前置生物膜污水处理系统中的数据融合技术研究

发布时间：2018-11-18 09:23

【摘要】：水是生命之源,但在世界范围内各种水资源正遭受着不同程度的污染。污水处理是解决水污染的主要途径。将曝气生物滤池与前置生物膜脱氮技术相结合组成的前置生物膜脱氮污水处理系统,不仅能够有效降解有机物,还能最大限度的去除污水中的含氮量。然而曝气系统能耗大,运行成本高,若能对出水COD浓度进行预测建模,可以减少测量传感器的使用,且有助于污水处理系统的自适应调节,进而降低系统能耗,节省开支。本文依托与某环保科技公司联合承担的安徽省重大科技攻关项目,以ABAF/OBAF前置生物膜脱氮污水处理工艺及其应用项目为基础,建立对出水COD浓度的预测模型。一方面,在物料守恒原理和Monod方程的基础上,依据采集数据和参数经验值确定有机物降解模型数学表达式。考虑到污水中不确定因素的随机干扰,采用随机过程理论对降解模型进行改进,建立对应的模型微分方程并求解,确定最终模型表达式,进而得到出水COD浓度的预测轨迹。另一方面,在贝叶斯理论和极大似然估计的数据融合思想上,建立基于高斯过程的出水COD浓度预测模型,并结合主成分分析法对模型输入参数初值进行设定,最后用Matlab编程软件验证模型的预测效果。对比以上两个模型的预测结果,得出结论:基于数据融合的高斯过程预测模型的预测效果更加接近实际测量值。
[Abstract]:Water is the source of life, but all kinds of water resources are being polluted in different degrees all over the world. Sewage treatment is the main way to solve water pollution. The biological aerated filter (BAF) combined with the pre-biofilm denitrification system can not only effectively degrade organic matter, but also remove nitrogen content in wastewater as much as possible. However, the aeration system has the advantages of large energy consumption and high operating cost. If the effluent COD concentration can be predicted and modeled, the use of measurement sensors can be reduced, and the adaptive regulation of sewage treatment system will be helpful, thus reducing the system energy consumption and saving expenses. Based on the important scientific and technological research project of Anhui province jointly undertaken with a certain environmental protection technology company, the prediction model of effluent COD concentration was established on the basis of ABAF/OBAF pretreatment biofilm denitrification wastewater treatment process and its application project. On the one hand, on the basis of the principle of material conservation and Monod equation, the mathematical expression of organic degradation model is determined based on the collected data and the empirical values of the parameters. Considering the random interference of uncertain factors in wastewater, the degradation model is improved by stochastic process theory, the corresponding differential equation of model is established and solved, and the final model expression is determined, and then the predicted trajectory of effluent COD concentration is obtained. On the other hand, on the basis of Bayesian theory and data fusion idea of maximum likelihood estimation, a prediction model of effluent COD concentration based on Gao Si process is established, and the initial values of input parameters of the model are set up with principal component analysis (PCA). Finally, the prediction effect of the model is verified by Matlab programming software. By comparing the prediction results of the two models, it is concluded that the forecasting effect of Gao Si process prediction model based on data fusion is closer to the actual measurement value.
【学位授予单位】：安徽工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：X703;TP202

【相似文献】