使用时列信号分解的地磁日变基值推算方法

Calculation method of daily geomagnetic variation benchmark based on STL and empirical mode decomposition

针对在地磁场长期变化作用下,地磁日变基值存在日变改正模型误差,导致海洋磁力测量精度难以提高的问题,提出STL+EMD相结合的算法:在提取地磁日变曲线短期变化趋势基础上,进一步分解得到地磁日变的长期变化,并以此建立具有时间维度的地磁日变基值的改正值。此算法优势是填补了对地磁场长期变化影响的客观评判,减小了日变改正中的模型误差。通过与IGRF参考模型的线性拟合方法相比较,STL+EMD算法的精度有显著提升,精度提高了42.27%。将STL+EMD算法应用到2020年8—11月某海域的实测数据表明,与传统平均值算法相比较,实测数据内符合精度提高了49.71%,系统误差降低了60.07%,同时明显改善了测线的交叉点差值。

The geomagnetic day-variation base value is an important basic data for marine magnetic measurement.Under the long-term change of geomagnetic field,the base value of geomagnetic daily variation has the error of daily variation correction model,which makes it difficult to improve the accuracy of marine magnetic measurement.In this paper,we propose a combined STL+EMD algorithm:on the basis of extracting the short-term trend of the geomagnetic diurnal variation curve,we obtain the long-term variation of the geomagnetic diurnal variation through further decomposition,and use this to establish the corrected value of the geomagnetic diurnal variation base value with a time dimension.The advantage of this algorithm is that it fills in the objective judgement of the influence of the long-term change of geomagnetic field and reduces the model error in the correction of the daily variation.By comparing with the linear fitting method of the IGRF reference model,the accuracy of the STL+EMD algorithm is significantly improved,with a 42.27%increase in accuracy.The application of the STL+EMD algorithm to the measured data in a sea area from August to November 2020 shows that,compared with the traditional averaging algorithm,the accuracy of the internal conformity of the measured data has been improved by 49.71 per cent,the systematic error has been reduced by 60.07 per cent,and the intersection point difference of the survey lines has been significantly improved.