淨水混凝之快混操作參數決定: 去穩機制與聚集動力解析

Abstract

摘 要 混凝/膠凝為淨水程序中廣泛運用於去除水中微粒的的物化程序，混凝可概分為三個階段，首先混凝劑加入水中首先進行一連串的化學反應之後形成有作用性的混凝作用基，接下來混凝劑作用基透過 ”快混”的操作 與水中穩定的微粒接觸，並進行膠體去穩，最後階段為已去穩的膠體微粒進行有效碰撞之後，形成粒徑較大的膠羽，形成膠羽可在後續沈澱與過濾處理單元之中被去除。快混單元的主要目的是在短時間內均勻地分佈混凝劑使其與水中顆粒碰觸且作用，進而達到顆粒去穩的作用，然而時至今日，此單元的操作仍是以經驗為主，沒有具體的準則可供操作或設計人員參考。 本研究主要在不同混凝的機制下，來評估快混操作因子對混凝的影響。首先利用不同劑量的硫酸鋁及聚合氯化鋁對3種低、中、 高濁度的水樣進行混凝，結果發現主要的混凝機制為電性中和以及沈澱絆除兩種機制。對於中、高濁度的原水混凝，此兩種機制皆可有效地去除顆粒，然而，當原水為低濁度時，沈澱絆除是唯一有效的混凝機制。 利用濁度量測以及顆粒計數對各種快混操作因子對顆粒去穩與移除的的影響進行評估。快混因子包含了混凝劑的注入溶液濃度、加藥點、攪拌器形式、攪拌強度與攪拌時間，為突顯快混因子的影響，此部分混凝試驗的水樣即採用高濁度原水，實驗結果顯示，注入溶液濃度、加藥點、攪拌器形式對於顆粒去除並無重大的影響，而攪拌強度以及攪拌時間則有明顯影響。若混凝以電性中和為主要作用機制，試驗結果顯示混凝劑的混合均勻與否為決定混凝效率的主要因素。若混凝以沈澱絆除為主，結果則顯示在快混激烈攪拌過久的情況下，顆粒去除效率反而降低。 接下來於兩種混凝機制下進行試驗，利用顆粒聚集動力的分析對快混攪拌強度與攪拌時間對混凝的影響進行探討。在電性中和的機制下，探討局部過量加藥的情況對水中顆粒聚集以及沈澱去除的影響，除了利用濁度量測觀察顆粒去除的情形，也利用顆粒計數器分析殘餘濁度中特定粒徑大小的顆粒。在顆粒聚集方面，則利用光纖膠羽偵測儀（PDA）觀察顆粒聚集的狀態，而針對混凝形成的膠羽則進行特性分析。結果顯示在足夠的快混攪拌強度操作下，顆粒去除效果好，上層液中微小顆粒數目較少，顆粒的聚集則呈現聚集體與聚集體的結合，但形成較鬆散的膠羽，沈降速度較慢；而在 不足夠的快混攪拌強度操作下，顆粒去除效果差，上層液中微小顆粒數目較多，顆粒的聚集則呈現顆粒與聚集體的結合，但形成較結實的膠羽，沈降速度極快。透過以上試驗的結果，我們建立一套模式來完整描述不同快混操作對後續顆粒聚集以及膠羽沈降的影響。 在沈澱絆除與電性中和的機制下，探討快混攪拌時間對原水中的顆粒聚集以及沈澱去除的影響，利用濁度量測觀察顆粒去除的情形。在顆粒聚集方面，除了利用光纖膠羽偵測儀觀察顆粒聚集的狀態，也利用粒徑分析儀對快混階段所形成的微膠羽粒徑進行分析。此外，我們也分析所形成膠羽的強度及再聚集能力。結果顯示在沈澱絆除的情況下，過長的激烈攪拌會導致殘餘濁度再上升，在電性中和的機制下則無此現象發生。這是因為沈澱絆除所形成膠羽的強度以及再聚集能力較差，長時間的快混攪拌會導致原本的膠羽破壞而形成較小且難以沈澱的膠羽。 最後，我們發展一套改良式PDA偵測方法以判定快混的操作情況，對應不同的混凝機制與原水濁度，PDA的輸出值與批次試驗所得的殘餘濁度呈現逆相關。比較此一方法、批次試驗與文獻中經驗公式所決定的最適快混時間，結果相當接近，顯示此一方法具有良好的可行性。

