PROBABILISTIC SIMULATION OF BATCH TRAY DRYING USING MARKOV CHAINS AND THE MONTE CARLO TECHNIQUE

The objective in food dehydration is to dry the product to a specified uniform moisture content. In practice, a distribution in final moisture content is unavoidable, arising from the intrinsic variability of food properties and the probabilistic nature of the food drying process. A methodology is outlined whereby the inherent stochastic dynamics of batch tray dehydration in cabinet or tunnel dryers can be modeled efficiently using Markov chains. By simulating the drying process with the model, the progressive transformation of the moisture distribution of the trays throughout the drying cycle can be quantified. Output from the model is validated with experimental drying results to test its accuracy. In addition the output of the Markov approach is compared with that of an equivalent Monte Carlo simulation model to provide inter-model comparisons. Markov analysis can predict the variation in final moisture content and thus has the potential to help maximize the value of the dried product.