Development of Equilibrium Moisture Relationships for Storage Moisture Monitoring of Corn

Abstract

Commercial systems are currently available to measure grain moisture during storage using relative humidity (RH) and temperature (T) sensors and equilibrium moisture (EMC) models. However, the variability of the EMC relationships between grain lots necessitates that a good model be selected or developed for each specific grain. The objective of this research was to develop modified Chung-Pfost equilibrium moisture models for seven corn samples, examine their prediction accuracy, and evaluate simplification of the experimental procedures used to develop these models. The models that were developed used corn conditioned over a broad range of five moisture levels (5-point models). Models were also developed using a reduced number of levels with a narrower moisture range; models with 4 and 2 moisture levels (4- and 2-point models, respectively). The 4-point model used all data except the highest moisture level while the 2-point model used the extreme moisture levels of the 4-point model. The 5-point models had the highest standard error of estimates (SEE) averaging 0.90 and 0.82 for adsorption and desorption, respectively. The 2-point model was used to predict all of the moisture levels of the 4-point data; the standard errors of prediction (SEP) for these predictions averaged 0.41 and 0.39 for adsorption and desorption, respectively, and are only marginally higher than 4-point models. These results show that a 2-point model can be used for accurate moisture measurement across a range common for stored corn, while reducing the amount of work required for model development. The moisture levels suitable for a 2-point model should be limited to upper and lower moisture contents corresponding to 85% to <90% relative humidity and 35% to 50% relative humidity, respectively.