大气可降水量预测模型的自适应Kalmam滤波改进

针对现有可降水量预报模型存在预报精度不高等问题,该文提出采用方差分量估计自适应卡尔曼滤波对可降水量数据进行预处理,用以提高径向基神经网络预测模型的预测精度,从而形成高精度预报模型。通过比较不同基站不同时间的数据,分析使用方法的预报精度。实验结果表明:将预测模型应用于全国7个测站进行实验,预测相对精度的平均值可达95%以上,预报残差在10-5左右,残差值小于0.001的占90%以上。在影响因素方面,使用较短时间作为模型原始数据进行预测会得到较好的预测结果。实验证明本预测方法在预报大气可降水量值方面具有较高的精度。

Improved prediction model of precipitable water vapor using adaptive Kalman filter

In view of lack of high precision precipitation prediction of current model,using the variance component estimation adaptive Kalman filter to preprocess the data was proposed in this paper in order to improve the prediction accuracy of RBF neural network model,so that high accuracy prediction model was formed.By comparing the different time data in different base stations,so as to analyze the prediction accuracy.The experimental results showed that the prediction model was applied to seven stations in the whole country,the average relative accuracy could reach more than 95 ％,the prediction residual is about 10 5,the residual value less than 0.001 accounts for more than 90％.In the aspect of influencing factors,it has better prediction results to use shorter time as the original data to predict and the results showed that the prediction method had high reliability in predicting precipitable water vapor(PWV)value.