【作者】 李振亮；

【导师】 张代钧；

【作者基本信息】 重庆大学 ， 环境科学与工程， 2014， 博士

【摘要】 活性污泥处理工艺是最广泛应用的废水生物处理技术。研究活性污泥的絮凝动力学，分析絮凝形成絮体的粒径分布及其演化过程、影响因素，对于完善活性污泥流体动力学，理解反应池中的污泥分布和二沉池中的泥水分离情况，以及研究江河、湖泊、海岸中自然絮体的形成与聚集等，都具有重要意义。基于群体平衡（Population Balance）的活性污泥絮凝动力学是在絮体质量和数量守恒的假设下研究絮凝过程中絮体的粒径分布及其变化，目前还存在一些需要解决的问题和改进的方面，如有效的絮体聚并效率模型和子颗粒粒径分布函数的构建，絮体粒径分布随时间演化过程的模拟，影响絮凝的关键因素（如EPS含量和Zeta电位）与絮凝动力学参数间的关系等。为此，本论文进行了以下的研究：①应用激光粒度仪测量了好氧活性污泥絮体在絮凝过程的粒径分布，研究了速度梯度、VSS/SS、EPS含量及Zeta电位对絮体粒径分布的影响。结果表明：絮体平均粒径与速度梯度显著负相关，与Kolmogrov尺度数量级基本相当，其间差异性与污泥VSS/SS、絮体强度等有关；相同的速度梯度下，絮体平均粒径与VSS/SS或EPS含量显著正相关，与Zeta电位负相关；有机质和EPS会增强絮凝效果，EPS中蛋白质比多糖对絮凝的促进作用更明显。②应用显微图像分析方法，计算得到活性污泥絮体的分形维数。结果表明：絮体二维分形维数与VSS/SS（或EPS含量）负相关，且与絮体的层次结构有关（大絮体与小絮体的分形维数有差异）；活性污泥絮体三维分形维数与絮体粒径负相关，符合幂函数关系式。③改进了基于群体平衡的活性污泥絮凝动力学模型：提出了由絮体破碎形成的子颗粒的二项式分布函数，用二项式分布参数Cp表征絮体破碎的主要方式（破裂或侵蚀）；在模型参数校核中，提出“平衡因子”——即破碎系数A与聚并效率α的比，不仅减少了模型需要校核的参数数量，更有助于简捷地实现絮凝过程中絮体粒径分布随时间演化过程的模拟。模拟结果表明：1）模型能够较好地模拟絮凝过程中絮体的粒径分布及其演化过程；2）平衡因子与平均速度梯度呈幂函数关系，随着速度梯度的增加而增加；3）聚并效率与平均速度梯度呈幂函数关系，随着速度梯度的增加而减小；4）破碎系数仅与污泥类型有关，与速度梯度无关；5）絮体颗粒区间划分数量不影响参数校核结果，区间划分数量愈多，絮体粒径分布的模拟精度愈高；6）子颗粒二项式分布较二元分布对絮体粒径分布的模拟精度更高。④对絮体聚并效率模型和破碎频率模型进行了研究和修正，结果表明：1）与聚并效率为常数相比，认为聚并效率与絮体粒径差有关的Friedlander模型和Pruppacher and Klett模型可提高絮体粒径分布的模拟精度；2）文中所构建的Pruppacher and Klett修正模型可以适用于不同速度梯度时絮体粒径分布演化过程的模拟；3）Kusters破碎频率模型较Pandaya and Spielman破碎频率模型可改善大粒径絮体的体积分数分布模拟，实现平均粒径的准确模拟；4）假设絮体的破碎频率与最大粒径成正比，且存在一个最大的破碎频率，文中所构建的Pandaya andSpielman修正模型，可实现絮体粒径分布和平均粒径的同时准确模拟。⑤基于絮凝动力学模型，分析了影响活性污泥絮凝的关键因素（EPS含量和Zeta电位）与絮凝动力学参数间的关系，结果表明：1）平衡因子A/α与EPS含量或Zeta电位线性负相关；EPS含量愈大或Zeta电位愈小，A/α愈小，更有利于絮凝，絮体的平均粒径愈大；2）子颗粒二项式分布可以更好地描述不同EPS含量或Zeta电位的絮体粒径分布的宽度；二项式分布参数Cp与EPS含量或Zeta电位线性正相关，说明不同EPS含量或Zeta电位的活性污泥絮体的破碎方式可能存在差异——当EPS含量增加时，絮体的破碎方式可能由侵蚀为主逐渐演变为破裂为主；添加Al3+时絮体可能以侵蚀的破碎方式为主，絮体的稳定性较强；添加Ca2+时絮体可能以破裂的破碎方式为主，絮体的稳定性较弱；相对于Ca2+，Al3+对絮凝的促进作用更为显著。更多还原

