从载人飞行以来，在飞机、导弹、宇宙飞船和航天飞机的发展中，实验空气动力学始终起着关键性作用。高超声速气动力试验是风洞实验中最基本的实验项目，也是实验空气动力学的重要组成部分，能够为高超声速飞行器研制提供可靠的气动力试验数据。风洞天平是气动力试验中最重要的测量装置，准确获取并处理好天平信号是确保试验数据精准度和可靠性的最基本保证。高超声速气动力试验技术的发展与应用有利于深入了解高超声速飞行器的气动力特性，对推动我国高超声速领域研究与发展至关重要，并为相关领域的科学研究和工程实践提供重要参考。

随着我国高超声速科技的发展，除了对高超声速气动力试验指标提出了更高要求，还需实现高温、高压环境下的高性能测力试验。激波风洞以其可模拟高焓环境流场和高超声速飞行器极端苛刻飞行条件，成为可获得接近真实飞行环境气动力结果的重要地面模拟设备。然而，在激波风洞中进行气动力测量时，面临着时间短、冲击强、干扰大等挑战，传统的常规风洞测力方法难以满足精准度要求，不适用于激波风洞测力。因此，发展适用于短试验时间激波风洞的高精准度气动力试验技术以及数据处理方法是国家航空航天领域面临的迫切需求之一。

本文面向国家空天科技发展的重大需求，探索力学与其他学科领域的交叉融合及应用，以脉冲型一体化测力系统（integrated Force Measurement System, iFMS）为研究对象，围绕“激波风洞天平各维分量间的动态非线性耦合特性”、“冲击载荷激起结构的振动形式和激振的机制问题”和“复杂结构受迫振动情况下对气动力干扰规律影响”等基础性科学问题，以风洞天平静态校准、信号处理和测力系统动态校准为技术路线，结合时频变换和深度学习等方法，提出适用于短试验时间激波风洞气动力测量技术与智能化天平数据处理方法，可获取高精度的高超声速飞行器气动力数据。本文主要研究内容和创新成果如下：

1）提出了基于卷积神经网络（Convolutional Neural Network, CNN）的风洞天平静态校准数据处理方法。基于深度学习对非线性问题的处理优势，建立CNN风洞天平静态校准模型，实现了天平电压输出值与加载载荷之间的高精准度智能映射。对现有脉冲型天平开展校准获得的数据显示，CNN静态校准模型相较于传统方法，其综合加载误差和重复性等指标达到了国军标的先进指标要求，个别指标性能提升近1个量级。

2）发展了用于激波风洞瞬态气动力测量的小波变换（Wavelet Transform, WT）和希尔伯特-黄变换（Hilbert-Huang Transform, HHT）天平信号处理方法。发展了适用于激波风洞测力信号处理的小波阈值降噪方法，将信噪比（Signal-Noise Ratio, SNR）提升20%，获得了更高质量的天平阶跃信号，支撑后续高精准度天平信号智能处理。基于小波变换和希尔伯特-黄变换，成功重构出天平信号中惯性特征和其他干扰成分，与理想阶跃信号的相对误差优于1%，为工程测力应用中激波风洞天平惯性干扰信号的有效处理提供了新思路。

3）提出了基于人工智能（Artificial Intelligence, AI）技术的气动力测量系统单矢量动态自校准（Single-Vector Dynamic self-Calibration, SVDC）方法。提出一种新的基于深度学习技术的单矢量动态自校准方法。该方法将人工智能深度学习技术引入脉冲型风洞测力系统的动态校准，对非线性项的智能处理将切实提高测力系统性能。新动态校准方法力求突破传统高精度测力系统因脉冲风洞极短试验时间而限制其航空航天领域应用的技术瓶颈，可使脉冲高焓风洞测力系统性能和测力试验技术指标在现有基础上大幅度提升。

4）提出了一种标准流程化的智能测力方法并成功应用于激波风洞气动力试验。单矢量动态自校准方法的应用打破了传统典型气动力试验流程，在试验前对安装于风洞内的气动力测量系统（模型-天平-支撑杆-弯刀机构及底座）开展一体化动态特性智能校测，探索了智能测力方法在复现高超声速飞行条件激波风洞（JF-12复现风洞）测力试验的应用。分别通过尖锥标准模型和两级入轨（Two Stage To Orbit, TSTO）复杂构型飞行器测力试验验证了方法的可靠性和精准度，未来可进一步开展工程应用。

Since the beginning of manned flight, experimental aerodynamics has been playing a key role in the development of aircraft, missiles, spacecraft and space shuttles. Hypersonic aerodynamic force test is the most basic experimental item in wind tunnel experiments and an important part of experimental aerodynamics. It can provide reliable aerodynamic force data for the development of hypersonic aircraft. The wind tunnel balance is the most important measuring device in aerodynamic force test. Accurate acquisition and processing of balance signals is the most basic guarantee to ensure the accuracy and reliability of test force data. The development and application of hypersonic aerodynamic force test technology is conducive to an in-depth understanding of the aerodynamic characteristics of hypersonic aircraft, is crucial to promoting research and development in the field of hypersonics in our country, and provides important reference to scientific research and engineering practice in related fields.

