GET (GEWISS Expert Tool)

The GEWISS Expert Tool (GET) is a simulation tool developed by Thomas Preisler, Nils Weiss and Antony Sotirov from the HAW Hamburg as part of the name giving GEWISS Project. The data model as well as the actual data is provided by Ivan Dochev from the HCU Hamburg. Additional contributers are Arjun Jamil (HAW Hamburg, simulation logic and UI/UX design), Ev Köhler (HAW Hamburg, simulation logic) and Hannes Seller (HCU Hamburg, simulation logic). The project is lead by Wolfgang Renz (HAW Hamburg) and Irene Peters (HCU Hamburg).

The GET-Icon is made by Freepik from www.flaticon.com.

Among other things the project aims at simulating the development of the building stock in the city of Hamburg (Germany) until 2050. GET intends to let expert users, like city planners, simulation different development scenarios for a given building stock under certain conditions. These user can define certain conditions for the building stock development and then simulate how these conditions might influence the building stock in the future. Thereby, the users can get an idea about what measuresments are needed to reach certain goals regarding the reduction of heat demand and CO₂ emission.

The tool consists of two projects get-simulator contains the simulation logic, the data model and the required input data while get-ui contains everything related to the user interface. Both projects are realizes using Java 8 and the UI is build using the JavaFX Framework.

Remark: This is not the repository used for the actual development of the tool. Due to privacy reasons, we're not allowed to share all of our input data with the general public (see section Building Input Data). We use a private repository for the actual development and use this one to share the source code without the sensitive input data with the public. Therefore, only major updates are commited to this repository.

The following describes briefly the simulation approach, the building input data and how to build and run the application on your computer.

Simulation Approach

The simulation approach is mainly based on a ranking of all buildings according to their likeliness for (hull) renovation. Currently, all buildings are ranked according to the last year of renovation, so buildings that have been renovated a long time ago or old building that have not been renovated before a more likely to be renovated in the current year of the simulation. (The simulation is a discrete one, where each simulation step is one year.) After the ranking a user configurable amount of buildings with the highest ranking score are selected for (hull) renovation in the current year of simulation.

The user can affect the ranking of buildings with the definition of so called Modifiers. These modifiers specify a modifier factor for buildings that match certain criteria. E.g.: One-family houses build before 1953 located in a certain city quarter. If a building matches this criteria its ranking score is multiplied with the given modifier factor, thereby increasing/decreasing the scoring value and influencing the ranking.

A more detailed description of the simulation approach can be found in the paper "Proposing Multiple-Criteria Ranking to Simulate Building Renovation in Cities" by Arjun Jamil, Nils Weiss, Thomas Preisler and Wolfgang Renz, that has been presented at the ASIM 2018 symposium in Hamburg.

Git Checkout

You can checkout the GIT repository either via command line:

git clone https://github.com/gewiss/gewiss-expert-tool.git

Or you can download it from your browser and then unzip it.

Building Input Data

The main objective of the GEWISS Expert Tool is to simulate the building stock development in the city of Hamburg, but the simulation approach is also suitable for other cities. Unfortunately, due to privacy reasons we are not allowed to publish the building data set of Hamburg as an example. Therefore, we created a small anonymized dataset for testing and publishing purposes.

To do that, we took a real constellation of buildings and moved them to a location in the North Sea, close to Neuwerk Island. The coordinates were affine transformed to make them unrecognizable, but still to resemble a real urban area. Attributes like street names, names of heating companies and so on were changed where needed. Also, the actual names of the quarters and districts were changed to real ones from other places in Hamburg. Finally, the building typology for non-residential buildings of our project partner Ecofys was changed to a generic one.

All the data required for the simulation is stored inside a SQLite database. So, in order to use your own data set you have to change the value of the DB_FILE constant in the SQLiteConnection Java file to the file name of your data set. The Java file can be found in the get-simulator project directory in the de.hawhh.gewiss.get.simulator.db.dao package folder. Your data file has to be copied to the get-simulator/src/main/resources/ folder. Be aware that whenever you change the data set file you have to rebuild the application! How to do so is described in the next section.

