Development of a low-cost volumetric moisture content sensor at 76 GHz for industrial drying applications using effective medium modeling

Current methods of production line compatible moisture content detection are too costly to allow for full process monitoring. Other key drawbacks of current methods include sensitivity to changes in sample color, poor sample penetration, and limited moisture content measurement range. Sensors operating in the THz range could provide a solution to these problems and by carefully picking the frequency, a low-cost sensor could be developed. This work covers the development of a low-cost volumetric moisture content sensor based on commercial off-the-shelf hardware, a TI IWR1642 76 GHz automotive radar module, coupled with a sample specific calibration based on effective medium theory. The resulting sensor is orders of magnitude less expensive than competing sensors while providing measurements of volumetric moisture content accurate to within 5% dry-basis moisture content. This work includes a study of two effective medium models, the Bruggeman method and the Looyenga method, and their applicability to the moisture sensing problem as well as a detailed description of the methodology used to build these models and assess their accuracy. These effective medium models utilize known microscopic material properties to describe the macroscopic properties of mixtures. The application space that this study explored was industrial pulp and paper production where precise measurement and fast measurement of volumetric water content is crucial for ensuring product quality and consistency. Using electromagnetic radiation in the THz band to measure the water content of paper is not completely novel but previous approaches have made use of empirical calibration models which require significant time to acquire and can be sensitive to changes in calibration and application environment.