| 摘要: | 地表原水所含的物質種類相當繁多，主要藉由在水中加入混凝劑，使水中顆粒凝聚形成膠羽，以利後續處理去除。目前淨水廠主要以操作經驗或瓶杯實驗決定藥劑量，常有過量加藥的問題產生，使得藥劑量成本及污泥量增加。本研究團隊利用影像分析技術監測膠羽在混凝過程之長成變化，發現在混凝過程中顆粒的變化會反應於RGB值上，可利用RGB值之變化程度來判斷混凝效果及膠羽長成之好壞。本研究則沿用此套監測系統，應用於地表原水之混凝程序監測，在不同初始濁度及藥劑量下，探討光訊號RGB值的變化情況，以及利用影像軟體分析膠羽之大小及結構，加以了解膠羽長成情形與RGB值之間的關聯性，以建立此套系統更加完善的可行性。 研究結果顯示，快混階段之RGB值下降趨勢越明顯及標準偏差計算結果越大時，混凝效果越好，而慢混階段之RGB值跳動幅度越大及標準偏差計算結果越大時，長成膠羽越大。地表原水呈現土黃色，導致在水樣變化過程中R值的變化程度較B值及G值明顯，且透過B－R及G－R快混階段的下降斜率值與加藥量的關係，可判斷出有效加藥範圍，利用殘餘濁度驗證後也確定此判斷方法是可行的。另外在混凝過程中所得到之影像藉由影像分析軟體分析膠羽粒徑與結構後之結果與RGB之標準偏差值在混凝過程中變化趨勢也相符，表示利用RGB之標準偏差值大小判斷混凝及膠羽長成好壞也是可行的。此套系統監測確實可運用於實場原水之監測，除了能夠適時的反應出混凝過程中顆粒變化及膠羽長成的情形外，也可藉由光訊號RGB值快速得知混凝效果好壞，以達控制加藥之目的。Coagulation is an important process for the removal of suspended particles and dissolved substance from water. The dosage of coagulant is usually determined by jar test or the experience of operators, which usually leads to overdose of coagulant. Overdose of coagulant causes high cost for chemicals and sludge treatment. A novel monitoring technique, floc image colorimetric analysis (FICA), have been developed in our research groups to monitor performance of coagulation. The evolution of the red, green, and blue (RGB) values, obtained by analyzing the image of the suspension, can be used to determine performance of the coagulation. In this research, river water was used to examine whether this technique can be applied when there are colored particles and dissolved substances. Also, more image analyses were conducted to see if FICA can provide more information to determine the performance of coagulation of source water. The flocs size and structure were also analyzed to understand correlations between the flocs growth and the RGB values. The results showed that when the coagulation was effective, the RGB values decreased significantly and the standard deviation of the RGB values increased within rapid-mixing. Then, as the flocs became bigger during the slow-mixing, the flocs passed through the observation window and scattered great amount of light. As the consequence, the RGB values jumped up and the standard deviation of the RGB values increased. Because of the color of clay, the color of the raw water is light brown, in which the changes in the R-value are more obvious than G- and B-values. It is also found that the optimal dosage of coagulant occurred when the slopes of B－R or G－R are the same for the neighboring dosages, which indicate the structure of flocs are similar. The evolution of the standard deviations of RGB values agreed with evolution of floc size obtained from image analysis. In conclusion, by conducting coagulation of river water, it is sure that FICA, which can determine the optimal dosage during rapid- or slow-mixing, can be a useful tool for the monitoring of coagulation processes. Also flocs can be observed and further analyzed to examine the floc size and shapes. |