Table structure

The following briefly describes the SQL-tables required for the simulation tool.

Table gewiss_buildings_v_1 (Building table)

Column Name	Type	Description
alkis_id	String	Unique Building ID
geomwkt	String	2D Geometrie in WKT format
wohnfl	Double	Living Floor Area in square meter
nwg_ngf	Double	Square meter of usage space for non-residential buildings
bezirk	String	The city district
stadtteil	String	The city quarter
stat_gebiet	String	Statistical area (smaller than quarter)
baublock	String	Urban block (smaller than statistical area)
bj_alk_dt	Integer	Year of construction
dt_san_year	Integer	Last year of renovation
dt_heiztyp	String	Heating system type
dt_eigentum	String	Building Ownership Information
iwu_typ	String	Residential building type
nwg_typ	String	Non-residential building type
bak_fin	String	Construction age class
cluster	String	ID of building cluster, not used for computation, but retained to allow publishing of results

Table baualtersklassen (Construction age class)

Column Name	Type	Description
klasse	String	Building class
von	Integer	From year
bis	Integer	To year

Table districts_and_quarters (City districts and quarters)

Column Name	Type	Description
id	Integer	Unique table id
district	String	Name of the district
quarter	String	Name of the quarter

Table heat_demand_and_load_at_generation

Column Name	Type	Description
id	Integer	Unique table id
BUILD_TYPE	String	Building type
RENOVATION_LEVEL	Integer	0,1,2 for baseline, EnEV 2014, Passive house
SPACE_HEATING_DEMAND	Double	Demand for space heating needed as generation output (kWh/m²)
WARMWATER_DEMAND	Double	Demand for domestic hot water as generation ouput (kWh/m²)
SPACE_HEATING_LOAD	Double	Peak heat load for space heating (kW/m2)
WARMWATER_LOAD	Double	Peak heat load for domestic hot water (kW/m2)

Table heat_demand_and_final_energy

Column Name	Type	Description
id	Integer	Unique table id
BUILD_TYPE	String	Building type
RENOVATION_LEVEL	Integer	0,1,2 for baseline, EnEV 2014, Passive house
HEATING_SYSTEM	String	type of heating system, e.g. "DISTRICT HEATING"
FINAL_ENERGY	Double	Demand for space heating and domestic hot water as final energy (kWh/m²)

Table primary_energy_factors

Column Name	Type	Description
id	Integer	Unique table id
HEATING_SYSTEM	String	type of heating system, e.g. "DISTRICT HEATING"
ENERGY_SOURCE_TYPE	String	type of energy source, e.g. "NATURAL_GAS", used only for information
PRIMARY_ENERGY_FACTOR	Double	Primary energy factor of the heating system type (ratio)
CO2_2019	Double	CO₂ factor for the heating system type (g/kWh_{final energy}) for 2019 (current)
CO2_2030	Double	CO₂ factor for the heating system type (g/kWh_{final energy}), projection for 2030
CO2_2050	Double	CO₂ factor for the heating system type (g/kWh_{final energy}), projection for 2050

Table costs_building_shell

Column Name	Type	Description
id	Integer	Unique table id
BUILD_TYPE	String	Building type
RENOVATION_LEVEL	Integer	0,1,2 for baseline, EnEV 2014, Passive house
COST_PER_SQ_METER	Double	costs for renovation of the building shell (euro/m²)

Table costs_heating_system

Column Name	Type	Description
id	Integer	Unique table id
HEAT_LOAD_kW	Integer	The peak heat load needed, e.g. 5, 15, 25 kW
HEATING_SYSTEM	String	type of heating system, e.g. "DISTRICT HEATING"
COST	Double	total cost (euro) for exchanging the old heating system with this system

Building and Running the Application

The application is based on Apache Maven so, once you have checked out the repository you can either import the Maven projects get-simulator (simulation logic) and get-ui (JavaFX UI) to the IDE of your choice (we used Netbeans, but Eclipse or IntelliJ work as well) to build and run the it or use Maven and the command line, as decribed in the following part.

