基于离散小波变换的海洋工程数据降噪研究
发布时间：2018-07-17 16:10
【摘要】：小波变换作为一种数学工具,凭借其良好的多分辨分析能力,在诸多研究领域被广泛地应用。物理模型试验是海洋工程领域重要的研究手段之一,但在物理模型试验中采集到的数据往往会受到噪声污染,因而在进一步处理这类数据时,需要对数据信号进行降噪处理以确保试验结果的准确性。基于上述情况,本文通过基于离散小波变换的降噪方法对从软钢臂单点系泊浮式生产储油卸油装置(FPSO)物理模型试验中采集到的数据进行降噪处理,从而提高试验数据的准确性。本文首先详细介绍了离散小波变换的基本概念和理论基础,然后对小波降噪方法的理论进行了细致的介绍,重点介绍了常用的几种小波函数和降噪过程中参数的选取规则以及数据降噪性能的评价指标。随后对单点软刚臂系泊FPSO水池物理模型试验进行了简单的介绍,通过对FPSO自由衰减信号中加入已知白噪声形成带噪信号,并应用小波降噪方法对其进行分析处理来评估不同小波函数、阈值选取以及重调方式的降噪效果,从而找出适合于船舶运动量信号的最优降噪参数选取组合。结果表明,db8和sym10小波函数,Stein无偏风险阈值选择方法,sln重调方式的组合降噪效果最佳。将此种降噪组合应用于单点软钢臂系泊FPSO试验采集的运动量数据进行降噪处理,结果表明降噪效果良好。最后应用小波降噪方法对系泊力数据进行了降噪处理。系泊塔关键节点处的受力是系泊系统设计的主要依据。采用小波降噪方法对几组工况下采集到的系泊力数据进行降噪处理,并与传统的低通滤波方法进行了对比。结果表明小波阈值降噪方法可以很好地解决低通滤波方法滤去高频瞬态系泊力的问题,其降噪结果较为合理,能够反映真实的试验现象。另外通过频谱分析,发现小波降噪方法能够准确地分辨出真实信号和噪声之间的差别,保留了信号中重要的高频信息。最后,通过选取不同参数,比较其对降噪效果的影响,得到对于实测系泊力信号的最优降噪组合。
[Abstract]:Wavelet transform, as a mathematical tool, is widely used in many research fields with its good ability of multi-resolution analysis. Physical model test is one of the important research methods in the field of marine engineering, but the data collected in the physical model test are often contaminated by noise, so in the further processing of this kind of data, The noise reduction of the data signal is needed to ensure the accuracy of the test results. Based on the above situation, the noise reduction method based on discrete wavelet transform is used to reduce the noise of the data collected from the physical model test of the flexible steel arm single point mooring floating oil storage and unloading unit (FPSO), so as to improve the accuracy of the test data. In this paper, the basic concept and theoretical basis of discrete wavelet transform are introduced in detail, and then the theory of wavelet denoising method is introduced in detail. Several commonly used wavelet functions, the selection rules of parameters in the process of noise reduction and the evaluation indexes of data denoising performance are emphatically introduced in this paper. Then the physical model test of a single point flexible rigid arm mooring FPSO pool is briefly introduced. The noise signal is formed by adding known white noise to the FPSO free attenuation signal. The wavelet denoising method is applied to evaluate the noise reduction effect of different wavelet functions, threshold selection and resetting mode, so as to find the optimal combination of noise reduction parameters suitable for ship motion signal. The results show that the combination of Db8 and sym10 wavelet function Stein unbiased risk threshold selection method has the best effect on noise reduction. The noise reduction combination is applied to the noise reduction of mooring FPSO with single point soft steel arm mooring. The results show that the noise reduction effect is good. Finally, wavelet denoising method is used to deal with mooring force data. The stress at the key nodes of the mooring tower is the main basis for the design of the mooring system. The mooring force data collected under several working conditions are de-noised by wavelet denoising method and compared with the traditional low-pass filtering method. The results show that wavelet threshold denoising method can solve the problem of filtering high frequency transient mooring force by low pass filtering method. The result of noise reduction is reasonable and can reflect the real experimental phenomenon. In addition, through spectrum analysis, it is found that wavelet denoising method can accurately distinguish the difference between real signal and noise, and retain the important high-frequency information in the signal. Finally, by selecting different parameters and comparing their effects on the noise reduction effect, the optimal noise reduction combination for the measured mooring force signal is obtained.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P75
本文编号：2130198
[Abstract]:Wavelet transform, as a mathematical tool, is widely used in many research fields with its good ability of multi-resolution analysis. Physical model test is one of the important research methods in the field of marine engineering, but the data collected in the physical model test are often contaminated by noise, so in the further processing of this kind of data, The noise reduction of the data signal is needed to ensure the accuracy of the test results. Based on the above situation, the noise reduction method based on discrete wavelet transform is used to reduce the noise of the data collected from the physical model test of the flexible steel arm single point mooring floating oil storage and unloading unit (FPSO), so as to improve the accuracy of the test data. In this paper, the basic concept and theoretical basis of discrete wavelet transform are introduced in detail, and then the theory of wavelet denoising method is introduced in detail. Several commonly used wavelet functions, the selection rules of parameters in the process of noise reduction and the evaluation indexes of data denoising performance are emphatically introduced in this paper. Then the physical model test of a single point flexible rigid arm mooring FPSO pool is briefly introduced. The noise signal is formed by adding known white noise to the FPSO free attenuation signal. The wavelet denoising method is applied to evaluate the noise reduction effect of different wavelet functions, threshold selection and resetting mode, so as to find the optimal combination of noise reduction parameters suitable for ship motion signal. The results show that the combination of Db8 and sym10 wavelet function Stein unbiased risk threshold selection method has the best effect on noise reduction. The noise reduction combination is applied to the noise reduction of mooring FPSO with single point soft steel arm mooring. The results show that the noise reduction effect is good. Finally, wavelet denoising method is used to deal with mooring force data. The stress at the key nodes of the mooring tower is the main basis for the design of the mooring system. The mooring force data collected under several working conditions are de-noised by wavelet denoising method and compared with the traditional low-pass filtering method. The results show that wavelet threshold denoising method can solve the problem of filtering high frequency transient mooring force by low pass filtering method. The result of noise reduction is reasonable and can reflect the real experimental phenomenon. In addition, through spectrum analysis, it is found that wavelet denoising method can accurately distinguish the difference between real signal and noise, and retain the important high-frequency information in the signal. Finally, by selecting different parameters and comparing their effects on the noise reduction effect, the optimal noise reduction combination for the measured mooring force signal is obtained.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P75
【参考文献】
相关期刊论文 前2条
1 张一,成礼智;小波变换图像压缩中最优小波基的选取方法[J];电视技术;2004年10期
2 吕瑞兰,吴铁军,于玲;采用不同小波母函数的阈值去噪方法性能分析[J];光谱学与光谱分析;2004年07期
相关硕士学位论文 前1条
1 原泉;小波分析在船舶测试信号方面的应用研究[D];大连理工大学;2005年
，本文编号：2130198
本文链接：https://www.wllwen.com/kejilunwen/haiyang/2130198.html
