猪场沼液UF-MBR+RO处理工艺浓缩液回流的盐积累模型

反渗透(ReverseOsmosis,RO)膜工艺在沼液深度处理中发挥重要作用,其浓缩液回流引起的盐积累会降低生化阶段的效能。该研究模拟猪场沼液超滤(Ultrafiltration, UF)-膜生物反应器(Membrane Bioreactor, MBR)+RO处理工艺中浓缩液回流,构建盐积累模型预测不同污泥停留时间(Sludge Retention Time, SRT)下UF-MBR中盐积累量,分析污泥吸附作用对盐积累模型准确度的影响。结果表明:构建的盐积累模型可预测盐积累量及达到盐平衡所需的回流次数,Ca~(2+)、Mg~(2+)的实际值与理论值的拟合决定系数R~2高于0.95,RMSE小于4.00mg/L,模型对Ca~(2+)、Mg~(2+)积累量预测的准确度高。SRT从60 d降低至30 d,盐度从4.83%降低至2.63%,达到盐平衡所需的时长从249 d降低至179 d,降低SRT可作为一种有效策略来降低MBR中盐积累量及达到盐平衡的时长。SRT控制在30 d以下可使MBR盐度低于1.00%,使MBR生化阶段发挥效能的高效性。此外,污泥的吸附可降低MBR中积累的K~+、Na~+的含量。但是,Ca~(2+)、Mg~(2+)累积量较高时,污泥吸附作用对模型的影响较低,该研究构建的模型可为猪场沼液UF-MBR+RO处理工艺的应用提供参考。

Salt accumulation model for the reflux of membrane concentrate from piggery liquid digestate UF-MBR+RO treatment process

High-concentration organic wastes are often found in the liquid digestate that is derived from anaerobic digestion of manure in large-scale swine farms. There is also a contradiction between the treatment load of liquid digestate and the available land for absorption, due mainly to the high content of ammonia nitrogen, while the C/N ratio is relatively low. Thus, it requires the integration of physical or chemical technologies with biological ones for a deep treatment, since biological treatment alone cannot meet the current requirement of large-scale liquid digestate. Reverse osmosis (RO) membrane can play an important role in the deep-treatment of piggery liquid digestate. However, the reflux of concentrate in the RO process can lead to the accumulation of salinity, leading to much lower efficiency of subsequently biological treatment. In this study, a salt accumulation model was established in the nanofiltration (UF)-membrane bioreactor (MBR) under the reflux of membrane concentrate. Four kinds of salt ions were fitted with the theoretical, where the root mean square error (RMSE) was measured to evaluate the accuracy of the model. The salinity accumulation was clarified in UF-MBR under three types of sludge retention time (SRT), considering the adsorption capacity of activated sludge. The results showed that the salinity accumulation model of UF-MBR was successfully established to predict the salinity equilibrium. The running cycles were needed to achieve the required salinity. Secondly, the fitting determination coefficient (R2) of the actual values of calcium ions (Ca2+) and magnesium ions (Mg2+) in the MBR and the theoretical values predicted by the salinity accumulation model were higher than 0.95, and the RMSE was less than 4.00. Therefore, the UF-MBR salinity model here can be expected to predict the accumulation trend of Ca2+ and Mg2+ with high accuracy. Thirdly, SRT decreased from 60 d to 30 d, and the salinity decreased from 4.83 to 2.63%. The required time decreased from 249 d to 179 d for reaching salt equilibrium. The reduction of SRT can effectively alleviate the salinity accumulation in UF-MBR, while reducing the time to reach the salinity equilibrium. In addition, when the SRT was 30, 45, and 60 d, the salinity accumulated in UF-MBR should be between 0.95-2.63%, 1.29-3.59%, and 1.59-4.83%, respectively, indicating the salinity equilibrium values were higher than 1.00%. When setting the SRT within 30 d, there was alleviated inhibition of salinity accumulation on the removal performance of activated sludge pollutants. Fourthly, both the R2 of the actual values of potassium ions (K+) and sodium ions (Na+) in this process and the theoretical values predicted by the salinity accumulation model were less than 0.95, while the RMSE was higher than 15.00. The low-accuracy prediction of the UF-MBR salinity model here may be attributed to the adsorption of activated sludge, thereby reducing the content of K+ and Na+ in MBR. The effect of adsorption on the model was low, indicating a feasible model when the accumulation of Ca2+ and Mg2+ was high. This work can provide a sound reference for the future application of UF-MBR+RO treatment in piggery liquid digestates.