Drying lumber in kilns: A methodology for stochastic analysis using Markov modelling

The value of dried lumber to the kiln operator is to a large extent dependent on the final moisture content distribution of the batch. This variation in moisture content arises from the intrinsic heterogeneity of wood and the stochastic nature of the timber drying process. Theoretical analyses of wood drying have hitherto relied upon deterministic models and have limited practical applicability. A methodology is proposed herein whereby the inherent probabilistic dynamics of batch drying can be modelled efficiently using Markov chains. The discretization of the variables of moisture and lime that is necessary for this approach is discussed. By simulating the drying process with the model, the progressive transformation of the batch moisture distribution throughout the drying cycle can be studied. Output from the model is compared with experimental drying results to test its validity. There are direct model applications in improving kiln control algorithms and optimising pre-drying sorting and conditioning strategies.