Investigation On Transmission And Dispersion Of Particles In Limited Space With Markov Chain Based Probabilistic Model

Indoor suspended particles are crutial factors influcing Indoor Air Quality(IAQ).And the effects can be classified into two main aspects: The first one lies in the impacts on occupant's health.The other one lies in the impacts on industrial manufacturing processes requiring the clean space.It is therefore crucial to acquire information on transmission and dispersion of indoor suspended particles for indoor environment control.One of the effective ways of obtaining abovementioned information is to develop predictive models.Developing models that can rapidly acquire real-time or even faster-than-real-time information on indoor suspended particles is currently one of the popular research topics in the indoor environment field.The present paper,which mainly investigated the prediction of the transmission and dispersion of indoor suspended particles by using the Markov chain based model,consists of five chapters.The primary contents of each chapter are briefly introduced below:The first chapter is the introduction chapter,which has summarized the up-to-date methods and research status all over the world.The analyzed results reveal that the Multi-node model cannot provide sufficient information on contaminant concentrations but with the fastest computing speed and a simple model configuration.The Computational Fluid Dynamics(CFD)method can provide abundant information on contaminant concentrations but sometimes with huge computational costs.Both Zonal model and the Markov chain model are considered as intermediate models between the Multi-node model and the CFD model regarding computational efficiency and accuracy.These models currently can simply predict convection and diffusion induced particle transmission instead of complicated behaviors of particles including infiltration,deposition and resuspension.Based on the shortcomings of these predictive models mentioned above,this paper tries to propose a set of Markov chain based models to rapidly predict the transmission and dispersion of indoor particles.Thus,a new computing system different from the conventional predictive models is constructed.The second chapter of this paper constructed the new storage matrices for the flow field data provided by the CFD softwares.Based on the matrices,we developed the CFD combined Markov chain model to rapidly predict the transmission and dispersion of indoor airborne contaminants released by constant/pulsed contaminant sources under steady-state/transient flow conditions.Furthermore,we introduce the basic CFD combined Markov chain model into multi-zone spaces,and the grid-merging operation is developed to merge grids and improve computational efficiency in the regions with less concerns.In addition,the effects of the time step size and grid merging rate on the accuracy of the proposed model are analyzed.Comparasion of computational efficiency and model accuracy between the proposed Markov chain model and traditional Eulerian/Lagrangian model is conducted.The implementation of the grid-merging operation and the multi-zone Markov chain model enables the capability of functioning as either the CFD model or the Multi-node model.Due to the wide range of distributions of indoor particle diameters as well as complicated thermal indoor environment,it is reasonable to observe deposition under specific conditions of particles with large diameters.The primary task of the third chapter is to develop the Markov chain based predictive model for particle deposition,which is constructed by firstly adding some extra supplementary cells adjacent to the boundary cells.Then the transfer probabilities caused by gravity or thermophoresis between the supplementary cells and boundary cells are calculated.The supplementary cells can also be used to receive deposited particles.The present model can be adopted to simulate gravity as well as thermophoresis induced deposition in non-isothermal environment of indoor suspended particles.Some parameters such as dimensionless deposition velocity,deposition/collect efficiency and escape efficiency are used to evaluate the predicted results.Meanwhile,the effects of temperature gradient and particle diameter on the accuracy of the proposed model are investigated.Particles deposited on indoor surfaces can get re-suspended due to various disturbances,which include aerodynamic as well as mechanical disturbances.The main purpose of the fourth chapter is to develop a Markov chain based predictive model for particle resuspension.The proposed resuspension model is in reality a combination of the basic Markov chain model and the turbulent burst theory,which is implemented to provide the methods of calculating the probabilities of particle transfer between the supplementary cells and the boundary cells.Then the original State Transfer Matrix can be extended and modified for further calculation.Moreover,the effects of the fraction of particles re-suspended under a single turbulent burst,particle diameter and inlet flow velocity on the accuracy of the proposed model are also investigated.To further extend the applied range of the previously proposed Markov chain model from steady-state conditions to transient conditions,the fifth chapter aims to develop a Markov chain based model to predict the transmission and dispersion of contaminants released by sources with dynamic strength/dynamic location and under the condition of dynamic flow distribution.Markov chain based predictive model for sources with dynamic strength is realized by changing the expressions of the state vectors within one single step according to the function of release strength.Markov chain based predictive model for sources with dynamic location is realized by changing the expressions of the state vectors within one single step according to the function of moving path.Dynamic flow distribution in the present study is the process of consecutively changed flow patterns.The Markov chain based predictive model under the condition of dynamic flow distribution is realized by multiplying all State Transfer Matrices regarding those consecutively changed flow patterns in chronological order during the state transfer process.The proposed model is proved to be valid through the verification of the experiment data from literatures in predicting the transmission and dispersion of contaminants released by dynamic sources and under the condition of dynamic flow distribution.The effects n,flow pattern and dynamic source moving path on the predicted results of the proposed model are analyzed.The Markov chain based predictive model for the transmission and dispersion of indoor particles investigated in this paper comprising the predictive models for airflow induced particle(passive)transmission,gravity/thermophoresis induced particle deposition,particle resuspension and dynamic source and flow field conditions.The proposed models in this paper combined the advantages of the CFD and Multi-node methods with respect to computational accuracy and efficiency,respectively.Thus,the proposed model provides a completely new method for real-time or even faster-than-real-time predictions on the transmission and dispersion of indoor particulate matters.