连: 大连理工大学, 2006: 12. (0) [2] 沙丹丹. 基于HHT和SVM的砰击检测算法研究[D]. 哈尔滨: 哈尔滨工程大学, 2015: 3. (0) [3] HUANG N E, SHEN Zheng, LONG S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the royal society A: mathematical, physical and engineering sciences, 1998, 454(1971): 903-995. DOI:10.1098/rspa.1998.0193 (0) [4] LEI Yaguo, LIN Jing, HE Zhengjia, et al. A review on empirical mode decomposition in fault diagnosis of rotating machinery[J]. Mechanical systems and signal processing, 2013, 35(1/2): 108-126. (0) [5] WU Zhaohua, HUANG N. Ensemble empirical mode decomposition: a noise-assisted data analysis method[J]. Advances in adaptive data analysis, 2008, 1(1): 1-41. (0) [6] YEH J R, SHIEH J, HUANG N. Complementary ensemble empirical mode decomposition: a novel noise enhanced data analysis method[J]. Advances in adaptive data analysis, 2010, 2(2): 135-156. DOI:10.1142/S1793536910000422 (0) [7] GUO Guodong, LI S Z. Contents-based audio classification and retrieval by support vector machines[J]. IEEE transactions on neural network, 2003, 14(1): 209-215. DOI:10.1109/TNN.2002.806626 (0) [8] 王军栋, 齐维贵. 基于EDM-SVM的江水浊度预测方法研究[J]. 电子学报, 2009(10): 2130-2133. DOI:10.3321/j.issn:0372-2112.2009.10.004 (0) [9] 张金涛. 光纤光栅传感器复用关键技术的研究与应用[D]. 合肥: 合肥工业大学, 2015. (0) [10] 曲健, 陈红岩, 刘文贞, 等. 基于改进网格搜索法的支持向量机在气体定量分析中的应用[J]. 传感技术学报, 2015, 28(5): 774-778. DOI:10.3969/j.issn.1004-1699.2015.05.027 (0) [11] QIAO Meiying, MA Xiaoping, LAN Jianyi, et al. Time-series gas prediction model using LS-SVR within a Bayesian framework[J]. Mining science and technology, 2011, 21(1): 153-157. DOI:10.1016/j.mstc.2010.12.012 (0) [12] 吕震宇. 基于ELMD和KNN分类器的船体砰击载荷检测算法研究[D]. 哈尔滨: 哈尔滨工程大学, 2016. (0)