富磷上清液铁接触除磷效果预测模型

Composite prediction model of phosphorus removal rate for anaerobic phosphate-enriched supernatant by iron-contactor process

富磷上清液铁接触除磷效果受多种因素影响,且各因素间关系复杂,单一的数学模型难于准确模拟.建立了灰色BP人工神经网络组合预测模型,该模型结合了进水总磷浓度、曝气强度、水力停留时间、pH值、水温等5个主要影响因素,可准确预测不同工况下富磷上清液铁接触除磷效果.通过组合预测模型计算可知:随着进水总磷浓度和pH值的升高,总磷去除率随之下降;而随着曝气强度的增大和水力停留时间的延长,总磷去除率随之升高;随着水温的升高,总磷去除率有波动,水温为27℃时,总磷去除率最大.

Phosphorus removal rate for anaerobic phosphate-enriched supernatant by iron-contactor process is influenced by many factors, and the relationship between these factors are complex. It is difficult to simulate accurately by single mathematical model. A composite prediction model of grey back propagation artificial neural networks （G-BP-ANNs） is established. Five main effective factors are considered in the model, namely total phosphorus content in effluent, aeration intensity, hydraulic retention time, pH value and temperature of influent. The results of prediction model fit very well with the effect of phosphorus removal under various practical conditions. It is shown that, TP （ total phosphorus） removal rate decreases with the increase of TP in the influent and pH, while increases with the increase of aeration intensity and HRT （hydraulic retention time） ; TP removal rate fluctuates with the increase of water temperature, and maximum TP removal rate is obtained at 27 ℃.