| Summary: | 碩士 === 國立中央大學 === 環境工程研究所 === 99 === Coagulation is an important solid-liquid separation process in potable water/wastewater treatment. Among all the control factors for coagulation, the dosage of coagulant is the most difficult to control. The dosage of coagulant is usually determined by jar test or the experience of operators, which usually leads to an overdose of coagulant. Overdose of coagulant causes high cost for chemicals and for on sludge treatment. So far, in our laboratory, a monitoring system for coagulation process, floc image colorimetric analysis (FICA), to resolve the problems mentioned above has been developed; however, the responses of the monitoring system for various coagulation conditions are not clear and the applications of the system in the control of coagulant dosage have not been clarified yet.
In this research, modeled turbid water is made by kaolin particles. The responses of the FICA to various coagulation conditions, such as initial turbidity, solution pH, and coagulant dosage were investigated. The evolution of RGB values of the suspension images were analyzed.
During coagulation, the solution images varied as the particles grew; as the consequence, the RGB values of the images changed correspondingly.. When the coagulation was effective, the RGB values decreased significantly within rapid-mixing. Then, as the flocs became bigger, the RGB values jumped up as the flocs passing through the observation window and dropped down due to the clear solution during the slow-mixing process. It was found that the slope of the RGB values during rapid-mixing could be used to monitor the coagulation efficiency. As the coagulation efficiency increased, the solution becomes clearer and, thus, the RGB dropped significantly and the slope of RGB values increased. .It is also found that the optimal dosage of coagulant occurred when the slopes of RGB values first hit the bottom. In other words, the optimal coagulation condition could be determined right after the rapid-mixing from jar test, which is much faster than the traditional method. In addition, the standard deviation of the RGB values increased with increasing particle size due to brighter images of flocs.
|
|---|
