Multimodal Machine Learning Guides Low Carbon Aeration Strategies in Urban Wastewater Treatment

The potential for reducing greenhouse gas(GHG)emissions and energy consumption in wastewater treatment can be realized through intelligent control,with machine learning(ML)and multimodality emerging as a promising solution.Here,we introduce an ML technique based on multimodal strategies,focusing specifically on intelligent aeration control in wastewater treatment plants(WWTPs).The generalization of the multimodal strategy is demonstrated on eight ML models.The results demonstrate that this multimodal strategy significantly enhances model indicators for ML in environmental science and the efficiency of aeration control,exhibiting exceptional performance and interpretability.Integrating random forest with visual models achieves the highest accuracy in forecasting aeration quantity in multimodal models,with a mean absolute percentage error of 4.4%and a coefficient of determination of 0.948.Practical testing in a full-scale plant reveals that the multimodal model can reduce operation costs by 19.8%compared to traditional fuzzy control methods.The potential application of these strategies in critical water science domains is discussed.To foster accessibility and promote widespread adoption,the multimodal ML models are freely available on GitHub,thereby eliminating technical barriers and encouraging the application of artificial intelligence in urban wastewater treatment.

Optimal Aerations in the Inverse Fluidized Bed Biofilm Reactor When Used in Treatment of Industrial Wastewaters of Various Strength

The aim of this work was the determination of the optimal aerations, and more specifically the corresponding optimal air velocities uopt, at which the largest COD removals were achieved in treatment of industrial wastewaters of various strength conducted in the inverse fluidized bed biofilm reactor. The largest COD removals were achieved at the following air velocities uopt and retention times ts, and (Vb/VR) = 0.55: i) for CODo = 72,780 mg/l at uopt = 0.052 m/s and ts = 80 h;ii) for CODo = 62,070 mg/l at uopt = 0.042 m/s and ts = 65 h;iii) for CODo = 49,130 mg/l at uopt = 0.033 m/s and ts= 55 h;iv) for CODo = 41,170 mg/l at uopt = 0.028 m/s and ts = 45 h;v) for CODo = 35,460 mg/l at uopt = 0.025 m/s and ts = 27.5 h;and vi) for CODo = 26,470 mg/l at uopt= 0.014 m/s and ts = 22.5 h. In the treatment operation conducted in a reactor optimally controlled at the above values of uopt, ts and (Vb/VR), the following decreases in COD were obtained: i) from 72,780 to 5410 mg/l;ii) from 62,070 to 3730 mg/l;iii) from 49,130 to 2820 mg/l;iv) from 41,170 to 1820 mg/l;v) from 35,460 to 1600 mg/l;and vi) from 26,470 to 1180 mg/l, that is, approximately a 93%, 94%, 95%, 96%, 95% and 96% COD reduction was attained, respectively.

Kinetic Performance of Oil-field Produced Water Treatment by Biological Aerated Filter

The biological aerated filter （BAF） was used to treat the oil-field produced water. The removal efficiency for oil, COD, BOD and suspended solids （SS） was 76.3%-80.3%, 31.6%-57.9%, 8.6.3%-96.3% and76.4%--82.7%, respectively when the hydraulic loading rates varied from 016m·h^-1 to 1.4m·h^-1. The greatest partof removal, for example more than 80% of COD removal, occurred on the top 100cm of the media in BAF. The kinetic .performance of BAF indicated that the relationship of BOD removal efficiency with the hydraulic loadingrates, in biological aerated filters could be described by c1/c1=l-exp（-2.44/L^0.59）. This equation could be used topredict the B OD.removal efficiency at different hydraulic loading rates.

Treatment of Slightly Polluted Wastewater in an Oil Refinery Using a Biological Aerated Filter Process

The slightly polluted wastewater from oil refinery contains some COD, oil pollutants and suspended solids （SS）. A small-scale fixed film biological aerated filter （BAF） process was used to treat the wastewater. The influences of hydraulic retention time （HRT）, air/water volume flow ratio and backwashing cycle on treatment efficiencies were investigated. The wastewater was treated by the BAF process under optimal conditions： the HRT of 1.0 h, the air/water volume flow ratio of about 5 ： 1 and the backwashing cycle of every 4-7 days. The results showed that the average removal efficiency of COD, oil pollutants and SS was 84.5%, 94.0% and 83.4%, respectively. And the average effluent concentration of COD, oil pollutants and SS was 12.5, 0.27, 14.5 mg·L^-1, respectively. The experimental results demonstrated that the BAF process is a suitable and highly efficient method to treat the wastewater.

