Supplementary Material
Real-World Benchmarking of Control Strategies in an Industrial Energy System

This item contains historical scenario data and experiment results for benchmarking
advanced control strategies in the ETA Factory Heating Network High Temperature
(HNHT).


Related Repository

Lademann, T., Clement, A., Witt, J.-N., Frank, M., Zangenberg, J., & Weigold, M.
(2026). ETA Factory Thermal System Operation (Version v1.1.0). Zenodo.
https://doi.org/10.5281/ZENODO.20050236


Contents

1. Scenario Data

The repository includes historical scenario data from 10 August 2024 to
8 December 2025 at 60-second resolution.

The dataset contains:
- Thermal loads
- Electrical demand
- Energy prices
- Emissions factors
- Ambient temperature
- Production mode information

For details on how the scenario data is organized and used within the repository,
please refer to the repository documentation.


2. Experiment Results

The experiment results include simulation and real-world experiments comparing
three control strategies:

- rl:
  Reinforcement learning for model-free optimization

- mpc:
  Model predictive control using a predictive system model

- mbc:
  Market-based control with multi-agent bidding and a local double auction


3. Folder Structure

scenario_data/
  Historical scenario data in CSV format for simulations and real-world
  experiments.

results/
  Experiment results.


4. Result Folder Naming Conventions

real_*
  Real-world experiment.

sim_*
  Simulation experiment.

*_control
  Results generated by the controller.

*_monitoring
  Results generated by the monitoring script with a 60-second sampling time.


5. File Format

- The files with extensions *.1 and *.pkl can be opened and used with the related repository.