Forecasting simulations of indoor environment using data assimilation via an Ensemble Kalman Filter

• The EnKF is applied to forecast indoor environment.
• The contaminant concentration can be predicted without using local measurements.
• The optimum number of ensemble members is approximately 70–80.
• Both number of observations and time step of observations affect the EnKF performance.
• Predictability of the indoor model is significantly improved with EnKF implementation.

Data assimilation is widely used in weather forecasting and other complex forecasting problems such as hydrology, meteorology, and fire dynamics. Among various data assimilation methods, the Ensemble Kalman Filter (EnKF) is one of the best solutions to large-scale nonlinear problems while the computational cost is relatively less intense than other forecasting methods. In this paper, a new application of EnKF to forecast indoor contaminant concentrations is presented. The first part of the paper introduces the fundamental theories of data assimilation. The second part is a case study of forecasting the concentrations of a tracer gas in a multi-zone manufactured house by using a mass balance model with an EnKF. The benefits of EnKF and several important parameters for EnKF were discussed including numbers of ensemble members and observations, time step of observations, and forecasting lead time. The EnKF method presented is one of the first studies applied to the indoor environment field. It was shown that by using EnKF, the predictability of the simple indoor air model for the multi-zone space was improved significantly.