Summer heat protection verification is required by the Building Energy Act (GEG) for all new building projects and renovations. This verification has been incorporated in the EnEV since the version issued in 2009, and its implementation is based on the German standard DIN 4108-2:2013-02 which describes two methods.
The first method is a simplified model where the solar gains of a room (or zone) are calculated. This is then compared to an allowable maximum parameter, which is a combination of different components such as location (climate region), construction, ventilation, and orientation.
The second methodology is a dynamic simulation with defined boundary conditions. The validation is determined by the over-temperature-hours (OTHs) in Kh/a. This indicator integrates the amount of time that the critical zone’s temperature is higher than a reference value, a graphic description is shown in Figure1. The reference value is given by the location.
It is understandable that these two methods are recognized, as each has advantages over the other. The static method is straightforward, fast and requires a few input parameters. Meanwhile, the dynamic simulation accounts for critical factors like internal gains, ventilation, and shading to provide a better understanding of the building’s energy balance.
Theoretically, the conclusion should be the same using either of the methods described. However, due to the simplification of the first method, it is expected a classification error that may lead to erroneous conclusions.
A Montecarlo analysis was used to compare the two approaches. The boundary conditions for the energy models were established as shown in the norm and the variable parameters were divided into 3 categories: transparent surfaces, construction materials and geometry. Examples of the energy models in 3D-view can be seen in Figure 2. Machine learning algorithms were also used to create an alternative verification model that calculates the over-temperature-hour with greater accuracy. Three different regressions were investigated with the goal of maintaining the static approach’s simplicity while also ensuring the reliability of dynamic simulations. The findings are promising, and we intend to publish them soon.
NEWSLETTER ABONNIEREN
Um über Neuigkeiten auf dem Blog informiert zu werden, können Sie sich hier mit Ihrer E-Mail-Adresse für den Newsfeed registrieren. Sie werden dann per E-Mail benachrichtigt, wenn neue Newseinträge veröffentlicht werden.