Performance of sensors for quality analysis of irrigation water

Abstract

Monitoring the quality of irrigation water can help in the maintenance of filters and irrigation systems, avoiding clogs and uniformity problems. The objective of this work was, thus, to evaluate the performance of sensor modules for monitoring irrigation water quality variables. For that, three sensors were evaluated, and their performance was rated from the adjustment of calibration equations, obtained through linear regression analysis (yi = b0 + b1xi + εi), using the ordinary least squares method (OLS) to estimate its parameters (β0 and β1). The first sensor evaluated was the Ph4502c for pH measurement. Direct methodology was used, using standard pH solutions (1.79; 4.5; 6.88; 12.13; and 13.99) and an electrode type BNC probe. The second evaluated sensor was turbidity model TSW30. To evaluate the total dissolved solids (TDS) sensor, the direct method was applied, using solutions with electrical conductivity of 0.50, 1.0, and 2.0 dS m-1. To investigate the assumptions of independence, homoscedasticity, and normality of the residuals of the linear regression models, the Durbin-Watson, Breusch-Pagan, and Kolmogorov-Smirnov tests were respectively used. In the evaluation of the statistical performance, the indicators of the root-mean-square error, coefficient of determination, correlation coefficient, confidence index, and index of agreement were adopted. The ordinary least squares method did not produce the best unbiased linear estimators for the calibration equations of the pH, turbidity, and TDS sensors, due to the violation of the regression assumptions. The adjustments showed good accuracy for water quality assessment, according to high performance statistics and models classified as ‘Excellent’.