| Groundwater is not only the valuable natural resources, but also the basic element of environment. Reasonable model applied in groundwater dynamic trend prediction is important both for properly exploiting of production department and properly decision making of management department. Defect against the indiscernible hydrologic geology information and aquifer system of old water resource area, research of groundwater random dynamic prediction model has been done. Based on the instance in actual project, models for groundwater dynamic prediction are built, one is the random coupling model of nonstationary time series, and the other is optimum neural network model of multiple forecast factors.According to the J2EE criterions, base on the idea of MVC design pattern and module developing, the Struts framework has been used to establish a groundwater dynamic prediction system in B/S Web circumstance for improving the models' practical application and real time application. In the meanwhile, two prediction algorithm models are integrated into the Web system with the technology of Matlab network function, which gives a new conception to build a general flatform for groundwater scientific management and on-line decision, and makes feasible to realize real-time data management and model prediction. Through the existed instances, it is show practicability and rationality in technique implementation. This system has advantages of better augmentability and portability, on which basis more practical business can be refined and extended in order to establish a compositive application flatform of groundwater dynamic prediction.The main research content and results are as follow:(1) Depart from the technique research, general researching framework of this paper is proposed, which threads three parts as theory research, practical model building and algorithm Web integration of groundwater random dynamic prediction;(2)On the basis of groundwater system generalized, this paper analyses the...
|
PDF Full Text Request