Treatment performance of integrated biological aerated filter by orthogonal experiments

Biological aerated filters have many advantages such as small volume and high treatment efficiency. This research focused on sewage treatment performance of integrated biological aerated filter (IBAF) under different conditions such as aeration, hydraulic retention time and the height of fillers layer, to identify the turn of marked affecting factor of removal performance through orthogonal experiments, optimize the function parameter of IBAF, reveal the regularity of sewage treatment of IBAF under different conditions, and adopt suitable measures to guarantee excess water quality of IBAF.

Effects of Aeration Rates and Patterns on Shortcut Nitrification and Denitrification

The effects of aeration rates and aeration patterns on the oxidation of ammonia-nitrogen into nitrite were investigated. The influent high ammonia-nitrogen synthetic wastewater resembled to those of the catalytic process of the petrochemical refinery. The method involved the biological shortcut nitrification and denitrification lab-scale’s sequencing batch reactor (SBR) process based on intermittent aerations and aeration patterns. All the operations were carried out in a 20 L working volume SBR bioreactor, and the influent synthetic wastewater’s concentration was always 1000 mg/L ammonia-nitrogen NH4-N concentration at a C/N (carbon/nitrogen) ratio of 2.5:1. Effective shortcut nitrification to nitrite was registered at 1.1 mg-O2/L (i.e. 9 L-air/min) with 99.1% nitrification efficiency, 99.0% nitritation rate and 2.6 mg-NO3--N/L nitrate concentration. The best results with 99.3% nitrification efficiency were recorded when operating at 1.4 mg-O2/L (i.e. 12 L-air/min). According to these experiments, it results that the nitrite accumulation rate was related to aeration rate and cycle’s duration. However, at 1.7 mg-O2/L (i.e. 15 L-air/min), the system was limited by an increase in nitrate concentration with more than 5 mg/L which could be a point of reverse to conventional nitrification. The best total nitrogen (TN) removal was about 71.5%.

Effect of media heights on the performance of biological aerated filter

The optimum media height of carbon oxidation and nitrification in a down-flow biological aerated filter was determined, and the distribution of the heterotrophic and nitrifying populations through studying the changes of organic carbon contents and ammonia concentration at different media height was got. The results showed that as a down flow BAF with granular media, the active layer of nitrifiers was deeper than heterotrophs in BAF. And the optimum media height for the removal of SS, COD Cr and NH + 4-N was 40 cm,60 cm and 80 cm respectively. The removal efficiency of SS, COD Cr and NH + 4-N was 79.1%, 63.9% and 96.4% respectively under the influent COD Cr and NH + 4-N of 122.1 mgCOD Cr /L and 14.84 mgNH + 4-N/L, the influent flux of 15.8 L/h, air to liquid ratio of 3∶1.

Common effluent treatment plant(CETP):Reliability analysis and performance evaluation

Common effluent treatment plants(CETPs) have been installed and are in operation at numerous industrial clusters throughout India. They serve to reduce effluent treatment cost, provide better collective treatment, and reduce land cost for small-scale industrial facilities that cannot afford individual treatment plants. Optimum working conditions for treatment of effluent to be at par with discharge standards is a major mandate for any CETP. In this study, the reliability and removal efficiencies(REs) of a CETP in the industrial area of Maharashtra State in India were examined. An established methodology was adopted to determine the effectiveness of the CETP in terms of biochemical oxygen demand(BOD), chemical oxygen demand(COD), total suspended solids(TSS), and oil and grease(O&G) concentrations. The CETP's compliance with respect to design standards and its operation were studied in detail. This paper highlights the results of RE and the coefficient of determination(R2) values obtained from the CETP data, estimates the pollutants removed at the highest and lowest rates over a period of time, and highlights the reasons for problem areas along with remedial measures. It was observed that, except O&G, all the parameters(BOD, COD, and TSS)showed fluctuations in removal efficiencies and their reliabilities. This situation can be improved by releasing effluent containing hydraulic and organic loading to the CETP as per standards and optimizing treatment processes, especially primary clari-flocculators and aeration tanks, both of which are important units of any CETP.

