考虑系统不确定性的舰船壳体结构中低频相关激励源识别

Ship hull structure plays a crucial role in both source excitation and vibro-acoustical transmission of ship noise except the direct radiated sound including the ship propeller noise, the fluid noise and so on. Of ship hull structure, multi-path vibration transmission is produced by the presence of characteristic spectrum overlap and ship hull structural mode coupling of both machinery vibration and the hull vibration excited by ship propeller and shaft system in low- and medium-frequency range, which causes low- and medium- frequency coupled vibration excited by machine(ry)-machine(ry) and the ship propeller and shaft system. When considering that the uncertainties of the model，system parameters and the low- and medium-frequency boundary conditions obey Independent and Identically Distributed (IID) Gauss distribution，a statistical analysis model of multi-source- correlation excitation and multi-path coupling vibration system is established for the above ship hull structure, and then eigenanalysis, response statistical analysis, transfer path analysis and excitation source analysis of random dynamical system are carried out on the above model. Therewith, features of correlated source statistics of the system are obtained by combining theoretical analysis, numerical simulation and model test. This project whose aim is, to propose a statistical-analysis-model-based random eigenanalysis method, and to establish the random dynamic stiffness/flexibility matrix that reveal the reverse/forward transmissibility of transfer path of random coupled vibration system, and then to present new ideas for identifying correlated sources from the statistical viewpoint based on statistically analyzing the random dynamic stiffness/flexibility matrix and spatially decoupling the elements of that matrix, will open up a new way for finding out, separating and identifying ship hull structural noise source, and promote the improvement of vibration and noise reduction technology.

在舰船噪声中，除螺旋桨、流体激励直接辐射声等之外，舰船壳体结构在其声源激励及声振传递中均表现极其重要作用。针对舰船壳体结构，中低频段内机械振动及桨轴激励船体振动在特征频谱上的叠合以及与船体结构的模态耦合，通过振动多途传递，形成机（械）-机（械）及桨轴激励船体中低频耦合振动。考虑模型、参数及中低频段内边界条件不确定性服从IID Gauss分布，建立其多源相关多通道耦合振动统计分析模型，继而进行随机系统本征分析、响应统计分析、通道传递分析及激励源分析，结合理论分析、数值仿真及模型试验，获取系统相关激励源统计特性。本课题旨在基于统计分析模型提出随机本征分析方法；建立随机动刚度/柔度矩阵，明确随机耦合振动系统逆/正向通道传递关系；基于随机动刚度/柔度矩阵统计分析及其矩阵元素解耦，提出以统计观点识别系统相关激励源的新思路；为查明、分离和识别舰船壳体结构噪声开辟新途径，推动减振降噪技术的发展。

针对舰船壳体结构机（械）-机（械）及桨轴激励船体的多源激励多通道耦合振动，考虑系统边界条件具有不确定性时，采用附加边界刚度传递边界条件I.I.D. Gauss分布随机特征，建立其Gauss边界动力系统，建立利用微观物理学理论解决宏观随机动力系统中低频振动问题的纽带。基于RMT理论，构建CCGE（列约束Gauss随机矩阵/系综）模型特征矩阵，在中低频段基于Saint-Venant原理的预应力公式发展DS方法估计固有频率PDF，高频段比拟求解Fokker-Planck方程给出固有频率JPD；而针对本征函数引入“权参数”概念，基于多重分形理论获取其一致性条件分布律及对应的模态振型概率模型，将宏观动力系统分析基础性研究向微观物理学领域探索。在获得固有频率、模态振型的统计特征基础上，基于随机DSM/DCM矩阵建立FEM节点FRF矩阵，基于FRF均值、方差的幅频特性及其样本各频点PDF等描述正/逆向传递通道统计特征，实现矩阵元素（通道）统计解耦；在中低频段提出PCE方法分析解耦通道统计特征；在高频段基于本征函数多重分形的Green函数分析得到通道传递统计特性，拓展振动噪声控制技术研究方向。基于随机动力系统响应、通道传递和激励源等统计分析提出中低频结构噪声源识别策略，结合基于自适应广义线性混合模型的源分离算法，并进行节点FRF矩阵相关/（偏）相干分析，获取激励源位置均值分布及其统计特征，初步以统计观点识别随机系统相关激励源，作为振动噪声源分离方法的补充与发展，将指导舰船减振降噪措施的正确实施。对于Gauss边界动力系统，在理论分析、数值仿真基础上，建立舱段单层圆柱壳结构试验模型，开展随机边界条件下舰船壳体结构振动试验，为研究具有模型、参数、边界条件等不确定性的舰船壳体结构减振降噪提出一种试验条件，探索在实船海试试验中降低对海况、辅机运行工况的苛刻条件并减少海试舰船航次的新技术途径。