Inverse Modeling Of Indoor Airborne Contaminant Source Location With Adjoint Probability-based Method Under Unsteady Airflow Field

Accurate and prompt identification of airborne pollutant source location in indoor environment is of vital importance to pollutant isolation and removal once there is accidental release of harmful agents.Successful inverse modelling to identify the location of the airborne pollutant by limited sensor readings relies on accurate flow field information as an input.For steady-state indoor inflow,such problem has been intensively investigated.However,in many real case scenarios air velocity direction and magnitude varies with time,which poses a great challenge to the original method of identifying airborne contaminant source location.This paper mainly focuses on the identification of the pollutant source location under dynamic airflow.The mathematical model and process is investigated and presented.In this study,an adjoint probability-based inverse tracking method was employed to identify the pollutant source location.On the basis of the original adjoint probability method under steady-state airflow,it is extended for dynamic airflow field.The applicability of the newly derived method is investigated by three case studies: a two-dimensional office case and subway station case and three-dimensional cabin environment case.The capability of the newly derived method is verified for air pollutant source location identification under unsteady indoor airflow through the two-dimensional office case.By comparing the performance of newly developed mathematical model under the fine grid and the coarse grid,it is found that this mathematical model doesn't need the very detailed thermo-flow fields.As long as the flow pattern is consistent with reality,the inverse algorithm will give acceptable source location prediction.Using coarse grid can greatly improve the computational speed of CFD,which lay a solid foundation for real-time application of identifying the location of pollutant source.In order to highlight the necessity of developing the mathematical model under dynamic airflow,this study selected two-dimensional subway station cases and three-dimensional cabin environment with strong unsteady characteristics of airflow to demonstrate and verify the if the algorithm is applicable.It is found that previously developed adjoint method for steady-state airflow field failed to identify indoor airborne contaminant location under such unsteady airflow field.The new algorithm consideringthe airflow field variation is necessary for airborne pollutant source location identification.In order to validate application of the adjoint method using physically measured contaminant concentration data,experiment was conducted in the aircraft cabin chamber.Obtaining the actual flow field by measurement is quite difficult and also very expensive.In such circumstance,we used an alternative option by performing a contaminant concentration comparison to make sure that air flow field by numerical simulation is consistent with the experiment.The experiment case study verified the accuracy of method for predicting the position of indoor pollutant source under dynamic airflow scenario and the air flow field by numerical simulation can be used as an alternative to the experimentally measured data in this inverse calculation process.