市政污水生化处理过程中曝气控制系统研究进展

为了促进污水处理厂(WWTPs)的低碳可持续运营管理,近年来致力于改善出水质量、削减运营成本和降低温室气体排放的智能曝气控制技术蓬勃发展,但关于可持续污水处理厂的最佳曝气控制策略尚未达成共识。系统综述了曝气过程中溶解氧浓度对于N2O释放的影响,通过分析对比不同曝气控制策略的差异及优缺点,提出曝气控制策略进一步优化的建议,为实现市政污水处理厂溶解氧的综合科学管理提供理论支撑,旨在为污水处理行业提供科学依据和技术支持,促进环境保护和可持续发展。在曝气控制系统中加入N2O排放量化模型来实现出水水质、曝气能耗、温室气体排放三者之间的平衡是未来的关键研究方向,需开发更准确、更合适的温室气体量化模型、灵活的控制方法和智能曝气策略,以满足污水处理厂更多的需求。

二元复合驱采出水达标处理工艺研究

二元复合驱采出水中存在较多的驱油剂、聚合物和表面活性剂,其处理难度高于常规水驱或聚驱的采出水。为实现此类采出水的达标处理,分析了聚合物、表面活性剂浓度对水质特性的影响,采用曝气-生化-物理过滤流程对水样进行深度处理,并进行了现场验证。结果表明,当聚合物浓度从0mg/L增加~1000mg/L时,水样浊度和Zeta电位绝对值逐渐增大;当曝气强度为12h,气水比为30:1,生化处理时间8d,并采用纤维球作为滤料时,其处理效果最佳,滤后采出水的油、悬浮物和粒径中值均满足规范要求,符合达标回注的要求。

数据驱动的污水处理曝气过程低碳优化控制方法

针对现有的污水处理过程存在碳排放机理不清且难以评估,无法通过有效的调控方式降低碳排放总量的问题,设计了一种数据驱动的污水处理曝气过程低碳优化控制方法。首先,通过深入分析碳排放影响因素及其与水质指标的相互关系,获得了曝气过程各水质指标和碳排放之间的关联关系;其次,采用数据驱动的方法,设计了曝气过程能耗与碳排放的优化模型,以获取曝气过程最优化的控制策略;最后,将获取的曝气过程低碳优化控制方法应用于基准仿真模型。测试结果说明该方法能够有效地跟踪控制曝气过程,降低能耗与碳排放量总量。

移动床生物膜反应器-生物滤池脱氮除磷中试研究

采用移动床生物膜反应器-生物滤池(MBBR-BF)一体化污水处理设备进行实际污水中试试验,考察曝气量对一体化设备挂膜启动阶段污染物去除效果的影响,探究低温环境下一体化设备适应性及污染物处理能力。结果表明:曝气量降低对COD和TN去除率影响较小,而NH_(3)-N与TP的去除率随之下降,NH_(3)-N平均去除率由87.46%降至48.64%,TP平均去除率由72.41%降至54.66%;常温环境下COD、NH_(3)-N、TN、TP平均去除率分别为92.48%、94.41%、49.39%和61.63%,低温环境运行对污染物去除效果有明显影响。中试应用试验状况良好,一体化设备在冬季低温环境下运行状况正常,可为农村生活污水的有效处理提供重要参考。

微生物强化曝气-抽提处理低含油污泥效能研究

采用微生物固定化辅以AS曝气抽提处理处置含油污泥。以现场油泥为来源,筛选了5株石油降解菌,室内实验结果表明,其中WJ-1菌、WJ-4菌,WJ-5菌降解效率相对WJ-2,WJ-3较好。5种菌株均随着培养基含油率增加,降解效率降低,但高于10%含油率时,5种菌株降解效率均显著降低。利用以石油降解菌为核心的生物技术,采用麦麸为原料,通过包埋固定化的方式制备生物菌剂,并设计曝气-抽提反应堆。在优化曝气参数的条件下,分别投加液体菌剂5 m^(3)和固体菌剂10 m^(3)进行实验,曝气量为15 m^(3)/h,连续处理40 d后,污泥含油量分别为1.76 mg/g和1.82 mg/g,含油率小于0.3%,实现了污泥的无害化处理。

生物炭强化生物滤池处理农村生活污水效果及碳排放

为研究生物炭对生物滤池强化作用及其在“双碳”背景下碳排放量与去除效果相结合的评估方法,以秸秆生物炭强化曝气生物滤池为研究对象,在中温环境(25~30℃)下,探究了生物炭强化生物滤池的挂膜启动时间、处理能力、稳定性以及日碳排放量。结果表明,滤池的挂膜时长为14 d;当水力负荷和进水流量分别为0.045m^(3)/(m^(2)·h)和0.9 m^(3)/d时,系统的最佳进水水力停留时间为8 h,化学需氧量、氨氮、总氮和总磷的平均去除率分别为95.0%、79.1%、62.5%、78.4%;当水力停留时间和进水流量分别为8 h和1.2 m^(3)/d,系统最佳的水力负荷为0.030 m^(3)/(m^(2)·h),平均去除率分别为95.3%、87.4%、68.1%、79.0%;在系统运行过程中,不同时期悬浮球与陶粒上胞外聚合物(extracellular polymeric substances,EPS)含量相比刚驯化污泥上EPS的含量都有所增加,表明了生物滤池在实际运行过程中比较稳定且污泥活性也比较好;在8 h的最佳水力停留时间以及进水流量0.9 m^(3)/d的条件下,曝气生物滤池使污水达标排放的碳排放量为9.08 kg/d,因此采用生物炭强化曝气生物滤池处理农村生活污水具有较高的应用前景。

