Optimizing the Measurement of Metal Content in Water Using the Kalman Filter Method

Abstract

Research on detection and screening of metal in the water using Atmega 8535 Microcontroller focused on designing a device to detect metal levels in the water and it could conduct metal screening in the water. There were 3 samples such as well water, gallon water, drinking water, and RO water as a neutralizer. It was obtained fluctuating or unstable result. It was because several factors like lack of sensor probe number used, sensor probe materials which were too reactive to the water, and unstable change of output value because of previous value saving. Because of these problems, this study was conducted an optimization of metal content measurements in water with TDS sensor with Arduino Uno base. The optimization was done by using the Kalman Filter method. There were refill water samples, tap water /drinking water, and well water which were tested 30 times each. The error percentage results without the method were 29.74%, 9.53%, 9.02%, while the results error percentage used the Kalman Filter method were 29.34%, 9.39%, 8.94%. In this test, the Kalman Filter method could reduce the error percentage in a row by 0.4%, 0.14% and 0.08%. In the test comparing eight samples with the TDS meter, the researcher got the average error percentage for testing without the method of 7.881%, and the average error percentage for testing with the Kalman Filter method of 6.705%. In this case, the Kalman Filter method can reduce the error percentage by 1.176%. From these data, it could be concluded that the Kalman Filter method could reduce the error percentage so that the reading of the measurement results would be more stable.