基于人工神经网络算法的某机场雾的模式识别预报  

Model Recognition and Forecast of Fog over the Airport Based on ANN

作　　者：袁志康[1] 张韧[1] 周树道[1] 李宁 赵奇峰 

机构地区：[1]解放军理工大学气象学院 [2]南京73602部队气象台

出　　处：《解放军理工大学学报（自然科学版）》2002年第4期82-86,共5页Journal of PLA University of Science and Technology(Natural Science Edition)

基　　金：高等学校骨干教师资助计划项目 ( 314 9);暨总装备部"低云大雾监测预报"课题资助

摘　　要：基于人工神经网络 (ANN)误差后传算法 (BP模型 ) ,通过对 1 0 0个有雾天气个例和 5 0个无雾天气个例进行模式识别训练 ,用前期 (前日 0 8、1 4、2 0时实测资料 )的温、压、湿、风等要素指标建立起了某机场雾的 ANN识别预报模型。该模型具有很好的拟合效果 ,独立样本的试验和检测结果均达到较为精确的预报效果 。Based on the BP model of ANN, a recognition model is built for fog forecasting over a the airpart. Through model recognition training and with one hundred fog cases and fifty non-fog cases and based on six prereguisite weather element indicators, such as temperature, pressure, humidity and wind etc, a ANN fog forecast model of a the airport is established. The results show that this ANN model has a good fitting effect. Both for prediction of independent cases and experiment in model tests show considerably accurate forecast skill. Therefore, this ANN model can be applied to the operational fog forecasting.

关 键 词：人工神经网络算法 机场 雾天气 BP模型 模式识别 天气预报 形成机理 

分 类 号：P457.7[天文地球—大气科学及气象学] TP183[自动化与计算机技术—控制理论与控制工程]