4 Tools

The purpose of the Renewable Energy Hub Optimizer (REHO) tool is to optimize energy systems at building-scale or district-scale, considering simultaneously the optimal design as well as optimal scheduling of capacities. It allows to investigate the deployment of energy harvesting and energy storage capacities to ensure the energy balance of a specified territory, through multi-objective optimisation and KPIs parametric studies.

It exploits the benefits of two programming languages, namely AMPL and Python:

The core optimisation model is written in AMPL: objectives, constraints, modeling equations (energy balance, mass balance, heating cascade etc.). This is the result of the PhD thesis of Paul Stadler (2019).
All the input and output data is passed to the model through a Python wrapper. This data management structure is used for initialization of the optimisation model, execution, and results retrieval. It was developed by Luise Middelhauve during her PhD thesis (2022).

The REHO tool allows to study the various configurations of multi-flows and multi-services energy systems for urban areas, in a district-level integrated approach. This tool help the IPESE lab to understand the benefits of sharing infrastructures and energy harvesting, conversion, storage capacities and to assess their optimal level of centralization vs. decentralization for the different types of districts. To recall, a building-scale optimisation treats all the considered buildings independently, i.e. each of them is optimised regardless of the presence of the others. In contrast, a district-scale optimisation considers the whole building stock as being able to interact, and exploits the synergies of the overall system defined by the buildings to be optimised. In Python, the building-scale optimisation is obtained using the decentralized option, whereas to obtain the district-scale optimisation the centralized option is chosen. Centralized optimisation provides a single optimisation for the whole district. Knowing that this is computationally demanding, a decomposed method has been developed, called Dantzig-Wolfe decomposition. This decomposition approach takes a longer time to obtain results than the classic centralized formulation for few buildings. However, it has been noted that on large districts the decomposition formulation has a linear increase of the CPU time while the centralized formulation has an exponential increase. Therefore, in the case of large districts, the decomposition formulation should be encouraged. To summary:

View of the energy hub at the building-scale: In this perspective, each building is considered independently. The district is modeled as a collection of individual building. This approach leads to a decentralized design strategy in which the multi-objectives optimisation of the building energy systems are considered with the individual perspective and performed sequentially.
View of the energy hub at the district-scale. In this perspective, all building are considered at the same time. The district is modeled as a single entity. This approach leads to a centralized design strategy in which the distributed energy systems are designed and operated to a central objective. The decomposition method presented above is applied to overcome some runtime issues.

In addition, another tool called QBuildings is developed. This is a GIS database used to characterize the territory from an energy point of view (end-use demand, building stock, endogenous resources). This database is built by gathering different public databases and combining them with SIA norms. The norms of the Swiss Society of Engineers and Architects (“Société suisse des Ingénieurs et des Architectes”) in French are Swiss national rules for the art of building.