Water is one of the indispensable things to human and we cannot survive without it. Still, many people have died from water-related diseases, where tap water is not supplied. Even the citizens with tap water wants cleaner and healthier water, which is thought to be endless desire. For this purpose, we propose a new control method using machine learning algorithms to cope with limitation of the existing manual operation or general feedback control methods.
Coagulants are injected in order to remove a variety of organic substances in the influent of a water purification plant. The proper dosage rate of coagulants can be determined by measuring sedimentation turbidity four to seven hours later, which make real-time feedback control impossible. In addition, manual operation in accordance with the Jar-Test carried out in the laboratory may cause experimental and human errors due to the changes of organic characteristics and water quality. The proposed intelligent control algorithm to decide optimal chemical dosage rate is expected to operate in real-time and reduce cost up to 25%.
In the pre-chlorination process, the resulting water quality depends largely on operator’s experience, because sedimentation basin is open to space and affected by climate and sunshine. From this background, a fuzzy system is proposed for pre-chlorine modeling to reduce the carcinogenic substance and decide the optimal chlorine injection rate, which is affected by chlorine evaporation rate in sedimentation basin depending on detention time, weather and water quality. The additional chlorine meter is installed in the inlet part of sedimentation to reduce the feedback time and implement cascade control, which leads to maintaining the residual chlorine concentration decided by fuzzy rules. The proposed fuzzy system helps to take a preemptive action about long time delay, the characteristics of the disinfection process, and reduce the variation of residual chlorine rate by 4.6 times and the chlorine consumption.
In the post-chlorination process, it is very important to maintain a constant chlorine concentration, which is the final step in the water treatment process before servicing water to citizens. Even though a flow meter between the filtration basin and clear well must be installed for the post-chlorination process, it cannot be installed due to bad conditions. In that case, a raw water flow meter has been used as an alternative and has led to dosage errors due to detention time. Therefore, the influent flow to the clear well is proposed to be estimated by a time series neural network for the plant without a measurement value and the modeling and controller is also proposed by the neuro-fuzzy algorithm to analyze the chlorine concentration change in the well. The proposed algorithm leads to the reduction of input and output chlorination rate standard deviation of 1.75 times and 1.96 times respectively when it was applied to an operating WTP. As a result, hygienically safer drinking water is supplied with preemptive response for the time delay and inherent characteristics of the disinfection process.
In this paper, the improvement effect of the coagulation and disinfection process have been confirmed by using the proposed algorithm in the main process of water treatment. In the future, those artificial intelligence algorithms must be examined continuously in other processes to supply water more hygienically, efficiently and economically.