| RBF与Elman神经网络在人工湿地复合基质去污效果预测中的应用 |
| |
| 引用本文： | 梁启斌,刘云根,田昆,王万宾.RBF与Elman神经网络在人工湿地复合基质去污效果预测中的应用[J].环境工程学报,2013,7(4):1368-1372. |
| |
| 作者姓名： | 梁启斌 刘云根 田昆 王万宾 |
| |
| 作者单位： | 1. 西南林业大学环境科学与工程学院,昆明,650224
2. 西南林业大学环境科学与工程学院,昆明650224;国家高原湿地研究中心,昆明650224 |
| |
| 基金项目： | 云南省应用基础研究面上项目(2010CD066) |
| |
| 摘 要： | 人工湿地的去污机理复杂、呈高度非线性,故利用神经网络技术构建模型预测其长期运行效果。通过构建人工湿地复合基质模拟槽系统进行为期4个月的实验,监测得到56组COD去除率数据样本,经Matlab小波去噪后分别利用RBF和Elman网络构建动态神经网络模型,预测该系统对生活污水中COD去除效果。结果表明,RBF和Elman神经网络预测值的均方根误差分别为0.0186和0.0163,精度较高,该系统后期的COD去除率在49.4%～59.0%之间。
|
| 关 键 词： | 人工湿地基质 Elman网络 RBF网络 预测 |
| 修稿时间： | 6/4/2012 12:00:00 AM |
Application of RBF and Elman neural network in prediction on pollutant removal efficiency of constructed wetland with different compound substrates |
| |
| Liang Qibin,Liu Yungen,Tian Kun and Wang Wanbin.Application of RBF and Elman neural network in prediction on pollutant removal efficiency of constructed wetland with different compound substrates[J].Techniques and Equipment for Environmental Pollution Control,2013,7(4):1368-1372. |
| |
| Authors: | Liang Qibin Liu Yungen Tian Kun and Wang Wanbin |
| |
| Institution: | 1. College of Environmental Science and Engineering, Southwest Forestry University, Kunming 650224, China;1. College of Environmental Science and Engineering, Southwest Forestry University, Kunming 650224, China;1. College of Environmental Science and Engineering, Southwest Forestry University, Kunming 650224, China;2. National Plateau Wetlands Research Center, Kunming 650224, China;1. College of Environmental Science and Engineering, Southwest Forestry University, Kunming 650224, China |
| |
| Abstract: | The neural network model was build to predict treatment efficiency of the constructed wetland because of the complex decontamination mechanism and nonlinear. In 4 month experiment, 56 groups of COD removal rate were obtained from constructed wetland with different compound substrates. To predict the COD removal rate, the models based on the radial basis function(RBF) and Elman neural network were presented after wavelet de-noising under the environment of Matlab. The results showed that the RMS error of RBF and Elman neural network are 0.0186 and 0.0163, respectively, which means that the precision of the model is high. The COD removal rates are 49.4%~59.0%. |
| |
| Keywords: | constructed wetland substrate Elman neural network RBF neural network prediction |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《环境工程学报》浏览原始摘要信息 |
| 点击此处可从《环境工程学报》下载免费的PDF全文 |
|
