GIS-AWBEM

GIS-based Automated White-Box Building Energy Modeling (GIS-AWBEM) is an open-source framework for conducting district-scale building energy analysis. Starting from publicly available GIS data, the framework automatically constructs Level of Detail 1 (LOD1) 3D building models.

This framework is primarily developed for urban planners, building energy analysts, energy suppliers, researchers, and scientists.

---

Motivation

Studying urban district energy transitions requires flexible and detailed building-specific energy performance analysis. Many existing tools and workflows simplify either the building envelope characteristics or the HVAC system representation.

GIS-AWBEM was developed to address these limitations by:

1. Employing a dynamic white-box modeling approach that includes detailed building physics and thermal mass properties
2. Providing detailed HVAC system models for the design and performance evaluation of specific energy systems
3. Using a single building energy simulator, EnergyPlus, thereby avoiding the need for multiple software integrations and interface developments
4. Offering a modular structure that requires only minor reconfiguration for future implementations
5. Being fully open-source, eliminating the dependency on commercial software

---

Workflow

---

Getting Started

Clone the repository:

git clone https://github.com/KIT-IAI/GIS-AWBEM.git

Install the package in editable mode:

pip install -e GIS-AWBEM

---

Examples

Two examples are provided to demonstrate the main functionalities of the framework:

• Example 1 models a district with an enrichment process for building envelope data
• Example 2 models a district without enrichment

In both examples, multiple HVAC systems can be applied.

---

District in Karlsruhe

This example models the Bergwald district (48.9723° N, 8.4648° E), which contains 241 buildings.

The following sections describe how to collect the required input data and run the simulation.

---

1. Collect Geospatial Data from OpenStreetMap (OSM)

Geospatial data must be queried from OpenStreetMap using Overpass Turbo:

https://overpass-turbo.eu/#

Set the bounding box around the district and run the following query:

[out:xml][timeout:25];
(
  way["building"]({{bbox}});
  relation["building"]({{bbox}});
);
out body;
>;
out skel qt;

Export the result as a GeoJSON file (e.g., Bergwald.geojson) and place it inside the Inputs directory.

If missing building-level values are detected (e.g., in District_geo.csv), they should preferably be corrected manually using tools such as Google Earth. Otherwise, missing values will be replaced with the district mean values.

---

2. Weather Data

Weather data should be obtained from the German Weather Service (DWD) using the station closest to the district, i.e., Rheinstetten.

https://www.dwd.de/EN/Home/home_node.html

The weather file must use the EnergyPlus weather format (.epw).

---

3. Internal Gains and Setpoint Temperatures

Daily schedules for internal gains/losses and setpoint temperatures are provided in:

Inputs/Internal gain profiles.xlsx

---

4. TABULA U-values

The overall heat transfer coefficients (U-values) for external walls, floors, roofs, and windows are adapted from the TABULA WebTool:

https://webtool.building-typology.eu/#bm

Wall, floor, and roof U-values are only used when enrichment is not enabled.

---

5. Enrichment

Building heights, archetypes, and construction years are enriched using the following ETHOS.BUILDA datasets:

Building Heights

https://zenodo.org/records/11845992

Archetypes and Construction Years

https://zenodo.org/records/12069755

The enriched data are stored in:

Bergwald_enrich.csv

Material properties, as well as layer numbers and thicknesses, are adapted from HUB4LCA:

https://hub4lca.e3d.rwth-aachen.de/

If a different enrichment methodology is required, modify the enrich() function in pre_process.py accordingly.

---

6. Simulation Configuration

After collecting all required input data, configure the enrich and config dictionaries in run.py.

The following HVAC systems are currently supported:

• Ideal HVAC
• Boiler and DX coil
• Packaged Terminal Heat Pump

Further details are available in the following publication:

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=11457217

---

7. Results

An overall district performance summary is printed to the Python console.

Depending on the selected HVAC system, transient HVAC component results are stored in the Results directory for detailed analysis.

Additional output variables can be added in post_process.py if required.

