efficiencies of all elements in the water network.
The basic elements of the Smart Water Grid is AMR (Automatic Meter Reading). AMR consists of a one-way real time network with field sensors, measurement and control devices that remotely and continuously monitor and diagnose problems in the water system. Smart water meters can monitor some key parameters such as flow, pressure, temperature, quality, consumption, and energy usage. The information gathered by these meters is wirelessly transmitted to a tower, which then transmits this information to a utility company.
A lot of research has arisen to leak detection utilizing AMR. According to statistics waterworks, 10% of the total supply occurred leaks annually. The amount of leakage has caused a loss of about 5,100 billion won in 2012. Because leak detection is a very important task.
A block system is defined as a discrete area of a water distribution network. It is usually created by closing boundary valves so that it remains flexible to changing demands. However, a block system can also be created by permanently disconnecting pipes to neighboring areas. Water flowing into and out of the block system is metered and flows are periodically analyzed in order to monitor the level of leakage.
Leak detection systems at block system detected leakage through the batch processing. Since the system detects leak through batch process, it is not possible to find a leak in real time. Leak detection method of the proposed study was to design a Statement for detecting leaks in real time using the CEP (Complex Event Processing).
In addition, if not collected flow data due to the failure of the meter or broken a network of remote meter reading system, there is a problem that as the estimated leak amount. This study raise accuracy in estimating the amount of leakage using the machine learning.