Predicting transient particle transport in periodic ventilation using Markov chain model with pre-stored transition probabilities

Abstract

Periodic ventilation has been proposed for effectively removal of air contaminants in indoor environments. To facilitate the design of periodic ventilation, it is essential to predict the transient particle transport. Compared with traditional models, the Markov chain model is faster as it does not require iteration in each time step. However, for unsteady airflow, since the transient transition probabilities need to be updated in each time step, the potential of the Markov chain model in terms of computing speed has been significantly limited. This study developed an improved Markov chain model with pre-stored transition probabilities to reduce the computing cost for predicting transient particle transport under periodic ventilation. The proposed model was validated with experimental data from chamber tests. The validated model was then used to investigate the particle removal performance of periodic ventilation under different conditions. The results show that the periodic ventilation removed particles more effectively than steady ventilation, especially under a high ventilation rate and when the particle source was located in the recirculation zone. Furthermore, the supply-air period and mode had little impact on particle removal. Finally, the computing costs for the two Markov chain models were compared. For the case with six supply-air periods investigated in this study, the pre-storage Markov chain model was 4.8 times faster than the existing Markov chain model. Furthermore, the computing cost ratio of the pre-storage model to the existing model decreased with the product of period and source number.