|
This paper according to the characteristics of the groundwater level dynamic data, using a variety of mathematical methods, the underground water level is decomposed into trend item functions, periodic item and random item three parts, and establish the mathematical model, the function application time regression method, the trend is expressed as a linear combination of the elementary function, application of periodic item Fisher criterion, isolate the main cycle for random items, use method of system modeling, establish regression... Moving average model (ARMA) model, through the nonlinear least squares model parameters can be estimated, using F criterion to judge the correct model order. When forecasting, trend and cycle time extension method is adopted, and the random items, by solving the green's function, using the minimum variance method to carry on the forecast. Test through an example, the modeling and forecasting theory has achieved good results. In addition, methods for the combination model to design a computer program, that this method has great popularization value, made contributions in the informatization of geological environmental monitoring work.
|