农村居民用水行为识别方法研究

发布时间：2019-02-25 13:39

【摘要】：随着经济的发展和居民生活水平的提高,人口增长、环境污染、以及城镇化等因素使得水资源供需之间的矛盾愈加突出,用水安全问题日益凸显。研究农村居民用水行为,可以提高当前农村居民的节水意识并改善水资源管理薄弱现状。本文提出的农村居民用水行为识别方法可以准确的识别居民用水行为,改善当前用水基础设施。本文通过分析几种典型的居民用水事件的流量特点。研究了农村居民用水行为的识别方法。具体的工作如下:从训练集合中提取不同用水行为的流量特征,采用由左到右隐马尔可夫模型(Hidden Markov Model, HMM)建立不同类型居民用水行为的识别模型,将测试数据输入训练好的HMM中对居民用水行为进行识别,根据居民此刻用水的流量序列识别居民此时的用水事件。为提高应用HMM的用水行为识别结果的准确度,本文将HMM和时间概率函数结合起来,得出该方法的识别结果。选定人工神经网络(Artificial NeuralNetworks,ANN)算法,设计BP神经网络(Back Propagation, BP)网络结构,确定BP网络训练参数,使用BP神经网络建立居民用水行为的识别模型,最后将测试数据输入训练好的BP神经网络模型中对居民用水行为进行识别,得出识别的结果。研究结果表明:对不同流量模式的用水事件采用HMM和时间概率函数的组合模型能得出更准确的识别结果;相似流量模式的用水事件采用BP神经网络模型识别能得到较高的识别准确度。
[Abstract]:With the development of economy and the improvement of residents' living standard, population growth, environmental pollution and urbanization make the contradiction between supply and demand of water resources more and more prominent, and the problem of water security is becoming more and more prominent. The study of water use behavior of rural residents can improve the awareness of water saving and improve the weak situation of water resources management. The method proposed in this paper can accurately identify the water use behavior of rural residents and improve the current water use infrastructure. In this paper, the flow characteristics of several typical water use events are analyzed. The identification method of rural residents' water use behavior was studied. The specific work is as follows: the flow characteristics of different water use behaviors are extracted from the training set, and the identification models of different types of residents' water use behavior are established by using left to right hidden Markov model (Hidden Markov Model, HMM). The test data are input into the trained HMM to identify the residents' water use behavior, and the residents' water use events are identified according to the flow sequence of the residents' water consumption at the moment. In order to improve the accuracy of the recognition results of water use behavior using HMM, this paper combines HMM with time probability function, and obtains the recognition results of this method. The artificial neural network (Artificial NeuralNetworks,ANN) algorithm is selected, the (Back Propagation, BP) network structure of BP neural network is designed, the training parameters of BP network are determined, and the identification model of residents' water consumption behavior is established by using BP neural network. Finally, the test data are input into the trained BP neural network model to identify the behavior of residents' water use, and the recognition results are obtained. The results show that the combination model of HMM and time probability function can obtain more accurate identification results for different water flow patterns. The BP neural network model can be used to identify water events with similar flow patterns.
【学位授予单位】：西安理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP183

【参考文献】