Ultrasonic measurement is an effective technical method for the bolt stress detection of the precise structural connection. This paper proposed an acoustic time difference algorithm based on adaptive hybrid extended Kalman filter to reduce the calculation error of acoustic time difference in bolt stress ultrasonic measurement. The initial value is firstly selected by particle filter algorithm, and the noise variance is measured by the Sage-Husa estimator. Then, a Kalman nonlinear filter is constructed to adaptively select initial values and observe noise. Thirdly, the filter and cross-correlation algorithm are used to accurately estimate the acoustic time difference. Finally, the signal simulation method and the ultrasonic measurement platform of bolt axial stress are built, and the proposed algorithm and the traditional algorithm are respectively used to calibrate and measure the correspondence between acoustic time difference and axial stress of the M8 coarse tooth half-thread bolt. The bolt tightening experimental and simulation results showed that the processed signal algorithm is verified and can effectively improve the measurement accuracy of bolt axial stress.

超声波测量是精密结构连接螺栓应力检测的有效技术手段。本文提出了一种基于自适应混合扩展卡尔曼滤波器的声时差算法，以减少螺栓应力超声测量中声时差的计算误差。初始值首先由粒子滤波算法选取，噪声方差由Sage-Husa估计器测量。然后，构建卡尔曼非线性滤波器以自适应地选择初始值并观察噪声。第三，使用滤波器和互相关算法来准确估计声学时间差。最后，建立了螺栓轴向应力的信号仿真方法和超声波测量平台，并分别采用所提算法和传统算法对M8粗牙半螺纹螺栓的声时差与轴向应力的对应关系进行标定和测量。螺栓拧紧实验和仿真结果表明，经过处理的信号算法得到验证，能够有效提高螺栓轴向应力的测量精度。