摘要:

The carbon source materials are important influencing factors in the progress of biological removal of nitrogen as the electron donors in denitrification. External carbon source materials are essential for the treatment of wastewater with low C/N ratio. For estimating the proper dosage of external carbon source, back-propagation (BP) neural network and radial basis function (RBF) neural network were introduced to develop a non-linear model between the dosage of external carbon source and influent conditions, using the experiment data from the cyclic activated sludge technology (CAST) on laboratory scale. Results show that both two networks prove to be effective in estimating the dosage of external carbon source; RBF neural network model turns out to be better in training speed and approximation capability, while BP neural network model shows higher prediction accuracy.碳源作为反硝化过程的电子供体,是影响生物脱氮过程的重要因素,低碳氮比污水需外加碳源以保证反硝化反应的顺利进行。为了优化控制碳源投加量,对实验室搭 建的CAST工艺污水处理装置的进水条件和外加碳源量的非线性关系分别进行了基于BP和RBF神经网络的模型研究,并对外加碳源量进行了预测。结果表明, 两种网络模型均能有效预测外加碳源量,RBF神经网络模型在训练速度和逼近能力方面优于BP神经网络模型,但在预测性能方面BP神经网络模型则有更高的预 测精度。

附注:

Times Cited: 01