摘要
针对传统的统计学方法难以精确刻画岩溶地下河日流量变化的非线性动态特性,引入有源自回归神经网络（NARX）技术,建立了基于NARX模型的岩溶地下河日流量预测模型,基于寨底地下河2013年1月15日~2014年6月30日的降雨量和流量数据,利用该模型对寨底地下河日流量进行了短期预测。结果表明,该模型预测效果较好,能够很好地预测岩溶地下河流量的变化趋势和极值等动态特性,另外该模型神经元个数越多,延迟阶数越大,神经网络对数据的学习能力和灵活性越强,但该模型不宜进行归一化处理。
The nonlinear dynamic characteristics of daily discharge for karst underground river is hardly depicted by using traditional statistical methods. In this paper, an artificial neural network of nonlinear autoregressive with exogenous inputs （NARX） has been developed to predict daily discharge of karst underground river. Based on the rainfall and dis- charge data from 15 January 2013 to 30 June 2014, the short-term daily discharge of Zhaidi underground river was fore- casted by using NARX model. The results show that the NARX model has good prediction effect and it can correctly sim- ulate discharge trend and extreme value; the number of neurons and time delay are in proportion to the network flexibility and learning ability, but the data normalization of the prediction model was inadvisable.
出处
《水电能源科学》
北大核心
2015年第5期19-21,25,共4页
Water Resources and Power
基金
中国地质大调查项目资助(1212011220959)
中国地质科学院岩溶地质研究所基本科研业务费项目(121237128100249)
关键词
NARX模型
岩溶地下河
日流量
预测
NARX model
karst underground river
daily discharge
prediction
