Implementation of an automated building model generation tool Sergio Leal, Stefan Hauer, Florian Judex and Stefan Gahr Energy Department - Sustainable Building and Cities AIT - Austrian Institute of Technology Vienna, Austria Email: [sergio.leal, stefan.hauer, florian.judex, stefan.gahr.fl]@ait.ac.at Abstract—Building modeling and simulation are resource in- tensive tasks, mostly with respect to personnel time and therefore costs, which could benefit from automation. In recent years, there has been an increased interest in automating the task of building modeling, but due to the uniqueness of each building, automated building modeling solutions are very demanding. Consequently, commercial as well as open-source solutions have been made available in the past years, but really reliable solutions are very scarce. This paper presents an approach to automate the generation of thermal building models. A tool has been developed which serves as a base for future studies and improvements in the auto- matic model creation and calibration field, reducing the modeling effort and resources for thermal building models were reduced significantly. The output of the model generator supports the validation and analysis phase of ISO 50001 regarding forecasting the building performance for validation purposes. While building planners benefit the most from this approach, it will also be the basis to support other modeling approaches automatically which can be used for model-based controls in the field of Building Management. I. I NTRODUCTION Increasing energy prices and energy demand pose new challenges for modern society. One possibility to control energy consumption in buildings is based on estimating the thermal loads of a building. This may be determined based on the Energy Performance Certificate (EPC, in most Euro- pean countries the mandatory legal implementation of Direc- tive 2010/31/EU, 2010 [1]) input and the energy consuming components in the building(such us heating cooling and air conditioning (HVAC) and lighting system loads, solar gains, plug-loads, and many more). Moreover, the typical physical properties of the construction type chosen, heating and cooling loads of the building as well as electric load profiles can be computed. Building performance may then be estimated by comparing the loads to actual amount of energy used for heating, cooling and electricity. This comparison gives first insights whether the object monitored and evaluated consumes a reasonable amount of energy. User awareness with respect to energy consumption, and the associated CO 2 emissions, may also profit from this assessment. This paper presents the implementation of an automated building energy simulation model generation tool, which re- duces cost related to building modeling and simulation. Advan- tages include a reduction in resources and complexity related to building model and simulation. Consequently, the employment of energy performance modeling and simulation on the entire building process is made more feasible. The results of the simulation are integrated with the estimation of the internal loads and compared the actual energy use, also ideally re- ported by the individual subsystems or components, leading to first assessments whether the object monitored and evaluated consumes a reasonable amount of energy. Initially, data that is not available for simulation purpose shall be replaced by general assumptions (for example, standard consumption profiles for offices). Later, further building data and system descriptions may be added by the user, improving the accuracy of the simulation results. This allows a definition of the cross- dependencies of existing data and the benefits thereof, for example getting a more precise energy consumption profile by adding ventilation specifications. It should be mentioned at this point that these simplified building models are not suitable for unsupervised control of buildings, due to the simplification of the envelope and system models. A modular architecture allows for energy components to be added to the baseline simulation based on building characteristics. As such, the baseline simulation software again allows different levels of detail, depending on existing building data. A visualization interface supports the validation process. A comparison between the simulated baseline and monitored data allows a better understanding of the building performance. Figure 1 is an example of an automatically generated building model using an vertical H layout. It shows how the final 3D model looks like ready for the thermal simulation. The thermal zones were generated automatically based on the user input. A detailed explanation of how the building model will be generated based on its inputs is shown in section III. Fig. 1. Example of an automatically generated building model using a H layout- 15 floors consisting of 129 zones in total (time to generate building model: 218.56 sec).