复合人工湿地磷去除混合模型研究

发布时间：2017-12-30 22:00

本文关键词：复合人工湿地磷去除混合模型研究　出处：《西南交通大学》2016年硕士论文　论文类型：学位论文

【摘要】：随着人口的增长与经济的快速发展,城市污水不断产生,人工湿地作为一种环境友好型污水处理技术得到广泛应用。但是人工湿地的设计和预测以黑箱模型为主,无法准确可靠的评价湿地的运行效果,限制了人工湿地技术的开发利用。为了推动人工湿地技术的发展,构建一个简便易行且准确可靠的预测模型势在必行,为此开展了以下工作：1、分析前人的研究成果,了解人工湿地的结构、特点、污染物去除机理及影响因素,选取研究较少的TN(总氮)、TOC(总有机碳)作为待研究的影响因素。2、以成都市凤凰河二沟人工湿地为研究对象,通过对实际工程运行效果的监测,分析了TN、TOC等因素对湿地中磷去除的影响及各子系统对磷去除的贡献,发现强化生物一级植物塘对磷去除量及去除率最高、二级生物塘次之、之后是植物砂滤池、二级植物碎石床最低,而一级植物碎石床不仅不去除磷,还在一定程度上增加污水中的磷含量；湿地对TP(总磷)的去除量总体上处于稳定状态；出水磷的浓度随进水磷浓度的增大而增大；TN、TOC的含量随着总磷含量的减小而减小。根据计算得出的相关系数以及决定系数说明出水磷浓度和进水磷浓度间、TOC与TP间、TN与1]P间具有很强的线性相关性。3、以Monod模型模拟湿地内部生化反应,结合径向基函数(RBF)人工神经网络模型考虑TN、TOC等因素对磷去除的影响,结合两者的优点,共同构建复合人工湿地的Monod-RBF混合模型。通过与单独的RBF神经网络模型和机理模型比较,混合模型的绝对误差机理模型RBF模型；混合模型的相对误差机理模型RBF模型；混合模型的预测标准误差SEP机理模型RBF模型,说明混合模型有更高的预测精度,能更好的预测人工湿地的处理效果。
[Abstract]:With the growth of population and the rapid development of economy, urban sewage is produced continuously. Artificial wetland as an environment-friendly sewage treatment technology has been widely used, but the design and prediction of constructed wetland is mainly based on black box model, which can not accurately and reliably evaluate the operation effect of wetland. In order to promote the development of artificial wetland technology, it is imperative to build a simple and accurate and reliable prediction model. Analyze the previous research results, understand the structure, characteristics, pollutant removal mechanism and influencing factors of constructed wetland, select TNN (Total nitrogen), which is less studied. TOC (Total Organic carbon) is the influencing factor to be studied. Taking the constructed wetland of Fenghuanghe River in Chengdu as the research object, the TN is analyzed by monitoring the actual engineering operation effect. The effects of TOC and other factors on phosphorus removal in wetland and the contribution of each subsystem to phosphorus removal were found to be the highest in the enhanced biological primary plant pond followed by the secondary biological pond and the plant sand filter. The secondary plant gravel bed was the lowest, while the first stage plant gravel bed not only did not remove phosphorus, but also increased the phosphorus content in sewage to a certain extent. The removal rate of TP- (total phosphorus) was stable in the wetland. The effluent phosphorus concentration increases with the increase of influent phosphorus concentration. The content of TOC decreased with the decrease of total phosphorus content. According to the calculated correlation coefficient and the determining coefficient, the relationship between effluent phosphorus concentration and influent phosphorus concentration between TOC and TP was explained. There is a strong linear correlation between TN and 1] P. The biochemistry reaction in wetland is simulated by Monod model, and the TN is considered by combining radial basis function (RBF) artificial neural network model. The effect of TOC and other factors on phosphorus removal, combined with the advantages of the two, the Monod-RBF mixed model of composite constructed wetland was constructed. The model was compared with the single RBF neural network model and mechanism model. The absolute error mechanism model of mixed model RBF model; The relative error mechanism model RBF model of hybrid model; The prediction standard error SEP mechanism model RBF model of mixed model shows that the mixed model has higher prediction accuracy and can better predict the treatment effect of constructed wetland.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：X703

【相似文献】