Abstract: In this paper, the simulation of external carbon addition was studied in the continuous flow anoxic/oxic(A/O) nitrogen removal process for domestic wastewater of low carbon∶nitrogen(C/N). It is difficult to develop a parametric model for the complex non-linear relationship among additional carbon dosage, total recycling rate and effluent total nitrogen concentration in a treatment system, which makes it hard to turn the control strategies into reality. Aiming at this problem, back-propagation neural network(BPNN) was introduced into the research.A non-parametric model was developed by the training of BPNN with the experimental data. The external carbon dosage could be optimized with the model based on neural network. It provided an alternative way to the realization of on-line optimal control of a wastewater treatment system.

Key words: 神经网络, A/O脱氮, 低碳氮比, 外加碳源

摘要： 对连续流缺氧/好氧（A/O）脱氮工艺处理低碳氮比（C/N）生活污水的外加碳源系统进行了仿真研究.由于处理系统的外加碳源量、总回流比和出水总氮（TN）之间存在的复杂非线性关系，很难用常规的参数型模型进行描述，给处理系统控制策略的实现带来较大的困难.针对该问题，引入了BP神经网络，通过神经网络对试验数据的学习建立系统的非参数型模型，通过该模型对系统进行仿真研究，可以达到优化碳源投加量的目的.研究结果表明，经过训练的BP神经网络模型可以很好地模拟处理系统，根据仿真分析结果可以实现碳源投加量的优化控制，这为污水处理系统在线最优控制的实现提供了一条可行的途径.

关键词: 神经网络, A/O脱氮, 低碳氮比, 外加碳源