数据驱动的脉冲风洞气动热试验数据分析方法探索  

Investigation of a Data-Driven Method for Analyzing Impulse Tunnel Aerothermal Test Data

作　　者：易翔宇 鄢明 陈星[1] 林键[1] 姚大鹏[1] 纪锋[1] 应明 张占一 YI Xiangyu;YAN Ming;CHEN Xing;LIN Jian;YAO Dapeng;JI Feng;YING Ming;ZHANG Zhanyi(China Academy of Aerospace Aerodynamics,Beijing 100074,China;China Orient Institute of Noise and Vibration,Beijing 100085,China)

机构地区：[1]中国航天空气动力技术研究院,北京100074 [2]北京东方振动与噪声技术研究所,北京100085

出　　处：《气体物理》2024年第4期65-82,共18页Physics of Gases

摘　　要：脉冲风洞的有效运行时间短,气动热测量信号易受干扰,通常采用人工判读的方式在大量的脉冲风洞测热试验数据中遴选有效数据,数据分析过程效率低、标准不统一。研究基于中国航天空气动力技术研究院1 m量级炮/激波双模式风洞气动热试验数据开展信号特征分析,确定了Fourier变换后的电压信号、总压-热流相关系数曲线、热流时间导数等参数对流动起止时间、运行模式分界时间、信号质量等关键信息的表征作用。在此基础上,基于卷积神经网络和密度基聚类算法,发展了传感器信号有效性判断模型、风洞有效运行时间段选取算法和传感器信号取值区间选取算法,实现了脉冲风洞气动热试验数据的智能处理。结果表明,智能算法对传感器信号有效性的判断准确度达到98%;针对3568条原始试验曲线进行处理,智能算法与人工方式得到的热流均值差异小于10%的数据约占91%,具有应用于风洞试验的价值。The test time of impulse tunnels is short,and the heat flux transducers can be easily interfered in the test time.Thus the aerothermal test data analysis of impulse tunnels is complex and usually carried out manually,which leads to low efficiency and nonuniform standards.In this paper,the signals of heat transducers used in the 1 m gun/shock dual-mode impulse tunnel at the China Academy of Aerospace Aerodynamics were investigated.A series of key parameters,e.g.,the FFT voltage profile,the correlation profile of nozzle total pressure and transducer heat flux,as well as the time derivative of the heat flux profile,were determined.The characterization capabilities of these parameters with respect to significant features,e.g.,the start/end time of the nozzle flow,the time boundary of different operation modes and the effectiveness of the transducer profiles,were certified.Moreover,a profile effectiveness estimation model,a test time interval selection algorithm and a transducer valid time selection algorithm were developed based on convolutional neural networks and density-based clustering algorithms.Thereby the aerothermal test data of impulse tunnels can be analyzed automatically with these data-driven methods.The results show that the accuracy of the profile effectiveness estimation model is as high as 98%.According to the comparison between algorithm and manual results in dealing with 3568 profiles,about 91%of the results show a difference of less than 10%,which implies that the data-driven method has the value of application to wind tunnel tests.

关 键 词：脉冲风洞 气动热试验 热流 卷积神经网络 密度基聚类 

分 类 号：O354.4[理学—流体力学]