Improved accuracy of radio frequency (RF) heating simulations using 3D scanning techniques for irregular-shape food

Many food products have irregular shapes and their geometries are difficult to create for computer simulation purposes. Three-dimensional scanning techniques could be utilized in the food industry to overcome this issue. In this study, potatoes, carrots and ginger roots were selected as representative vegetables and subjected to RF heating to obtain their heating pattern. At the meantime, the processes were simulated with the COMSOL Multiphysics (R) software package with both an imported 3D scanned & reconstructed geometry and a similar regular-shaped geometry, which was close to the original shape of the various food samples for comparison. Results showed that computer simulation with geometry obtained by using 3D scanning could accurately reproduce the experimental results for the hot/cold spot locations and temperature values. For the potato samples, the maximum temperature difference between experiment and real-shape simulated temperature on the potato surface was 1.8 degrees C, and the temperature difference from regular-shaped simulated geometry was 4.9 degrees C after 40 min of RF heating. The carrot and Ginger samples showed similar results. In general, the model with 3D scanned geometry provided an accurate prediction with an accuracy > 95%. Thus, using 3D scanning created geometry has potential advantages in many food processing simulations due to its improved prediction accuracy. Industrial relevance: The 3D scanning technique is convenient to use for complicated irregular-shape geometry reconstructions. Importing the 3D scanned geometry into a built-up RF heating model could accurately predict the temperature distribution of irregular-shaped vegetables during RF heating. This method could be adopted by the food processing industry for a virtual scaled-up continuous food processing design, especially for large pieces of materials with complex geometries.