Study on key technologies of construction of marine geomagnetic filed model based on polynomial method

摘要

There is an important relationship between the research of geomagnetic field model and the digital earth. The earth information includes that of geomagnetic field. Geomagnetic field model is to describe the geomagnetic field in mathematic way. It can provide the precise geomagnetic field information to help the development of the digital earth. At the same time, the geomagnetic field model is also used in the land and underwater geomagnetic navigation, and acts as an important role in finding underwater targets. Thus the paper will focus the research how to build a local marine geomagnetic field model. Firstly, we use the polynomial model, which includes Taylor polynomial model and Legendre polynomial model, to build the local marine geomagnetic model. Because geomagnetic field model expresses the steady geomagnetic change of the earth core. Thus we study Robust Kalman filter to eliminate the influences of abnormal geomagnetic data secondly. Besides, we also study the determination of model truncation order in using polynomial modeling, Because model truncation determines model accuracy and the efficiency in building local marine geomagnetic model. Model accuracy can be estimated by the standard deviation in interior test and exterior test. In the research, we find that model accuracy changes with model order, when the accuracy become stability with model order, we regard the order as model truncation order. Using above conclusions, we implement a experiment in the situations of geomagnetism mild change and geomagnetism complex change. We find that both Taylor polynomial model and Legendre polynomial model have smaller model truncation order in the situations of geomagnetism mild change than that in the situations of geomagnetism complex change. We also compare and analyze Taylor polynomial modeling and Legendre polynomial modeling in a large water area. By comparing and analysis, we find that Taylor polynomial modeling has higher accuracy and faster convergence spe-ed than Legendre polynomial molding. Thus Taylor polynomial is recommended as a optimum polynomial method in building local marine geomagnetic model. Finally we use both Taylor polynomial method and Legendre polynomial method to construct the local geomagnetic field model in the experimental water area of Bohai Bay in China. Taylor polynomial model achieves lower model truncation order and satisfying accuracy relative to Legendre polynomial model, which further prove above conclusions.