BP神经网络水华预测模型的敏感性分析  

Sensitivity Analysis of BP Neural Network for Algal Bloom Prediction Model

作　　者：殷高方[1] 张玉钧[1] 胡丽[2] 于绍惠[1] 肖雪[1] 王欢博[1] 石朝毅[1] 段静波[1] 刘文清[1] 

机构地区：[1]中国科学院环境光学与技术重点实验室中国科学院安徽光学精密机械研究所,安徽合肥230031 [2]合肥学院建筑工程系,安徽合肥230022

出　　处：《北京理工大学学报》2012年第12期1288-1293,共6页Transactions of Beijing Institute of Technology

基　　金：国家"八六三"计划资助项目(2009AA063005);合肥学院科研发展基金资助项目(12KY05ZR);安徽光学精密机械研究所所长基金资助项目(Y03AG31144)

摘　　要：敏感性分析能够定量地评价模型输入变量的变化对输出结果产生的影响,是揭示模型蕴含规律的有效途径.本文将敏感分析方法应用于BP神经网络巢湖水华预测模型中,分析结果表明巢湖水华形成受诸多环境因子共同影响,水温、溶解氧、浊度、气温、光照强度等环境因子变化与藻类质量浓度变化相关,其中气温是最大影响因素,相对贡献率达到17.01%;气压的上升则不利用于藻类质量浓度的增加;pH值的上升对藻类质量浓度的影响有正有负.Sensitivity analysis, which can quantitatively estimate the contribution of input variable to the output, is an effective way to reveal the inherent laws of the model. In this paper, sensitivity analysis was applied to the algal blooms forecast model based on BP neural network in Chaohu Lake. The result of the analysis indicates that algal blooms in Chaohu are affected by many factors. There was a positive correlation among the change of water temperature, dissolved oxygen, turbidity, atmospheric temperature, illumination and the change of mass concentration of algal. Among these factors, atmospheric temperature is the most important, with a relative contribution up to 17.01%; on the contrary, the rise of atmospheric pressure does harm to the algal; and the influence of high pH on the algal concentration is uncertain.

关 键 词：水华 预测模型 BP神经网络 敏感性分析 

分 类 号：X824[环境科学与工程—环境工程]