基于最优化插值算法的室内温度场模型

Indoor Temperature Field Model Based on Optimized Interpolation Algorithm

本文提出了一种改进的空间插值方法,将插值公式中的一些原本是经验值的常数值参数化,然后将部署在空间的传感器网络划分为建模网络和校验网络,并结合最优化的方法求解使校验网络误差最小的参数,从而可以校验得到的温度场精度。本文在一个温差较大的小机房实验了该方法,证明该方法在该环境中可以取得最大误差仅为0.27度的良好精度。

In this paper,an improved spatial interpolation method is proposed,which parameterizes some constant values which are originally empirical values in the interpolation formula.Then,the sensor networks deployed in space are divided into modeling network and calibration network,and the parameters with minimum error of calibration network are solved with optimization method,so as to verify the accuracy of temperature field.In this paper,the method is tested in a small computer room with large temperature difference.It is proved that the method can achieve a good accuracy of only 0.27 degree in this environment.