Artificial neural network-based fluid flow prediction in a solar chimney

Solar chimney is a potential solution in ventilation buildings. The key benefit is that it gains great energysaving up to 50% of manual ventilation energy. Many works relating to numerical simulation of the solar chimney demands a huge computational cost. Recent years, Machine Learning technique has been utilized as an alternative solution to deal with this issue. In this study, we obtain a simple Artificial Neural Network to predict the fluid field observed in a typical solar chimney. The trained model is then applied to predict the temperature distributions as various heat fluxes. As the result, the prediction values achieve a good accuracy.