定量构效模型预测苯系物对活性污泥的生物毒性

Predicting bio-toxicity of benzene and its derivatives to activated sludge by QSARs model

由于缺少有机化合物对活性污泥处理系统毒性方面的数据,所以促生了用定量构效模型(QSARs)来评价和预测废水中有机物对废水生化处理过程的影响。有机物对于活性污泥的毒性采用半抑制率浓度(C50%)来表征,通过在原水中分别添加不同的苯系物,并利用活性污泥的中期(24h)和短期(30min)呼吸速率的改变,计算出各物质对污泥活性的半抑制率浓度。利用试验所获得的各物质的半抑制率浓度建立QSARs模型,并预留一部分数据以检测模型的可靠性。检测结果表明中期模型比短期模型更稳定更精确。在所有的模型参数中,logKow和Elumo是最具有显著性影响的参数。多维二次回归模型被证明比线性模型有更好的内部和外部可靠性。通过此方法建立的QSARs模型对其它几种苯系物的半抑制率浓度的预测误差平均在10%左右。

Due to the deficiency of the toxicity data of organic compounds to activated sludge treatment system, the influence of organic compounds on wastewater biochemical treatment was appraised and predicted by QSARs model. The half inhibitory coneentrations（Cs0~） of different compounds to sludge activity were calculated through adding different benzene derivatives into the raw water and consulting the change of respiration rate of activated sludge in mid-term（24 h） and short-term（30 min）, and were used to characterize the toxicity of organic compounds to activated sludge. Based on the half inhibitory concentrations obtained from the experiment, QSARs model was established, and its reliability was examined by the reserved data. The results showed that, the mid-term model were more stable and accurate than the short-term model. Among all the model parameters, logKow and Elumo were two of the most significant impact parameters. Multi-dimension quadratic regression model was proved to have better internal and external reliability than linear model. The prediction errors of C50% of QSARs model established by the said method to several other benzene derivatives were about 10% averagely.