Building and Running with Maven

The following assumes that you have installed Apache Maven and can access it via the command line. Use the following command to check if Maven is installed:

mvn -version

If Maven is installed correctly use the command line to open the get-simulator folder and build the simulator:

cd get-simulator/
mvn clean install

This may take some time and in the end you should see something like this on the command line:

[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 48.676 s
[INFO] Finished at: 2018-11-27T15:47:18+01:00
[INFO] Final Memory: 23M/275M
[INFO] ------------------------------------------------------------------------

Now change to the get-ui folder and build the JavaFX UI:

cd get-ui/
mvn arcgis-java:arcgis
mvn clean install

Again it should display BUILD SUCCESS. The mvn arcgis-java:arcgis command is required to download the ArcGIS Java SDK's native libraries to the user's home directory. It is only required to be run once. We us the ArcGIS Java to show the buildings in the simulation on a map.

In order to start the application stay in the get-ui folder and type the following Maven command in the command line:

mvn jfx:run

Building a native executable

To build a native executable for the platform you are running on (Windows, Mac or Linux) first you have to copy the ArcGIS Java SDK's native libaries from your home directory to the get-ui folder. To do so, you have to create a new folder called arcgis inside the get-ui folder. Then you have to copy the jniLibs folder from e.g. /home/thomas/.arcgis/100.2.1/jniLibs to the newly created arcgis folder.

Also you have to get a free licence key for the ArcGIS SDK, you obtain one is described here: https://developers.arcgis.com/java/latest/guide/license-your-app.htm

Once you have obtained the key you have to uncomment line 47 in the GETAPP.java file found in get-ui/src/main/java/de/hawhh/gewiss/get/fx and enter your licence key here:

ArcGISRuntimeEnvironment.setLicense("YOUR-LICENCE-KEY");

Afterwards you can build the native executable with the following command:

mvn jfx:build-native

Keep in mind that this might require additional third party tools depending on your operating system. Maven will state on the command lines which tools are needed if not already installed and a web search should quickly lead to the required tools and instructions on how to install them.

Afterwards you can find the native executable as well as an installer for your platform in the folder get-ui/target/jfx/native.

Miscellaneous

The names of the SQLite databases for the GET are named "GEWISS_buildings.sqlite", "GEWISS_buildings_anonymized_for_testing.sqlite".
Updating the values in the GUI tables ("CO₂ Factors in g/kWh" and "Heating System Exchange Control Matrix") requires double clicking a single cell and then pressing enter after the new value has been inserted.
The transition rate to district heating systems depends on the dt_fw_dist field in the DB, which values of 0 (more than 50m away from an existing pipeline) and 1 (less than 50m to an existing pipeline).
- This eliminates the possibility of buildings with district heating in areas without access to district heating grids. The user can reclassify the DB field and control the behaviour.
- Internally, when the dt_fw_dist value of a building is 1, the transition matrix is taken as it is. If dt_fw_dist is 0, the district heating probability becomes 0% and the rest of the matrix values are normalized to 100%. So the final split of the heating systems might not be the same as the split given in the input matrix!
The starting year of the simulation is currently hard-set in the SimulationParameters class (FIRST_YEAR = 2019).
When data is exported using the GUI (either *.csv or *.xlsx file format), the SEED for the particular simulation run is appended at the name of the file name.
The possible renovation level transitions for the "Heating System Exchange Control Matrix" are 0 to 1, 0 to 2, and 1 to 2. The renovation level transitions from 0 to 2 and 1 to 2 have the same rate of change. Below is a small table showcasing the different acronyms used for the Renovation Levels:

Acronyms for Renovation Levels

UI Table ENUMs	Other names	Type	Description
N/A	Unrenovated	Renovation Level 0	Not renovated
***_ENEV	EnEV 2014	Renovation Level 1	Basic renovation
***_PASSIVE	Passive House	Renovation Level 2	Good renovation

***: RES for Residential buildings; NRES for Non-Residential buildings.

gewiss / gewiss-expert-tool