Joint state estimation of indoor thermal dynamics with unknown inputs using augmented fading memory Kalman filter

Abstract

An intelligent and efficient utilization of a heating, ventilation, and air conditioning system can be instrumental to reduce the building energy consumption, which in turn, is expected to reduce the green-house gases. The energy profiling requires modelling and estimation of the building environment with uncertainties. This paper proposes a strategy to estimate indoor thermal dynamics at multiple walls using a forgetting factor-based fading memory Kalman filter (FMKF) in presence of unknown inputs. This work also proposes a joint state estimation scheme based on FMKF which considers augmentation of the unknown heating energy inputs along with the thermal parameters of the thermodynamic model developed for indoor environment. The contribution of unknown inputs in the process of state estimation have been studied in the context of measuring node distribution. The proposed scheme has been implemented for multiple real-life thermal scenarios and results outperformed the conventional Kalman filter-based estimation scheme.