Towards The Integration Of The Urban Heat Island In Building Energy Simulations

Weather conditions in an urban environment differ from the conditions in a rural environment. This phenomenon is known as the urban heat island (UHI) effect. In this study the urban climate was monitored at five locations in Rotterdam (the Netherlands) for a period of 1.5 years. The urban heat island intensity was subsequently calculated by the difference between these results and measurements of the Royal Dutch meteorological institute (KNMI) at a rural area 5 km from the centre of Rotterdam. A data-driven method based on a neural network was used for the prediction of the UHI intensity. In this method the UHI intensity at a specific time is calculated as a function of eight weather parameters at that specific time as well as the preceding three hours. The results show a mean squared error on the test data of 0.18 °C. The results therefore indicate that the model reproduces the transient behaviour of the UHI intensity in an accurate manner. This approach can therefore be used to convert weather data of a rural area in weather data for an urban area and subsequently be used in building performance simulations.