Mathematical Models for Predicting Moisture Content of Aggregate Based on Electrical Resistance Sensors

Abstract

This technical paper aimed to develop mathematical models for predicting moisture contents using electrical resistance sensors for both coarse and fine aggregates. The mathematical models were in form of a polynomial function for multiple variables based on regression for each aggregate source. The unknown coefficients were determined from relationships between moisture contents and the values of electrical resistance measured by the sensors. The aggregate used in this study consisted of one source of coarse aggregate and two sources of fine aggregate from the western-central part of Thailand. The fineness modulus of aggregates varied within the specified range. Moisture contents varied from dry to damp states. The results from the study showed that the mathematical models in form of polynomials for multiple viables based on regression were of high accuracy with the adjusted coefficients of determination at least 0.99. When these mathematical models were verified and validated the predicted values of moisture content compared with those obtained from the oven-dried method, the accuracy was at least 95.6 percent. Applying these mathematical models for predicting moisture contents in coarse and fine aggregates in the concrete batching plant would result in more uniform quality of concrete and indirect reduction in material cost.