Uncertainty estimation for potentiometric sensor-based measurements

The authors have been involved in uncertainty analyses of multiple-sensor measurement procedures and Data Fusion (DF) algorithms under development in the 6TH FP WARMER (Water Risk Management in EuRope) project [1]. The main goal of this uncertainty study was evaluation of measurement procedure-dependent factors that determine basic uncertainty characteristics, i.e. accuracy and precision. Several uncertainty sources were taken into account, most important being: read-out/sensor modeling inaccuracy and dosing imprecision. Results of uncertainty analyses will be used to optimize measurement procedures for potentiometric sensor-based measurement heads and to perform rational cost/accuracy trade-off of in-situ measurement probe components. The paper presents the most important results of the uncertainty study. Accuracy and precision of different variants of in-situ measurement system were estimated with Monte Carlo method, using realistic estimates of uncertainty sources. First type of results concerns analysis of measurement uncertainty sensitivity to different sources of imprecision. Next, dependence of measurement uncertainty upon measurement scenario is shown. Single- and double-sensor measurements with different selectivity coefficients are compared. Advantages and limitations of multiple sensor Data Fusion are discussed and recommendations formulated.