【Abstract】 Activated sludge process has been widely used in wastewater biological treatment.Study on the flocculation dynamics of activated sludge, which focuses on the evolutionand effect factors of Floc Size Distribution (FSD) during flocculation process, has greatsignificance to the development of activated sludge fluid dynamics, the understandingof the distribution and sedimentation behavior of activated sludge in reactor orsedimentation tank and the investigation of the aggregation and formation of flocs innatural rivers, lakes and coasts.Flocculation dynamics of activated sludge based on population balance frameworkaims at the evolution of FSD during flocculation process under the assumption of theconservation of mass and quantity of flocs. However, some problems in this fieldremain unresolved and further research is needed, including the effective model forcollision efficiency and daughter particle distribution function, the modelling of theevolution of FSD during flocculation process, and the relationship between kineticparameters of kernel structures for aggregation and breakage and the key factors (EPScontent and Zeta potential) that affect the flocculation. In this thesis, the followingwork was conducted and conclusions were drawn:①FSD of aerobic activated sludge during flocculation process was measured bylaser particle size analyzer, and the influence of velocity gradient, VSS/SS, EPScontent and Zeta potential on FSD was investigated. The results showed that the flocvolume-average size was negatively correlated to velocity gradient, and theorder-of-magnitude of the floc volume-average size was equivalent to that ofKolmogorov scale (their differences were depended on sludge VSS/SS, floc strengthand etc). As fixing velocity gradient, the floc volume-average size was positivelycorrelated to VSS/SS and EPS, whereas negatively to Zeta potential. Organic matterand EPS play important roles on the flocculation of activated sludge by enhancing thefloc strength and improving the flocculation effect. Compared with polysaccharide,protein in EPS seems to be more beneficial to the flocculation of activated sludge.②The equivalent and maximium diameter of aerobic activated sludge floc weredetermined using microscopy and image analysis and thus the2D fractal dimension D2were obtained by fitting the maximum diameter and equivalent diameter. It was foundthat2D fractal dimension was negatively related to VSS/SS (or EPS content), and seemed to be related to the level of floc structure (there are non-ignorable differencesbetween the fractal dimension of large floc and that of small ones). The3D fractaldimension of aerobic activated sludge was also negatively related to floc size,displaying a power function form.③Different from the previous studies, the population balance framework-basedmodel of flocculation dynamics were modified as follows:(i) according to theknowledge that the floc were form through the aggregation of basic particles, abinomial distribution function was proposed for daughter particle distribution due tobreakage and the parameter Cpof binomial distribution was used to characterise themain mechanism of breakage, erosion or splitting;(ii)“balance factor”, defined as theratio of breakage coefficient (A) to collision efficiency (α), was proposed for thecalibration of the model, which contributes to the reduce of the parameters needed tobe calibrated and the efficient simulationof the evolution of FSD. The simulationresults show that:(i) not only the FSD of activated sludge at stable states, but also theevolution of FSD during flocculation process could be simulated well;(ii) the balancedfactor displays power function relation to the average velocity gradient, increasing withthe increasing average velocity gradient,(iii) collision efficiency determines the speedof the flocculation process, and diaplays power function relation to average velocitygradient, decreasing with the increasing average velocity gradient;(vi) breakagecoefficient seems only related to the character of activated sludge not velocity gradient;(v) geometric grid factor (or number of classes) used in discretization equation of PBMhas not affect the calibrated value of balanced factor, and more accurate results isobtained with more geometric grids;(iv) compared with binary distribution, moreaccurate results was obtained with binomial distribution.④The models for collision efficiency and breakage frequency were investigatedand modified. It was concluded that:(i) compared with the rectilinear model, whichsupposed collision efficiency is constant, the curvilinear models treating collisionefficiency as the function of particle size ratio, proposed by Friedlander (1965) andPruppacher and Klett (1978), could improve simulation accuracy of FSD;(ii) themodified Pruppacher and Klett (1978) model could be used to simulate the evolution offloc average size during flocculation process;(iii) Kusters (1991) model for breakagefrequency could improve simulation accuracy of large-size floc distribution, andsimulate the evolution of floc average size well;(vi) By supposing that breakagefrequency has positive correlation with the maximum size of floc and increases with floc size until arriving at a maximum value, the modified Pandaya and Spielman (1983)model for breakage frequency was proposed and not only the evolution of FSD but alsothat of floc average size during flocculation process at different velocity gradient couldbe simulated well by this model.⑤The relationship between the parameters of flocculation dynamics and somekey factors (EPS content and Zeta potential) affecting the flocculation was investigatedusing the model of flocculation dynamics. It was found that:(i) The balanced factor(A/α) is positively correlated to EPS content or Zeta potential, decreasing with theincreasing EPS content or the decreasing Zeta potential (in absolute value); averagevelocity gradient with the later and EPS present significant linear inverse proportionrelationship; the smaller value of balanced factor, the more beneficial to flocculationand the lager floc average size obtained;(ii) the wide or narrower shape of FSD withdifferent EPS content or Zeta potential could be simulated very well using the binomialdistribution for daughter particle, and the parameter Cpof binomial distribution ispositively related to EPS content or Zeta potential, which indicates that breakagemechanism of different activated sludge may be not the same (activated sludge floccontaining high EPS content seems to represent splitting mechanism, whereas thatcontaining low EPS content seems to represent erosion mechanism; activated sludgefloc adding Al3+has strong stability and seems to represent erosion mechanism,whereas that adding Ca2+has weakly stability and seems to represent splittingmechanism).更多还原