基于灰色关联度和BP网络的SO2浓度软测量模型Soft Measurement Model of SO2 Concentration Based on Gray Relation Degree and BP Networks
吴林峰
WU Linfeng
摘要(Abstract)：
燃煤发电厂SO2排放量的监测是进行大气污染源控制的基础性工作。但监测烟气环境恶劣,安装监测设备费用高、维护困难。为此,提出了一种SO2浓度预测的方法。SO2的产生受很多因素影响,利用灰色关联度分析法提取影响大的因素,然后利用选优后的参数建立BP神经网络预测模型。试验结果表明预测模型具有较高的准确性。
The monitoring of SO2 emissions from coal-fired power plant is the basic work to control atmospheric pollution sources. However, the environment of flue gas monitoring is harsh and the monitoring devices are costly and difficult in maintenance. Therefore, the paper brings forward a method of SO2 concentration prediction. The generation of SO2 is influenced by multiple factors. Therefore, factors that owns the most influence is picked up by gray relational analysis; then, the selected parameters are used to establish BP neural network prediction model. The test result shows that the prediction model owns higher accuracy.
关键词(KeyWords)：
灰色关联度;SO2排放量;BP神经网络;预测
gray relation degree;SO2 emission;BP neural network;prediction
基金项目(Foundation):
作者(Author):
吴林峰
WU Linfeng
DOI: 10.19585/j.zjdl.2015.03.012
参考文献(References)：
- [1]刘文颖,王维洲.基于灰色关联度与LSSVM组合的月度负荷预测[J].电网技术,2012,36(8):228-232.
- [2]刘庆龙,魏夕合,史俊伟.基于灰关联和神经网络的通风系统评价模型[J].矿业研究与开发,2012,32(2):63-66.
- [3]田建艳,张鹏飞,刘思峰.基于灰色关联分析的神经网络轧制力预报模型的研究[J].应用力学学报,2009,26(1):164-168.
- [4]卢建昌,王柳.基于时序分析的神经网络短期负荷预测模型研究[J].环境科学学报,2006,26(9):1553-1558.
- [5]杨国栋,赵琛瑜.环境规划中大气SO2预测的神经网络模型[J].干旱区资源与环境,2007,21(9):85-88.
- [6]郎君,苏小红,周秀杰.基于有机灰色神经网络模型的空气污染指数预测[J].哈尔滨工业大学学报,2004,36(12):1598-1601.