PENGGUNAAN NEURO FUZZY PADA SISTEM MONITORING KETINGGIAN AIR SUNGAI

Abstract

Flood is a natural phenomenon that usually occurs in an area that is often flowed by rivers. The impact of flooding can be reduced if the community is better prepared to face the coming flood. One of the solution is to create a flood early warning system. In this study, the concept of early warning system is a river water level sensor that consist of 2 conductor plates that will produce a capacitance value when drowned into the water with a certain depth. These water level measurement values will use 3 different methods. Furthermore, those three methods will be compared to determine the best method that produces a small measurement error. The three methods are linear regression, polynomial regression, and neuro fuzzy. The results show that the measurement error which generated by linear regression was 25.45%, the measurement error that generated by polynomial regression was 13.61%, and the measurement error that generated by neuro fuzzy was 3.76%. According to these error values, the study finds that neuro fuzzy is the best method for water level measurement.