摘要
工业园区的VOCs状态不稳定会严重威胁环境和人类的健康。因此,需要组建VOCs预测模型对工业园区的VOCs进行预测。但是采用当前模型进行工业园区VOCs预测时,VOCs受外界因素影响较大,无法详细地拟合出VOCs不同状态,存在预测误差大的问题。为此,提出一种基于遗传算法和BP神经网络的工业园区VOCs预测模型。该模型利用BP神经网络将半导体气体试验、色谱柱试验、VOCs的浓度信息各部分对应数据和VOCs的"稳定状态"、"不稳定状态"、"超常状态"、"严重状态"作为神经网络的输入层和输出层。基于遗传理论思想优化BP神经网络构建工业园区VOCs预测拓扑结构。给出污染指数指标样本数据,组建工业园区VOCs预测模型。实验仿真证明,该模型预测可靠性高,分辨力强,为工业园的VOCs全面治理提供思路。
The unstable state of VOCs in industrial parks seriously threatens the environment and human health. Therefore, VOCs prediction model needs to be constructed to forecast VOCs in the industrial parks. However, when the current model is adopted to predict VOCs in industrial parks, VOCs are greatly influenced by external factors, so it is unable to fit different states of VOCs in details, and there is a big prediction error. To solve these problems, we proposed VOCs prediction model based on genetic algorithm and BP neural network. The model used BP neural network to take the corresponding data of semiconductor gas test, chromatographic column test, and VOCs concentration information as the input layers, and to take the “stable state”, “unstable state”, “abnormal state” and “serious state” of VOCs as the output layers. Based on the genetic theory, we optimized BP neural network to build VOCs predicted topology in industrial parks. Sample data of pollution index were given, and VOCs prediction model in industrial parks was established. The simulation results show that the proposed model has high reliability and strong resolution, which provides a way for the comprehensive governance of VOCs in industrial parks.
作者
任伟
牛玉霞
Ren Wei;Niu Yuxia(Nantong Vocational College of Science and Technology,Nantong 226007,Jiangsu,China)
出处
《计算机应用与软件》
北大核心
2018年第12期274-277,324,共5页
Computer Applications and Software
基金
南通市科技局科技计划项目(MS12016028)
