Software tools
Our Laboratory develops and maintains open-source tools that support cutting-edge research and real-world applications in energy systems analysis, building performance simulation, and machine learning for system identification.
ehubX – An optimization tool for the strategic planning of energy systems
More information coming soon.
CESAR-P – A Python-based urban building simulation engine
CESAR-P (Combined Energy Simulation And Retrofitting - Python) is an open-source urban building energy simulation engine. It performs bottom-up modeling of building energy demand using archetype-based parameters and simulates each building individually using EnergyPlus.
- Supports city-scale simulation and retrofit scenario analysis
- Models based on building geometry, age, and use
- Considers passive cooling strategies and calculates emissions and costs
🔗 Access the Tool:
📄 Related Publications:
- Wang, D., Landolt, J., Mavromatidis, G., Orehounig, K., & Carmeliet, J. (2018). CESAR: A bottom-up building stock modelling tool for Switzerland to address sustainable energy transformation strategies. Energy and Buildings, 169, 9–26.
- Orehounig, K., Fierz, L., Allan, J., Eggimann, S., Vulic, N., & Bojarski, A. (2022). CESAR-P: A dynamic urban building energy simulation tool. Journal of Open Source Software, 7(78), 4261.
Contributors: Léonie Fierz, Aaron Bojarski, Ricardo Parreira da Silva, James Allan, Sven Eggimann (with contributions from Danhong Wang, Jonas Landolt, Georgios Mavromatidis, and Kristina Orehounig to the original CESAR MATLAB version)
SIMBa – A machine learning toolbox for stable system identification
SIMBa (System Identification Methods leveraging Backpropagation) is a Python toolbox for identifying discrete-time linear state-space models using PyTorch's automatic differentiation framework. It ensures model stability and allows prior knowledge integration, such as matrix sparsity or structural constraints.
- Ensures stability via LMI-based parametrization
- Supports structured and interpretable models
- Works with MATLAB and is pip-installable
🔗 Access the Tool:
📄 Related Publications:
- Di Natale, L., Zakwan, M., Svetozarevic, B., Heer, P., Ferrari-Trecate, G., & Jones, C. N. (2024). Stable linear subspace identification: A machine learning approach. In 2024 European Control Conference (ECC), 3539–3544. IEEE.
- Di Natale, L., Zakwan, M., Heer, P., Ferrari-Trecate, G., & Jones, C. N. (2024). SIMBa: System identification methods leveraging backpropagation. IEEE Transactions on Control Systems Technology.
Contributors: This project is jointly led by Loris Di Natale and Muhammad Zakwan, with the participation of Bratislav Svetozarevic, Philipp Heer, Giancarlo Ferrari Trecate, and Colin N. Jones.