对非均匀温度场内定位方法进行研究，以解决炉膛热态情况下压力管道泄漏定位问题。使用广义互相关方法进行远距离时延值测量实验并进行低信噪比仿真；使用球型插值法进行三维空间定位仿真；在CFD数据库的基础上，假设声波直线传播，使用初值迭代寻优模型对泄漏定位。实验中，传感器相距10 m，仿真环境为-6 db，三维空间深9 m、宽9 m、高5 m。单峰温度场下，将温度场平均与假设声波直线传播的时延值误差对比。结果表明，广义互相关测量时延值在大空间、低信噪比环境下可行；单峰温度场情况下，假设声波直线传播、温度场平均的情况下所得时延值误差分别在0.2和2.2 ms范围内，由此导致的定位误差分别为0.4和8 m范围内，初值迭代寻优模型能够有效解决非均匀温度场中声速不固定而导致被动定位算法时延值无法反推距离差的问题；假设声波直线传播的时延值误差及由此导致的定位误差满足工程精度要求。

To solve the problem of pressure pipe leakage in furnace,the locating method was studied in non-uniform temperature field GCC experiment of long distance was carried out,and -6 db SNR simulation was performed.Spherical interpolation method was used to determine the location in three-dimensional space.Based on CFD database,an optimization model based on initial value was proposed to locate the leakage position with straight line hypothesis.In the experiment,the distance of two sensor was 10 m.The SNR was6db.The scale of location model was 9 m×9 m×5 m.On the basis of single temperature field,TDOA values in two conditions were compared,which were the value of straight line hypothesis and the value of average temperature field.The result showed that the GCC method was feasible in large scale and low SNR condition.TDOA errors were in the range of 0.2 ms and 2.2 ms in straight line hypothesis and average temperature field,and the corresponding location errors were in the range of 0.4 m and 8m.The TDOA and location errors in straight line hypothesis meet the accuracy requirement .

TK223.1

国家自然科学基金(11674093)