摘要
以黄河龙门—潼关河段作为研究区域,在分析区间来沙来水对含沙量影响的基础上,建立了基于BP神经网络的潼关站含沙量过程预报模型。同时,为提高预报模型的预报精度,利用误差序列建立了相应的误差自回归模型对预报结果进行校正。校正前后泥沙过程的对比分析表明,校正后的泥沙过程预报精度有较显著的提高,5场验证泥沙场次的平均确定性系数由校正前的0.35提高到校正后的0.76。
Taking the reach from Longmen to Tongguan in Yellow River as study area, a BP neural network model is built to forecast the duration of sediment concentration in Tongguan Hydrological Station after analyzing the impact of sand and water runoff on sediment concentration. At the same time, an error self-regression model is also built based on error sequence to calibrate forecasting results. The durations of sediment concentration before and after calibration are compared, and the results show that the forecasting precision of calibrated duration of sediment concentration is significantly improved and the average uncertainty coefficient of five sediment delivery processes is increase to 0.76 from 0.35.
出处
《水力发电》
北大核心
2013年第1期23-26,66,共5页
Water Power
基金
水利部公益性行业科研专项经费基金资助项目(2009101016)
江苏省高校优势学科建设工程基金资助项目
关键词
含沙量
BP神经网络
误差自回归
水文预报
黄河中游
esediment concentration
BP neural network
error self-regression
hydrological forecasting
middle reaches of Yellow River
