指数SVD和Kriging模型的声波传感器三维温度场重建研究

Research on three-dimensional temperature field reconstruction based on exponent-SVD and Kriging

针对以声学CT法重建三维温度场时,由于网格数目有限,边缘温度值无法直接进行重建等问题,提出了一种指数SVD法与Kriging模型相结合的重建方法。采用指数SVD法对典型单峰、双峰、四峰模型温度场进行了重建;引入Kriging模型对重建温度场进行了内插、外推,验证Kriging模型插值精度。结果表明,指数SVD算法和Kriging模型结合的重建温度场重建精度高。将Kriging模型引入到重建温度场的内插、外推中,为解决重建温度场内部温度细化与边缘温度重建,以及网格剖分数目较少的温度场重建问题,提供了一种有效的手段。

Aiming at the problem in three-dimensional temperature field reconstruction by acoustic computer tomography-- limited mesh number leading to edge temperature failure in direct reconstruction, a reconstruction method based on the combination of exponent-SVD and Kriging is proposed in this paper. Typical single-peak, offset single-peak, two-peak and four- peak field models are reconstructed with exponent-SVD at first. And then the detailed description of the whole measured temperature field is acquired by Kriging interpolation and extrapolation and the interpolation accuracy is verified. The results show that the reconstructed temperature field by combination of two algorithms is high in reconstruction accuracy. So an effective approach is provided by introducing Kriging into the interpolation and extrapolation of the temperature field reconstruction for solving the problems in temperature field reconstruction such as inner temperature refining, edge temperature reconstruction, less mesh number, and so on