An Automated System for Determination of Hydration Kinetics of Dry Legumes: Improvement, verification, and modeling of commercial hydration of navy beans (Phaseolus vulgaris)

The species of Phaseolus vulgaris, which contains many commonly consumed cultivars of beans, can be found in grocery stores across the United States as a canned product. Beans are often hydrated to reach equilibrium moisture to ensure the safety and quality of canned bean products. The hydration kinetics of specific species or cultivars of beans must be analyzed to optimize soak time. Existing methods for assessing hydration kinetics are both time-consuming and labor-intensive. An automated system, called the Bean Volume Automated Tester (BVAT), was developed to monitor changes in the weight and volume of legumes as they absorb water. This system was tested against the AACC method 44-17.01 for moisture determination to calibrate the BVAT to provide percent moisture content readings consistent with the standard method (Cereals and Grains Association, 1999c) The two moisture determination methods had a linear relationship (y =1.09x-6.04). Using the Peleg model, the BVAT as able to distinguish hydration patterns at different temperatures (P < 0.0001). The BVAT also showed significant differences in hydration patterns across three varieties of black bean using Peleg model parameters (P=0.0013 [k₁], P=0.0347 [k₂]).

Hydration patterns from a commercial hydration process (soak tank) were tested for uniformity. Three depths within a soak tank were sampled and their moisture contents over time were fitted to the Weibull model. The beans at the top (3 ft depth) of the tank hydrated significantly more slowly than the middle (6 ft) and bottom (9 ft) levels (P = 0.0133). The three depths showed no significant difference between the Weibull model parameter related to the time it takes to reach 63% of equilibrium hydration. The depths did not show a significant difference in initial hydration rate (alpha = 0.05) after adjusting the soaking time for the time it takes to fill the soak tank. The BVAT accurately predicted hydration at the top of the soak tank (RMSE = 0.370, R² = 0.992). When soak time was adjusted, the BVAT also accurately predicted hydration at the middle of the soak tank (RMSE = 1.0, R² = 0.980). This automated system will allow bean manufacturers to better understand the hydration kinetics of beans before canning. Understanding hydration ensures a uniform product and optimizes processing and quality while maintaining microbial safety and cutting costs.