Inverse identification of a pollutant source in an apartment with multiple rooms by joint multi-zone model and CFD

In case a hazardous agent is accidentally released inside a building, both the pollutant source location and the temporal release rates must be determined in order to guide emergency actions. Inverse modeling based on either a multi-zone model or computational fluid dynamics (CFD) may help to find the pollutant source. This investigation proposed the inverse identification of both the release location and dynamic release rate profile of a pollutant source in an apartment with multiple rooms. In the first stage of the solution, the room containing the pollutant source was identified by the inverse multi-zone model; then, in the second stage, the occupant who released the pollutant was determined by the inverse CFD model. Both models solved for the possible pollutant release rate profiles with a regularized, inverted matrix at all candidate zones/positions based on information from a sensor. Next, the occurrence probability at each candidate zone/position was interpreted by the Bayesian probability model with input from another sensor. The results revealed that the two types of models could be used in conjunction to efficiently identify both the room in which the pollutant was released and the exact occupant who released it, as well as the dynamic release rate profile.