硝化/反硝化膜生物反应器（A/O-MBR）广泛应用于市政污水处理，膜污染问题是制约A/O-MBR进一步发展和应用的主要瓶颈，而溶解性微生物产物（SMP）是膜污染的主要贡献因子。已有研究利用人工神经网络（ANNs）方法建立MBR膜污染预测模型，模型输入指标涉及有操作参数、进水水质、水力条件、膜组件特征等，但均未以SMP作为建模因子。因此，研究SMP与膜污染间的理论逻辑关系对实现MBR系统的长期稳定运行具有重要意义。由于SMP会受到操作参数、菌群组成特征的变化影响，本项目首先研究多项操作参数对A/O-MBR系统中SMP含量及其特征的综合影响，评估各项参数影响SMP的权重分配；其次研究A/O-MBR系统中菌群组成与SMP之间的内在联系，从而获得对应最小膜污染的最佳菌群比例；基于获知的操作参数、菌群比例与SMP之间的相互作用，以ANNs为基础建立膜污染预测模型，最终揭示SMP与膜污染间的理论关系。

A/O-MBR has been widely used in municipal sewage treatment, but membrane fouling problem is the main bottleneck of A/O-MBR further development and application, and soluble microbial products (SMP) is the main contribution factor of membrane fouling. It was researched that artificial neural networks (ANNs) were utilized to establish membrane fouling prediction models about MBR, and input indicators included operating parameters, influent quality, hydraulic conditions, membrane characteristics and so on in these models, but SMP was not selected as modeling factors. Therefore, the study of theoretical relationship between SMP and membrane fouling is of great significance to the long-term stable operation of MBR system. Since SMP would be affected by operation parameters and characteristics of the flora composition, in this project, the comprehensive influence of operation parameters will be studied firstly for SMP to assess effect weights of various parameters; secondly, the internal relation will be studied between flora composition and SMP to acquire the best proportion of microbial population at that membrane fouling is minimized; finally, based on interactions among operation parameters, flora composition and SMP, the membrane fouling model will be established by using ANNs to uncover the theoretical relationship between SMP and membrane fouling.