Study of flow prediction and optimal solution for water quality improvement and energy reduction = 수처리 수질 및 에너지 절감을 위한 최적운영기법 및 수요예측에 관한 연구

In the water treatment process, the main objective is to improve water quality and minimize costs. To achieve these, an integrated monitoring and control system has tried through flow prediction and pump and valve scheduling but frequently failed. This paper proposes an integrated solution for prediction, optimization and the controller can be solved by learning algorithm. Flow prediction has usually been studied for long-term estimation, which was improper to control water treatment facilities, and an hourly based algorithm to follow trends is proposed to track the steady increase of flow demand, such as population surge, without learning. Unlike electricity, water can be stored in huge tanks for more than a dozen hours, which can be used for saving energy and increasing water quality. One of the general optimization problems is load shift operation, whose purpose is to charge various electric fees according to the hour and season. The other is pump on/o minimization to improve water quality. If inputted water to water treatment plants (hereafter WTP) varies, than output turbidity and particles are increasing, which could possibly be supplied to citizens. The proposed on/o minimization is expected to prevent those particles from leaking and to secure public health. To solve these optimization problem, a valve controller should be made beforehand to control the scattered tanks like as controlling one tank, which can make maximum usage of tank. The proposed controller will solve the target levels in WTPs and tanks simultaneously. The proposed modeling and controller have been tested to check its availability for real water treatment process.
Lee, Ju-Jangresearcher이주장researcher
Issue Date
455136/325007  / 020084104

학위논문(석사) - 한국과학기술원 : 로봇공학학제전공, 2010.08, [ viii, 60 p. ]


Load Shift; Flow Prediction; On/Off Minimization; On/Off 최소운전; 전력요금최소운전; 수요예측

Appears in Collection
Files in This Item
There are no files associated with this item.
  • Hit : 46
  • Download : 0
  • Cited 0 times in thomson ci


  • mendeley


rss_1.0 rss_2.0 atom_1.0