基于BP神经网络的土壤氮素运移模型

Model of Soil Nitrogen Transport Based on BP Neural Network

随着淡水资源的日益紧缺,再生水灌溉已成为人们日益瞩目的研究方向,而再生水灌溉条件下土壤氮素运移规律与模拟成为这个研究的关键环节之一。以往对土壤氮素运移的模拟主要聚焦在数值模拟上,鉴于数值模拟在应用上的复杂性,为了寻找一种简便实用的模拟方法,尝试引入人工神经网络技术对土壤氮素运移进行模拟,经模拟计算得出,拓扑结构为10∶12∶7的BP网络模型可以较精确地模拟再生水灌溉条件下的土壤氮素运移,此研究为土壤氮素运移的研究开辟了新方向。

With the increasing scarcity of fresh water resources, reclaimed water irrigation has become an increasingly attention-getting research area, in which the simulation of soil nitrogen transport was one key problem. Formerly, the simulation of soil nitrogen transport was mainly focused on numerical model, which was very complicated. In order to seek a simple and applied simulation method, BP neural network was introduced to simulate soil nitrogen transport in this paper. The simulation result indicated that BP network model, in which the topology structure was 10 ： 12 ： 7, could simulate soil nitrogen transport accurately. So this study would supply a new approach for the soil nitrogen transport research.