基于模拟退火算法的航向实测磁场数据建模分析

Modeling and Analysis of Directional Field Data Based on Simulated Annealing Algorithm

为了提高考虑测量数据的真实性时测量的精度以及稳定性,采用模拟退火算法对磁偶极子的位置进行了优化处理,再通过逐步回归方法来进一步优化磁矩,并去除测量点中存在粗大误差的数据。分别开展实验室数据和海况实测数据的误差分析:针对实验室数据,该模型可以得到很高的精度,得到拟合误差的最小值约为0.01。增加干扰信号之后,采用当前方法计算得到的拟合误差约为0.02,表明该模型具备较高的精度,得到更稳定的模型,达到更优的鲁棒性。针对海况实测数据,先对10.5m深度处得到的数据求解得到所需的模型,同时换算20.5m处的磁场,数据误差发现,采用该方法可以得到95%的较高建模精度,完全满足实际应用的精度条件。

In order to improve the accuracy and stability of measurement when considering the authenticity of measurement data,simulated annealing algorithm is used to optimize the position of magnetic dipole,and then stepwise regression method is used to further optimize the magnetic moment and remove the data with gross errors in the measurement points.Error analysis of laboratory data and sea condition measured data are carried out.For laboratory data,this model can obtain high accuracy,and the minimum fitting error is about 0.01.After increasing the interference signal,the fitting error calculated by the current method is about 0.02,indicating that the model has higher accuracy,more stable model and better robustness.Aiming at the measured data of sea conditions,the required model is obtained by solving the data obtained at a depth of 10.5m,and the magnetic field at a depth of 20.5m is converted at the same time.The data error is found that 95%of the higher modeling accuracy could be obtained by using this method,fully satisfying the accuracy condition of practical application.