摘要
从微观和宏观模型两个方面阐述了给水管网状态模拟技术的最新进展。通过采用参数校正,采样定位,对控制元素的模拟等方法,能够使微观模型达到较高的精度;经验与理论公式相结合的方法可以弥补宏观模型的不足;采用神经网络方法对模型的不稳定情况进行分析,从而得出状态模型的不准确程度;采用模型驱动的神经网络方法建立给水管网宏观模型,建成的模型具有更好的学习能力和容错能力。
Some new developments of the modeling techniques for the municipal water distribution network are expounded with both the microscopic model and the macroscopic model concerned herein. The accuracy of the microscopic model is improved with the methods of the parameter calibration, the data collection and the simulation of the control elements; and the shortages of macroscopic model can be covered with the combination of both the theoretical equation and the empirical one as well. The analysis on the instability of the model concerned is made with the method based on neural network, with which the uncertainty of the status model is known. The macroscopic model for the water distribution network is then established with the method of model driven neural network, which has better learning and fault-tolerant abilities.
出处
《水利水电技术》
CSCD
北大核心
2005年第7期129-131,共3页
Water Resources and Hydropower Engineering
基金
国家自然科学基金资金项目(50278088)。
关键词
给水管网
微观模型
宏观模型
water distribution network
microscopic model
macroscopic model
neural network
