BP神经网络对蚯蚓滤池处理COD的模拟预测

郭飞宏; 张继彪; 郑正

环境工程学报

, ,

郭飞宏, 张继彪, 郑正. BP神经网络对蚯蚓滤池处理COD的模拟预测[J]. 环境工程学报, 2012, 6(4): 1197-1201.

引用本文:

郭飞宏, 张继彪, 郑正. BP神经网络对蚯蚓滤池处理COD的模拟预测[J]. 环境工程学报, 2012, 6(4): 1197-1201.

Guo Feihong, Zhang Jibiao, Zheng Zheng. Simulation and prediction based on BP neural network for COD treatment by earthworm filter[J]. Chinese Journal of Environmental Engineering, 2012, 6(4): 1197-1201.

Citation:

Guo Feihong, Zhang Jibiao, Zheng Zheng. Simulation and prediction based on BP neural network for COD treatment by earthworm filter[J]. Chinese Journal of Environmental Engineering, 2012, 6(4): 1197-1201.

BP神经网络对蚯蚓滤池处理COD的模拟预测

1.
上海环境科学研究院，上海 200233； 2.南京大学环境学院污染控制与资源化研究国家重点实验室，南京 210093
3.
复旦大学环境科学与工程系，上海 200433
基金项目:
国家水体污染控制与治理科技重大专项(2008ZX07101-004)
中图分类号: X703.1

Simulation and prediction based on BP neural network for COD treatment by earthworm filter

1.
Shanghai Academy of Environmental Sciences， Shanghai 200233， China； 2. State Key Laboratory of Pollution Control and Resource Reuse， School of the Environment， Nanjing University， Nanjing 210093， China
3.
Department of Environmental Science and Engineering， Fudan University， Shanghai 200433， China
Fund Project:

摘要: 基于蚯蚓滤池处理去除污染物的非线性特点，利用BP神经网络建立了蚯蚓滤池处理COD的基本模型结构。同时对实验数据进行了验证和预测，通过权值贡献率分析确定了各种输入因素对COD出水浓度的影响。结果表明：COD的出水模型预测值与实际值平均误差较小，模型稳定，预测效果好。输入神经元为4，隐含神经元为8，输出神经元为1，学习速率为0.1，动量为0.1，训练次数为10 000的BP神经网络模型，预测的COD出水值最接近真实值。COD进水浓度对COD出水影响最大，符合理论研究结果。BP神经网络模型建立的成功为后续生活污水智能化控制的研究提供了相应的理论基础。
- 蚯蚓滤池 /
- BP神经网络 /
- 模型 /
- 预测
Abstract: Based on nonlinear character of contaminants removal by earthworm filter， the basic simulation and prediction model of COD removal was established under the use of Back Propagation (BP) neural network. Experimental data was also forecasted and validated in this research， by analysis of weight contribution rates， identified a variety of input factors on COD effluent concentration effect. Results showed that the mean error of model predictive value and actual value was small， model was stable and had a good forecasting effect. This kind of BP neural model， that 4 input neurons， 8 hidden neurons， 1 output neuron， 0.1 learning rate， 0.1 momentum and 10 000 training times， predicted results closest to the true value. The greatest impact factor was inflow COD， corresponding with other theories. BP neural model provided appropriate theory for the successful application of intelligent control in sewage treatment.
- earthworm filter /
- BP neural network /
- model /
- prediction

点击查看大图

计量

文章访问数: 1721
HTML全文浏览数: 607
PDF下载数: 1163
施引文献: 0

郭飞宏, 张继彪, 郑正. BP神经网络对蚯蚓滤池处理COD的模拟预测[J]. 环境工程学报, 2012, 6(4): 1197-1201.

引用本文:

郭飞宏, 张继彪, 郑正. BP神经网络对蚯蚓滤池处理COD的模拟预测[J]. 环境工程学报, 2012, 6(4): 1197-1201.

Citation:

BP神经网络对蚯蚓滤池处理COD的模拟预测

1. 上海环境科学研究院，上海 200233； 2.南京大学环境学院污染控制与资源化研究国家重点实验室，南京 210093
3. 复旦大学环境科学与工程系，上海 200433

收稿日期: 2011-03-24

基金项目:

国家水体污染控制与治理科技重大专项(2008ZX07101-004)

关键词:

蚯蚓滤池 /
BP神经网络 /
模型 /
预测

摘要: 基于蚯蚓滤池处理去除污染物的非线性特点，利用BP神经网络建立了蚯蚓滤池处理COD的基本模型结构。同时对实验数据进行了验证和预测，通过权值贡献率分析确定了各种输入因素对COD出水浓度的影响。结果表明：COD的出水模型预测值与实际值平均误差较小，模型稳定，预测效果好。输入神经元为4，隐含神经元为8，输出神经元为1，学习速率为0.1，动量为0.1，训练次数为10 000的BP神经网络模型，预测的COD出水值最接近真实值。COD进水浓度对COD出水影响最大，符合理论研究结果。BP神经网络模型建立的成功为后续生活污水智能化控制的研究提供了相应的理论基础。

English Abstract

Simulation and prediction based on BP neural network for COD treatment by earthworm filter

1. Shanghai Academy of Environmental Sciences， Shanghai 200233， China； 2. State Key Laboratory of Pollution Control and Resource Reuse， School of the Environment， Nanjing University， Nanjing 210093， China
3. Department of Environmental Science and Engineering， Fudan University， Shanghai 200433， China

Received Date: 2011-03-24

Fund Project:

Keywords:

Abstract: Based on nonlinear character of contaminants removal by earthworm filter， the basic simulation and prediction model of COD removal was established under the use of Back Propagation (BP) neural network. Experimental data was also forecasted and validated in this research， by analysis of weight contribution rates， identified a variety of input factors on COD effluent concentration effect. Results showed that the mean error of model predictive value and actual value was small， model was stable and had a good forecasting effect. This kind of BP neural model， that 4 input neurons， 8 hidden neurons， 1 output neuron， 0.1 learning rate， 0.1 momentum and 10 000 training times， predicted results closest to the true value. The greatest impact factor was inflow COD， corresponding with other theories. BP neural model provided appropriate theory for the successful application of intelligent control in sewage treatment.

全文HTML

参考文献 (0)

下载: 全尺寸图片幻灯片

返回文章

用微信扫码二维码

分享至好友和朋友圈

BP神经网络对蚯蚓滤池处理COD的模拟预测

Simulation and prediction based on BP neural network for COD treatment by earthworm filter

计量

出版历程

BP神经网络对蚯蚓滤池处理COD的模拟预测

English Abstract

Simulation and prediction based on BP neural network for COD treatment by earthworm filter

全文HTML

目录

全文HTML