基于遗传算法优化BP神经网络的曝气量预测

针对城市污水处理厂生化池曝气量的冗余问题,从节能降耗角度出发,提出了一种精确预测生化池曝气量的方法。首先,对从山东某污水厂采集的数据做预处理;其次,经相关性分析,建立以进出水指标和历史曝气量为输入、以当前曝气量为输出的反向传播(back propagation,BP)神经网络模型,并利用遗传算法优化BP神经网络的权值和阈值;最后,为了验证所提方法的有效性,将其与传统BP神经网络模型进行对比。实验结果表明,所提方法的均方根误差减小了39.88%,计算时间缩减了31.00%;且与实际测试数据相比,每月的曝气电费可节约892.80元。因此,所提方法对生化池曝气量的预测是有效和可行的。

纤维素乙醇废水生化出水深度处理中试研究

考察了“酸析→电化学氧化/中和沉淀→曝气生物滤池(BAF)”深度处理工艺处理纤维素乙醇生产废水生物处理出水的可行性。结果表明:酸析最佳反应pH为3,电化学氧化反应最佳停留时间为0.5 h,最佳的COD与H_(2)O_(2)(35%)质量比为1∶1。“酸析→电化学氧化/中和沉淀→曝气生物滤池(BAF)”工艺可用于纤维素乙醇生产废水深度处理工艺,运行稳定可靠,工艺最终出水COD平均质量浓度为(63.8±29.4) mg·L^(-1),NH_(4)^(+)-N平均质量浓度为(5.8±2.5) mg·L^(-1),该组合工艺最终出水可稳定达标。

Effect of Intermittent Aeration on the Treatment Performance in a Submerged Membrane Bioreactor

In order to improve removal for nitrogen in a pilot-scale submerged membrane bioreactor（SMBR）,intermittent aeration was conducted,and the effect on the treatment performance under four kinds of operation condition（run 1,continuous aeration;run 2,60/60 min aeration on/off time;run 3,60/90 min aeration on/off time;run 4,60/75 min aeration on/off time） was evaluated.The results showed that depending on the specific on/off of the aeration time ratio,removal efficiency of nitrogen could be improved significantly,and the removal rates of total nitrogen（TN） under different operation conditions were 28.0%,59.5%,66.8% and 70.7%,respectively.There were no obvious differences for removal rates for CODCr and ammonia among different operation conditions.In general,intermittent aeration could be used as a feasible way to improve treatment performance for nitrogen in the SMBR.

Effects of germination and aeration treatment following segmented moisture conditioning on theγ-aminobutyric acid accumulation in germinated brown rice

Brown rice was treated by segmented moisture conditioning to reach the suitable water content and aerated with air for germination.The effects of germination and aeration treatment on theγ-aminobutyric acid(GABA)content in germinated brown rice were studied.The results showed that the germination rate,glutamic acid,glutamate decarboxylase activity and GABA content in germinated brown rice increased significantly and then decreased with the increase of germination and aeration treatment parameters.Correlation analysis also revealed that there was a significant positive correlation between GABA accumulation and glutamic acid content,glutamate decarboxylase activity.These results suggested that the aeration treatment during germination following segmented moisture conditioning could contribute to a high GABA content of germinated brown rice.

融合SVMD和IGJO-LSTM的污水处理曝气量预测

污水处理过程中,曝气量数据波动大、周期性不明显,现有模型难以对曝气量进行准确的预测。因此,提出了一种优化模型,此模型利用逐次变分模态分解提取曝气数据特征,并采用改进的金豺算法优化长短期记忆网络的超参数,以提升模型预测能力。首先,针对实际污水数据复杂的问题,利用逐次变分模态分解算法分解重构原始曝气数据序列。其次,用长短期记忆网络分别对每个序列依次预测,并采用柯西反向学习混合变异策略改进金豺算法对长短期记忆网络参数进行优化。最后,将各个序列预测结果进行重组,得到最终预测值。利用实际污水水质数据对该模型进行验证,结果表明该模型有效提高了曝气量的预测精度,具有很好的应用前景,能很大提升污水处理厂的经济效益。