Probability-based inverse modeling algorithm for indoor pollutant source tracking

Building indoor environment quality (IEQ) has received growing attentions lately because of the extended time people spend indoors and the increasing reports of health problems related to poor indoor environments. Recent alarms to potential terrorist attacks with airborne chemical and biological agents (CBA) have further highlighted the research needs on building vulnerability and protection. To maintain a healthful and safe indoor environment, it is crucial to identifying contaminant source locations, strengths and release histories. Accurate and prompt identification of contaminant sources ensures that the contaminant sources can be quickly removed and contaminated spaces can be effectively isolated and cleaned. This paper proposed a probability concept based inverse modeling method − the adjoint probability method that can identify potential indoor pollutant sources with limited pollutant sensor ouputs. The paper introduces the principles of the method and presents the corresponding adjoint equations for computational fluid dynamics (CFD) model. A CFD based adjoint probability inverse modeling algorithm and program have been developed. By using an office building and an aircraft cabin as examples, the study demonstrates the application of the program for identifying indoor airborne pollutant source characteristics (location and release time) with few sensor measurement outputs. The research verifies the feasibility and accuracy of the adjoint probability method for indoor pollutant tracking. The paper indicates the further research directions with the goal of developing an intelligent and integrated building environment management system that can promptly respond to building pollution conditions with  effective detection, analysis and control.