An Intelligent System for Actuating Windows of Naturally Ventilated Residential Houses
Authors: Pokhrel, Manoj. K.; Anderson, Timothy ; Currie, Jonathan; Lie, Tek T.
In New Zealand’s (NZ) mild climatic conditions, most residential houses are ventilated naturally,
mainly by opening windows. However, maintaining the indoor thermal comfort characteristics of a house
by modulating natural ventilation is particularly challenging, as the solution is not explicit. Determining a
solution requires a technique that adjusts openable window area while encapsulating the complexity,
dynamics, and nonlinearity associated with the natural ventilation driving forces and building thermal
behaviour. This work examined the feasibility of predicting the air temperature time‐series of an occupied
naturally ventilated house by an Artificial Neural Network (ANN) model. It was found that the ANN
technique could be used to predict the occupied space indoor air temperature time‐series of the naturally
ventilated house and that this can be used as part of an intelligent window control strategy for naturally
ventilated residential houses.