内波对水下声信号时间相关性影响的数值和实验研究

发布时间：2018-01-11 06:23

本文关键词：内波对水下声信号时间相关性影响的数值和实验研究　出处：《中国海洋大学》2015年博士论文　论文类型：学位论文

【摘要】：海洋中内波的存在会使声信号的时间相关半径变小。受物理介质影响的信号时间相关半径表明了信号由完全相关变为不相关时经过的时间尺度,它是声纳目标的检测与识别、匹配场声源定位以及水下通讯的重要技术指标。本文通过数值仿真和海上实验数据处理,分析了内波对声场时间相关性的影响,得到了典型声传播条件下信号的匹配场时间相关半径以及时间相关半径在距离-深度空间上的分布。基于抛物方程(PE)理论,Monte-Carlo数值仿真了线性内波存在时声场的时间相关函数,得到时间相关性的空间分布。在前人PE仿真所采用的冻结声学环境模型的基础上提出了一种非冻结的环境模拟方法,时间相关性的数值结果表明,在一定的频率和接收距离范围内,冻结模型依然可靠,但频率越高,接收距离越远,可靠性越低。对距离-深度空间各个位置处声场时间相关性的理论和数值研究发现,时间相关性在空间上的分布具有与声场干涉条纹类似的结构,声场会聚区的时间相关性优于影区,随机噪声的存在使这个结论更加明显。此外,声源频率f、接收距离R和声速扰动的统计标准差|δc|等参数的改变引起声场干涉结构的变化,导致时间相关半径对其不再是对数坐标下的线性依赖。在Monte-Carlo研究的同时,提出了基于绝热近似和耦合简正波理论的统计方法。研究结果表明：深远海声传播条件下,绝热近似得到的匹配场时间相关半径随R、f和|δc|的幂指数关系与Monte-Carlo仿真和单向耦合方法的结果近似相同,但它高估了信号的时间相关半径,且不能正确描述时间相关半径的空间分布；浅海近距离声传播条件下,绝热近似的计算结果具有一定的可靠性。本文发展的二维统计模型得到的匹配场时间相关半径大于前人的三维模型,在本文参数范围内,数值计算中一般考虑前5号内波模态和几十号简正波就可以保证数值结果的可靠性,但要得到时间相关半径的空间分布需要考虑更多的简正波。对ASIAEX2001实验的声学数据处理后得到,当声传播路径上出现强孤立子内波时,声场的时间相关半径明显降低,某些时刻300Hz信号的时间相关半径低于20s,500Hz信号的时间相关半径甚至降到l 0s内；信号频率越高,时间相关半径越短：且声场的时间相关半径在深度上有明显起伏。由两个位置不同的声源S1和S2发射的400Hz信号在接收位置R处的时间相关半径不同,S2-R路径的时间相关半径比S1-R路径长。这可能是由于内波在S2-R路径上的相位传播速度小于S1-R路径,声源深度和地形的不同也可能对上述结果有所影响。PE仿真采用的冻结声学环境忽略了内波在传播过程中波形和强度的变化,导致计算得到的时间相关半径比实验数据处理的结果要长。因此,建立更加完善的非冻结环境模型需要在发射与接收间布放足够多的温度链以监测内波传播过程中的演变。
[Abstract]:The existence of internal waves in the ocean will make the time correlation radius of acoustic signal smaller. The time correlation radius of the signal affected by physical medium indicates the time scale of signal from complete correlation to irrelevant. It is an important technical index of sonar target detection and recognition, matching field source location and underwater communication. In this paper, the influence of internal wave on the time dependence of sound field is analyzed by numerical simulation and offshore experimental data processing. The time correlation radius of the matching field and the distribution of the time correlation radius in the distance-depth space are obtained under the typical sound propagation condition. The theory of the parabolic equation (PE) is used to obtain the distribution of the time-dependent radius of the signal in the range-depth space. The time correlation function of sound field in the presence of linear internal wave is simulated numerically by Monte-Carlo. Based on the frozen acoustic environment model used in previous PE simulation, a non-frozen environmental simulation method is proposed, and the numerical results of time correlation show that. In a certain range of frequency and receiving distance, the freezing model is still reliable, but the higher the frequency, the longer the receiving distance. The lower the reliability is, the more the theoretical and numerical study on the temporal correlation of sound field at each position in the range-depth space shows that the spatial distribution of temporal correlation has a structure similar to that of the interference fringes of sound field. The temporal correlation of the acoustic field convergence region is better than that of the shadow area, and the existence of random noise makes this conclusion more obvious. In addition, the sound source frequency f. The change of receiving distance R and the statistical standard deviation 未 c of the acoustic velocity disturbance leads to the change of the sound field interference structure. Therefore, the time-dependent radius is no longer a linear dependency in logarithmic coordinates. A statistical method based on adiabatic approximation and coupled normal wave theory is proposed. The results show that the time-dependent radius of the matching field obtained by adiabatic approximation depends on R. The power exponent relation of f and 未 c is approximately the same as that of Monte-Carlo simulation and unidirectional coupling method, but it overestimates the time-dependent radius of the signal. And the spatial distribution of time-dependent radius can not be correctly described. The calculated results of adiabatic approximation are reliable under the condition of near distance acoustic propagation in shallow water. The time correlation radius of the matching field obtained by the two-dimensional statistical model developed in this paper is larger than that of the previous three dimensional model. In the parameter range of this paper, the reliability of the numerical results can be guaranteed by considering the first 5 internal wave mode and the dozens of normal waves in the numerical calculation. But in order to obtain the spatial distribution of time-dependent radius, we need to consider more normal waves. After processing the acoustic data of ASIAEX2001 experiment, we can get that, when the strong soliton internal waves appear on the sound propagation path. The time correlation radius of sound field decreases obviously, and the time correlation radius of 300 Hz signal at some times is lower than that of 20 s ~ 500 Hz signal, even within 10 s. The higher the signal frequency. The shorter the time-dependent radius:. The time correlation radii of sound field fluctuate obviously in depth. The time correlation radii of 400 Hz signals transmitted by two different sound sources S1 and S2 are different at the receiving position R. The time-dependent radius of the S2-R path is longer than that of the S1-R path, which may be due to the fact that the phase propagation velocity of the internal wave in the S2-R path is lower than that in the S1-R path. The difference of sound source depth and topography may also affect the above results. The frozen acoustic environment used in PE simulation ignores the variation of the waveform and intensity of the internal wave in the process of propagation. The calculated time correlation radius is longer than that of experimental data processing. To establish a more perfect non-freezing environment model, enough temperature chains should be placed between the transmitter and receiver to monitor the evolution of internal wave propagation.
【学位授予单位】：中国海洋大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：P733.2;TB56

【相似文献】