ABSTRACT Coagulation/flocculation is an essential physicochemical process for particle removal in water treatment. The process involves three continuous sequential steps. The first step involves the addition and the activation of the coagulant in water through a series of chemical reactions to form active coagulant species. Next, the coagulant species react with the suspended colloidal particles to destabilize the colloids with the aid of rapid-mixing. Finally, the destabilized colloidal particles collide form aggregates which can then be removed in the subsequent sedimentation and filtration. The purpose of rapid-mixing is to disperse the coagulants in the reactor within a short period of time. Since no concrete guideline concerning the engineering design of the operation of rapid-mixing in water treatment can be offered, the practice of rapid-mixing relies mostly on experiences. This study investigated the effects of rapid-mixing parameters on coagulations of various destabilization mechanisms. Three synthetic turbid water samples from low to high turbidity were prepared from Ball clay for this study. Alum and PACl were employed as coagulants, and their predominant coagulation mechanisms under different dosages were determined. Two mechanisms, the ACN and the sweep coagulation, dominated the destabilization of the colloid under such experimental conditions. Although both mechanisms were effective in removing particles from high and medium turbid waters, only sweep coagulation was effective for the coagulation of low turbid water. The effects of operational parameters of rapid-mixing on particle destabilization and removal were studied by monitoring the changes in turbidity and particle count. Parameters investigated were concentration of the coagulant working solution, dosing site, mixer type, mixing intensity and mixing time. And the coagulation was performed on high turbid water. The concentration of the working solution, dosing site and mixer type had no significant effect on particle removal. Mixing intensity and time, on the other hand, exhibited great impact on coagulation efficiency. For the ACN coagulation, the transportation of the coagulant species was important. For the sweep coagulation, prolonged mixing at high intensity had adverse effect on particle removal. These were further studied by the aggregation kinetics approach. For ACN coagulation, the effects of local overdosing on the subsequent flocculation and sedimentation were studied. The residual turbidity, particle count, dynamics of the particle aggregation, flocs characteristics such as size, density and fractal dimension were determined for this purpose. The results showed that insufficient rapid-mixing were unable to induce aggregation to effectively coagulate small particles. The flocculation induced was cluster-particle aggregation, which formed denser and more settleable flocs. Sufficient rapid-mixing induced cluster-cluster aggregation which effectively coagulated most particles. The flocs formed were looser and less settleable. An aggregation and sedimentation model was proposed to illustrate the significance of rapid-mixing intensity on coagulation. For both sweep and ACN coagulation, the effects of rapid-mixing time on the turbidity and the particle counts of the supernatant were investigated. The dynamics of the particle aggregation was evaluated by monitoring the PDA ratio output of the coagulated suspension. The size of the microflocs was examined with a particle sizer (PS 2400 PC). Floc strength and its reaggregation capacity under sweep and ACN coagulations were also studied. The results showed that in sweep coagulation longer mixing time resulted in higher residual turbidity, while in ACN coagulation the same or lower residual turbidity. The flocs formed from prolonged rapid-mixing under sweep coagulation contained low reaggregation capacity. The flocs were small and hard to settle after flocs breakup and reaggregation. Finally, an optical technique (modified PDA system) was explored to determine the performance of the rapid-mixing. For each coagulation mechanism and turbid condition, the aggregation index of PDA was measured and compared with the residual turbidity of the supernatant under various rapid-mixing times. The residual turbidity and the PDA ratio output demonstrated a significantly inverse relationship. The optimal rapid-mixing times were determined from the jar tests, the PDA method and an empirical model. Consistent results were discovered, which suggested the feasibility of the modified PDA system in optimizing the coagulation operation.