With the development of our country's hypersonic technology, in addition to higher requirements for hypersonic aerodynamic force test indicators, it is also necessary to achieve high-performance force measurement tests in high-temperature and high-pressure environments. The shock tunnel has become an important ground simulation equipment that can obtain aerodynamic results close to the real flight environment because it can simulate high-enthalpy environmental flow fields and extremely harsh flight conditions of hypersonic aircraft. However, when conducting aerodynamic force measurement test in a shock tunnel, it faces challenges such as short time, strong impact, and large interference. The traditional conventional wind tunnel force measurement method cannot meet the accuracy requirements and is not suitable for shock tunnel force measurement. Therefore, the development of high-precision aerodynamic force test technology and data processing methods suitable for short test-duration shock tunnels is one of the urgent needs faced by our national aerospace field.

This article faces the major needs of the development of national aerospace science and technology, explores the cross-integration and application of mechanics and other disciplines, takes the pulse-type integrated Force Measurement System (iFMS) as the research object, and focuses on fundamental scientific issues, such as "dynamic nonlinearity coupling characteristics between the various dimensional components of the shock tunnel balance", "the form of vibration and excitation mechanism of structures excited by impact load" and "the impact of aerodynamic force interference on complex structures under forced vibration", and takes the static calibration and signal processing of wind tunnel balances, dynamic calibration of the force measurement system as a technical route, combines time-frequency transformation and deep learning methods, proposed a shock tunnel aerodynamic measurement technology and an intelligent balance data processing method suitable for short test duration, which can obtain high-precision aerodynamic data of hypersonic aircraft. The main research contents and innovative results of this article are as follows:

1) A static calibration data processing method for wind tunnel balances based on Convolutional Neural Network (CNN) is proposed. Based on the advantages of deep learning in processing nonlinear problems, a CNN wind tunnel balance static calibration model was established to achieve high-precision intelligent mapping between the balance voltage output value and the loading load. Data obtained from the calibration of existing pulse-type balances show that compared with traditional methods, the comprehensive loading error and repeatability of the CNN static calibration model have reached the advanced index requirements of the national military standard, and the performance of individual indicators has improved by nearly 1 order of magnitude.

2) Developed Wavelet Transform (WT) and Hilbert-Huang Transform (HHT) balance signal processing methods for shock tunnel transient aerodynamic force measurement. A wavelet threshold noise reduction method suitable for shock tunnel force measurement signal processing was developed, which increased the signal-to-noise ratio (SNR) by 20%, obtained higher-quality balance step signal, and supported subsequent intelligent processing of high-precision balance signals. Based on WT and HHT, the inertial characteristics and other interference components in the balance signal are successfully reconstructed. The relative error with the ideal step signal is better than 1%, which provides new ideas for effective processing of the inertia interference signals of shock tunnel balance in engineering force measurement applications.

3) A single-vector dynamic self-calibration (SVDC) method for aerodynamic measurement systems based on Artificial Intelligence (AI) technology is proposed. A new SVDC method based on deep learning technology is proposed. This method introduces artificial intelligence deep learning technology into the dynamic calibration of the pulse-type wind tunnel force measurement system, and the intelligent processing of nonlinear terms will effectively improve the performance of the force measurement system. The new dynamic calibration method strives to break through the technical bottleneck of the traditional high-precision force measurement system that limits its application in the aerospace field due to the extremely short duration of the pulse-type wind tunnel testing. It can significantly improve the performance of the pulse high-enthalpy wind tunnel force measurement system and the technical indicators of the force measurement test on the existing basis.

4) A standard process-based intelligent force measurement method is proposed and successfully applied to shock tunnel aerodynamic tests. The application of the SVDC method breaks the traditional typical aerodynamic test process. Before the test, the aerodynamic force measurement system (model/balance/support rod/machete mechanism and base) installed in the wind tunnel is carried out with integrated dynamic characteristics intelligent calibration, and the intelligent force measurement method is explored in the the long-test-duration hypervelocity detonation-driven shock tunnel (JF-12 shock tunnel) force measurement test. The reliability and accuracy of the method were verified through the force measurement test of the cone calibration model and the two stage to orbit (TSTO) complex configuration aircraft respectively, and further engineering applications can be carried out in the future.