基于源/场模式映射关系的近场声全息理论与测试方法研究

Near-field acoustic holography (NAH) is a powerful tool for the analysis of vibration and noise control. To promote the application of the NAH for the engineering problems, the project is proposed to develop a mapping relationship based NAH which can explicitly determine the necessarily required measurement to obtain a well reconstructed result of the acoustic sources and supply a way to evaluate the reconstructed results based on the given noise level. The project is going to explore the construction theory of the patterns for the NAH-suitable decomposition of the vibration source as well as the radiated filed, form a stable mapping relationship of the patterns between the source and the radiated sound field, investigate the influence of the position of the hologram and the number of the participant patterns on the characteristics of the reconstruction operators, build a determination theory of the hologram position and number of participant patterns in constructing an optimized reconstruction operator. Consequently, a recognition algorithm for the radiated patterns on the hologram is going to be developed. To build a robust and accurate pattern recognition algorithm as well as experimental theory, influences of the measurement, inevitably included errors as well as exterior noise source on the reconstruction of the coefficients of participant patterns. At last, a mapping relationship based NAH whose theory of reconstruction algorithm and experiment implementation are well-matched is developed. It is expected to deliver a powerful sound source identification tool with advantages of efficient, precise and easy implementation features.

近场声全息方法是分析振源及控制噪声的有力工具。为了便于指导近场声全息的工程应用，保证采取必要的测试代价以获得良好的声源重建结果,本项目拟建立基于模式映射关系的近场声全息理论:通过探索适用于结构振速分解及声场重构的分布模式构建理论，建立结构表面振源模式和辐射声场分布模式的映射关系，研究全息面测试位置及分布模式参与项数对重构算子特性的影响，建立最优化重构算子特性的全息面测试配置及模式参与项数的判定理论和方法；进而研究声场分布模式识别算法，明确传声器配置方式、测试误差和外部噪声源对于模式参与系数识别的影响，发展声场分布模式的稳健识别理论及其阵列测试方法。最终，建立一套基于模式映射关系的近场声全息理论和测试方法，为噪声源特性分析提供一种重建算法、全息面测试位置和及其空间采样配置三者均衡匹配的成像方法，满足声敏感设备降噪设计对高精度、稳健噪声源分析工具的需求。

近场声全息方法在噪声源可视化与噪声控制方面具有广泛的应用。在传统的近场声全息方法中，声源重建的精度受到了声场传播机理、全息测试面与环境噪声等限制。本项目针对近场声全息方法声源重建的精度限制这个基本问题，研究基于模式映射关系的近场声全息理论和测试方法。通过研究基于边界元法的近场全息理论，构建立了从复杂结构到任意测量面的数值映射模型。围绕封闭舱室混响中环境噪声源重建方法、基于机器视觉的实时声源重建方法和逆块传递函数法的实时声源重建方法，解决了混响的存在对重建声源精度发生畸变的问题，实现了不规则形状声源表面的全局三维精准重构。最后，利用信号同步的声源重建方法，有效地解决了测量中传声器数量的限制，并开发了一种自动化三维全息阵列系统，实现待测物体全息面上声压地高效采集，开展了近场声全息的工业应用实验研究。