Multivariate Standard Addition Cobalt Electrochemistry Data Fusion for Determining Phosphate Concentration in Hydroponic Solution

Abstract

Phosphate is one of the major elements affecting agricultural production. The accurate determination of phosphate concentration essential for plant growth, especially in a hydroponics system, allows regulating the balanced and suitable range set of nutrients to plants efficiently. This study proposed a data fusion model based on 70 samples for calibration and 30 samples for predicting concentrations of phosphate in an eggplant nutrient solution. Three multivariate analysis methods i.e. partial least squares model (PLS), Gaussian process regression (GPR), and artificial neural network (ANN) were studied and compared for their performance efficiencies. The results showed that combining the multivariate standard addition method (MSAM) in acquiring data from cobalt electrodes and ANN data fusion model came up with satisfactory outcomes. Both the method provided good performance with R2 values of 0.98 and 0.96, and the root mean square error (RMSE) of 50 and 66 mg. L-1 respectively in calibration and evaluation tests. These values were much higher than those of conventional processing techniques. Moreover, the normal direct calibration method in acquisition signal from cobalt electrodes was also applied, which provided R2 values of 0.7 to 0.8. These high values are sufficient for development to measure phosphate concentration in hydroponic solutions.

Keywords

• Phosphate sensing
• multi-sensor data fusion
• multivariate standard addition method (MSAM)
• partial least squares model (PLS)
• Gaussian process regression (GPR)
• neural network-ANN