This dataset serves as training data for modeling the temperature field emanating from open-loop groundwater heat pumps.
The dataset was simulated in 2D with Feflow using cut-outs from interpolated hydrogeological measurements of the Munich, Germany, region. Heat pump locations are chosen based on realistic positions, and extraction rates are adapted to fit the available groundwater.
To prepare the data for machine learning, it was transformed from unstructured to structured data (320x1280 cells) with Python. Inputs include hydrogeological parameters such as hydraulic conductivity [m/d], hydraulic head [m a.s.l.], aquifer thickness [m], and Darcy velocities [m/d], as well as 2D-embedded operational pump parameters such as maximum flow rate [m^3/d]. All hydrogeological parameters were extracted prior to heat pump operation. Heat pumps are operated under seasonal load as depicted in `general/normed_flow_injection_series.npy` and `general/temperature_injection_series.npy`. The former contains the normalized flow rates of the heat pumps, while the latter contains the corresponding injection temperatures.
The prepared data is split into training and test data. For training data, each datapoint consists of a set of inputs and labels; for test data, the labels are hidden. The summarized min and max values for normalization are contained in `general/normalization_info.yaml`. The order of inputs is specified by `index` in the info-yaml.
During testing, we only evaluate the last four timesteps, i.e., an annually repeating "steady state". For training, users can choose to train with the same data format, i.e., "training_data/"training_inputs_and_labels_steadystate", or use the additional information of 36 prior timesteps since the initial installation of the heat pumps, i.e., "training_data/"training_inputs_and_labels_timeseries", or utilize the simulated end-time velocity fields, which are not available as inputs for testing but during training in training_data/"training_inputs_and_labels_interimvelocities". For the interim velocities, additional normalization and dataset information are summarized in "general". For more specialized data, please refer to the raw data in https://doi.org/10.18419/DARUS-5920.
