遥测速变信号敏感特征的多尺度非线性挖掘与分析技术研究

Telemetry vibration signal contains a great deal of information that reflects the status of the aircraft during the test, and is an important basis for condition monitoring, structural design improvement and fault detection. Therefore, it is of great theoretical and practical value to study advanced signal processing methods to accurately extract signal features that reflect the operating status of the system. For the signal has characteristics of strong noise,non-linear, non-stationary, transient impact and so on，a multi-scale nonlinear signal analysis technology is researched. Firstly, a multi-scale nonlinear signal analysis technology is researched, and a noise reduction method based on phase space multi-scale mainstream shape recognition is proposed, which can effectively reduce noise and maximize ensure the integrity of the impact component of the original signal without changing dynamics of the system; Then the mechanism of multi-scale band-limited noise on stochastic resonance is studied, which can achieve different scales of noise energy transfer from the stochastic resonance system to the drive frequency through multi-scale noise adjustment and optimization of nonlinear stochastic resonance. Therefore, the feature information is enhanced; Finally, a multi-scale manifold learning and analysis method is proposed, which can deeply digging the intrinsic relationship between different scales of the signal. Through the analysis of sensitive features, the multi-scale non-linear sensitive features that can reflect the nature of the operating status state are extracted, and the accuracy of condition monitoring and fault diagnosis are improved. Through the research of the project can not only provide evidence for the analysis of aircraft state, structural design improvement, fault detection, But also can be extended to other areas of fault detection.

遥测速变信号反映了飞行器在试验过程中的状态信息，是状态监测、结构设计改进和故障检测的重要依据，因此研究先进的信号处理方法，准确提取反映系统运行状态的信号特征具有重要的理论和应用价值。针对信号强噪声、非线性、非平稳性、瞬态冲击性等特点，拟对多尺度非线性信号分析方法进行深入研究，提出基于相空间多尺度主流形识别降噪方法，在不改变系统动力学特征和保证信号中冲击成份完整性的前提下，实现有效降噪；探索多尺度带限噪声对随机共振的影响机理，通过多尺度噪声调节和优化非线性随机共振，实现不同尺度的噪声能量通过随机共振向驱动频率转移，对特征信息进行增强；研究多尺度流形学习分析方法，深入挖掘信号不同尺度信息之间的内在联系，通过敏感特征分析提取最能反映系统运行状态本质的多尺度非线性敏感特征，提高故障诊断的准确性。本课题的研究可为飞行器状态分析、结构设计改进、故障检测提供理论和技术支撑,也可推广到其它故障检测领域。

遥测速变信号反映了飞行器在试验过程中的状态信息，是状态监测、结构设计改进和故障检测的重要依据，因此研究先进的信号处理方法，准确提取反映系统运行状态的信号特征具有重要的理论和应用价值。课题以飞行器系统状态监测与精确故障诊断为目标，针对速变信号强噪声、非线性、非平稳性、瞬态冲击性等特点，对多尺度非线性信号分析方法进行了深入研究，首先，提出了相空间多尺度主流形识别降噪方法，将相空间信号分解到不同的尺度空间，根据尺度空间的信噪分布特点，实现有针对性的降噪，保证信号冲击成分的完整性不受影响、波形不失真以及不改变其反应的系统动力学特征；然后对多尺度优化非线性信号特征增强方法进行了深入研究，提出两种解决方案：一种是基于自适应可调品质因子小波稀疏优化的冲击特征增强方法，以信号的时频分布为基础，设计了小波关键参数的自适应选取策略，将稀疏优化思想引入到小波分解过程中，进一步凸显信号中稀疏的冲击成分；另一种是基于双树复小波多尺度噪声调节随机共振方法，在理论上探索多尺度带限噪声对非线性随机共振的影响机理，利用白噪声和带限噪声通过非线性随机共振系统后呈现洛伦兹分布这一特点，设计了多尺度噪声类型和水平的优化调节方法，使得噪声能量通过随机共振向驱动频率转移，有效地对信号的微弱特征信息进行了增强；接下来提出了基于时频流形自适应稀疏重构的信号分析方法，深入挖掘信号不同尺度信息之间的内在联系，有效提高了信号瞬态特征提取的准确性；最后是多尺度特征选择方法，提出了参照化流形空间融合学习的敏感特征提取与异常检测方法，通过建立专属参照化模型单元，采用级联流形学习的模式实现对信号多尺度特征信息的多层挖掘再学习，有效提高了特征信息的敏感性和故障识别的精度。本课题的研究可为飞行器系统状态分析、结构设计改进、故障检测提供理论和技术支撑,也可推广到其它故障检测领域。