基于非定常流场脉动特征的数据融合方法及其在压气机角区失稳机理研究中的应用

对压气机角区复杂流动的深刻认识与有效控制是航空压气机气动性能提升的关键，然而时空多尺度、高度非线性耦合的流动特征和多源、海量的流场数据给角区流动失稳机理的认识带来巨大挑战。项目基于多尺度动力学模态分解方法、相关性分析理论和逻辑回归模型，突破不同时空特征的多源数据难以融合的问题，建立基于脉动特征的高分辨率多源（实验源和数值源）数据关联机制，发展面向脉动模态稳定性分析的数据融合方法；在此基础上，以高负荷压气机角区高精度的数值流场和非定常实验结果为基础，突破单一数据源难以在时空层面全面、准确地分析复杂流场的难题，将数值源数据时空分辨率高、解析远壁面流场精细的优势和实验源数据采样时间长、捕捉近壁流动准确的优势相融合，建立空间拟序结构与流动失稳的关联机制，从多源综合信息中阐释角区流动失稳诱因。为高负荷航空压气机研制中面临基础科学问题提供理论支撑，为高分辨率数值流场和实验结果的充分挖掘提供新思路。

A deep understanding and effective control of the complex flow of compressor corner separation is the key to improving the aerodynamic performance of air compressors. However, space-time multiscale, highly nonlinear coupled flow characteristics, as well as multi-source and mass flow field data bring great challenges to the understanding of flow instability (flow field deterioration causes a drop of pressure ratio) mechanism in the corner region. Based on the multi-resolution dynamic modal decomposition method, correlation analysis method, and logistic regression model, the project intends to build a multi-source data (experimental and numerical sources) relation association method based on fluctuation characteristics and develops a data fusion method for fluctuation modal stability prediction, which breaks through the problem that it is difficult to integrate multi-source data with different temporal and spatial characteristics. Based on this, taking the high-precision numerical flow field of the corner area obtained by the detached eddy simulation and the unsteady experimental results of the corner area obtained by the sensor measuring as the multiple source data, and by integrating the advantages of numerical and experimental data (numerical simulation: high space-time resolution and predicting flow field far away from the walls ; experimental results: long sampling time and capturing the flow field close to the wall), which breaks through the difficulty of a single data source to comprehensively and accurately understand the complex flow field in the corner area. Then the mechanism which induces the flow instability of the corner area is analyzed from multi-source comprehensive information. This project not only provides theoretical support for the basic scientific problems faced by the development of high load aviation compressor, but also provides a new idea for the full mining and utilization of numerical simulation and experimental data with high spatial and temporal resolution.