Water Quality Monitoring with Regression Based PPM Sensor for Controlling Hydroponic Dissolved Nutrient

Abstract

Hydroponic cultivation requires rigorous monitoring and control of several parameters, such as turbidity, electric conductivity, acidity (pH), dissolved oxygen and nutrient, which usually be measured once a day manually. Therefore, automation in hydroponic cultivation requires those water quality information as the controlled variable. The dissolved nutrient is especially important because it significantly affects the hydroponic plant growth. Acquiring the dissolved nutrient can be done by using a PPM (parts per million) sensor, but most of the time the sensor needs further processing to obtain the desired measurement. This study presents a reading correction of a PPM sensor based on a regression method so the desired measurement can be done. Sample water with different PPM, such 309 PPM, 290 PPM, 762 PPM, 1910 PPM and 2420 PPM are measured first using a standard PPM meter. Then, the sample PPM is measured by using the PPM sensor. The study also investigates the best regression method to map the PPM sensor measurement to the standard PPM meter measurement by comparing several line equations, such as linear, exponential, polynomial and logarithmic. The function coefficient and bias is chosen by using least square methods. After comparing, the result shows that the polynomial function provides the best reading correction with average error of 76 PPM. The error is especially few when measuring the higher PPM (more than 500 PPM), which is suitable with hydroponic cultivation. Therefore, the PPM sensor with the polynomial function shown in this study can be used to measure the dissolve nutrient accurately in the automation of hydroponic activity compare to other line equations. This study is limited to small sample sizes to prove the concept. The generalization can also be considered in the future study.