The Comparison of One-Variable and Two-Variable Polynomial Regression Models to Measure the Cellular Concrete Moisture Using the Time Domain Reflectometry Method

In the paper there are presented models for moisture assessment applying the two reflectometric sensors in the cellular concrete samples. The readouts express the dependence between the cellular concrete moisture, measured in gravimetric way and the apparent permittivity values achieved by the Time Domain Reflectometry method and two surface sensors. According to observed relationships, the two types of calibration models were derived – the first model is a traditional one-variable model covering only time of signal propagation and the second one two-variable model which together with signal propagation time takes into account signal attenuation. The aim of this paper is to verify the efficiency of multiple regression to improve the accuracy of moisture estimation using the TDR technique. The applied models that consider amplitude attenuation are used for this type of analysis for the first time. With the conducted research and analyses, it was shown that the measurement quality of the method could be improved by obtaining more favorable values of the determination coefficient, Residual Standard Error, Root Mean Squared Error. Also the correlation analysis shows a better fit of two-variable models than one-variable to the obtained data.