Skip to main content
ARS Home » Research » Publications at this Location » Publication #105437

Title: DIELECTRIC MOISTURE SENSOR FOR CEREAL GRAINS AND SOYBEANS

Author
item Lawrence, Kurt
item FUNK, DAVID - INSPECTION SYSTEMS ENGIN
item Windham, William

Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/26/2001
Publication Date: 6/1/2001
Citation: LAWRENCE, K.C., FUNK, D.B., WINDHAM, W.R. DIELECTRIC MOISTURE SENSOR FOR CEREAL GRAINS AND SOYBEANS. TRANSACTIONS OF THE AMERICAN SOCIETY OF AGRICULTURAL ENGINEERS. 2001.

Interpretive Summary: Grain moisture content (m.c.) is important in the sale and storage of cereal grains. Commercial electronic moisture meters are available to rapidly sense the m.c. of batch samples of grain. However, there is still a need to improve the accuracy of these electronic-type moisture meters. Since cereal grains are sold by weight, errors in m.c. measurements can result in substantial discrepancies between the true and purchase price of a lot of grain. Therefore, a rapid parallel-plate dielectric moisture sensor was used to predict the m. c. of wheat (hard red winter, hard red spring, soft red winter, hard white, soft white and durum), yellow-dent field corn, long- and medium-grain rough rice, and soybeans. The sensor uses complex electrical measurements and sample weight to predict the moisture content of batch samples and was tested at room temperature (~22 (C). The sensor was calibrated with samples of differing lots, grown throughout the U. S. in 1997, and independently verified with 1998 crop-year samples for all commodities except rice (1997 and 1998 crop years were combined for calibration). Generally, the sensor had lower errors when compared to commercial moisture meter readings for similar populations and moisture ranges. This research demonstrates the potential for improving the accuracy of grain m.c. measurements.

Technical Abstract: A shielded parallel-plate moisture sensor was designed, calibrated, and verified to sense the moisture content (mc) of wheat (hard red winter, hard red spring, soft red winter, hard white, soft white and durum), yellow-dent field corn, long- and medium-grain rough rice, and soybeans. The sensor uses complex reflection coefficient measurements and sample weight to predict the moisture content of static samples at room temperature (~22 C) with a sample volume of 850 cc. Multivariate analysis techniques were used to eliminate outliers. The sensor was calibrated with samples of differing lots, grown throughout the U. S. in 1997. The standard error of calibration and the coefficient of determination ranged from 0.37% mc and 0.986 for yellow-dent field corn to 0.09% mc and 0.986 for durum wheat, respectively. The sensor was independently verified with 1998 crop-year samples with bias and standard deviation values reported for all commodities except rice (1997 and 1998 crop years combined for calibration). Generally, the sensor had a similar bias and lower standard deviation compared to commercial moisture meter readings for similar populations and moisture ranges.