中文摘要：

生物反应器（ MBR）工艺在现代污水处理中扮演着重的角色，然而膜污染严重影响了MBR工艺的性能，膜污染导致的最直接的后果就是膜通量的下降，膜通量的下降直接影响MBR污水处理的效果。为了有效、快速地预测MBR膜通量，利用随机森林（ RF）算法建立MBR膜通量预测模型。选取影响膜通量的主因子作为随机森林预测模型的输入，膜通量作为输出，建立MBR膜通量影响因素和MBR膜通量之间的非线性关系。首先利用训练集在随机森林预测模型上进行训练，然后用训练好的随机森林预测模型进行膜通量预测。通过预测数据和实验数据的对比，得出该算法对膜通量有较高的预测精度；为了进一步验证该算法的有效性，建立了BP神经网络预测模型，将两者进行比较，对比结果表明随机森林预测模型具有更高的预测精度。

英文摘要：

Membrane Bio-Reactor （ MBR ） technology plays an important role in wastewater treatment, but the performance of the MBR technology is seriously affected by the membrane fouling. In general, the result of membrane fouling is decline of MBR membrane, so the size of the membrane flux determines the extent of contamination. To identify MBR membrane flux accurately, this paper used Random Forest （ RF） algorithm to build MBR membrane flux prediction model, selected the important factors that affect the membrane flux as the input of random forest prediction model, flux as output, to build the non-linear relationship between the influence factors and MBR membrane flux. Firstly, the training set is trained on Random forest prediction model, and then the trained random forest prediction model was used to predict the membrane flux. The study can obtain that the algorithm has higher prediction accuracy by comparing the predicted results and experimental results. To verify the effectiveness of random forest, by comparing it with BP neural network model, the results of comparison show that the proposed prediction model has higher prediction accuracy.