Available EnergyPlus output variables are listed in the Input Output Reference:

https://energyplus.net/assets/nrel_custom/pdfs/pdfs_v26.1.0/InputOutputReference.pdf

---

District in Heidelberg

This example models a district in the southern area of Heidelberg (49.37386° N, 8.68910° E), containing 58 buildings.

---

1. Collect Geospatial Data from OpenStreetMap (OSM)

Geospatial data must be queried from OpenStreetMap using Overpass Turbo:

https://overpass-turbo.eu/#

Set the bounding box around the district and run the following query:

[out:xml][timeout:25];
(
  way["building"]({{bbox}});
  relation["building"]({{bbox}});
);
out body;
>;
out skel qt;

Export the result as a GeoJSON file (e.g., Heidelberg.geojson) and place it inside the Inputs directory.

If missing building-level values are detected (e.g., in District_geo.csv), they should preferably be corrected manually using tools such as Google Earth. Otherwise, missing values will be replaced with the district mean values.

---

2. Weather Data

Weather data should be obtained from the German Weather Service (DWD) using the closest station, i.e., Mannheim.

https://www.dwd.de/EN/Home/home_node.html

The weather file must use the EnergyPlus weather format (.epw).

---

3. Internal Gains and Setpoint Temperatures

Daily schedules for internal gains/losses and setpoint temperatures are provided in:

Inputs/Internal gain profiles.xlsx

---

4. TABULA U-values

The overall heat transfer coefficients (U-values) for external walls, floors, roofs, and windows are adapted from the TABULA WebTool:

https://webtool.building-typology.eu/#bm

---

5. Enrichment

This example does not include an enrichment process.

---

6. Simulation Configuration

After collecting all required input data, configure the enrich and config dictionaries in run.py.

The following HVAC systems are currently supported:

• Ideal HVAC
• Boiler and DX coil
• Packaged Terminal Heat Pump

Further details are available in the following publication:

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=11457217

---

7. Results

An overall district performance summary is printed to the Python console.

Depending on the selected HVAC system, transient HVAC component results are stored in the Results directory for detailed analysis.

Additional output variables can be added in post_process.py if required.

Available EnergyPlus output variables are listed in the Input Output Reference:

https://energyplus.net/assets/nrel_custom/pdfs/pdfs_v26.1.0/InputOutputReference.pdf

---

Minimum Required Data

• Geospatial data from OpenStreetMap in GeoJSON format
• Daily profiles for internal gains/losses and setpoint temperatures
• EnergyPlus weather data file (.epw)
• Overall heat transfer coefficients based on TABULA building archetypes

Please refer to Example 2 for a minimum-data simulation workflow.

---

Requirements

Python Packages

• numpy
• pandas
• utm
• eppy
• sympy

Energy Simulator

• EnergyPlus 25.1

---

Acknowledgment

We acknowledge the financial support of the Helmholtz Association of German Research Centres (HGF) within the framework of the Program-Oriented Funding POF IV under the Energy System Design (ESD) program.

---

License

The GIS-AWBEM framework is developed and released by the Institute for Automation and Applied Informatics, Karlsruhe Institute of Technology, under the MIT License.

---

How to Cite

@article{GIS-AWBEM2026,
  author={Tajalli-Ardekani, Erfan and Cheng, Haozhen and Kocher, Alexander and Kovačević, Jovana and Waczowicz, Simon and Çakmak, Hüseyin K. and Delibra, Giovanni and Corsini, Alessandro and Hagenmeyer, Veit},
  booktitle={2026 Open Source Modelling and Simulation of Energy Systems (OSMSES)},
  title={GIS-AWBEM: GIS-based Automated White-Box Building Energy Modeling},
  year={2026},
  pages={1-6},
  keywords={HVAC;Three-dimensional displays;Heat pumps;Computational modeling;Atmospheric modeling;Buildings;Urban planning;Ventilation;Computational efficiency;Glass box;Building Energy Modeling;District Energy Modeling;HVAC;Automated Framework;EnergyPlus},
  doi={10.1109/OSMSES69376.2026.11457217}
}