Completed research projects

Completed PhD theses                                                                                                                          



Completed research projects


COMTES - Combined development of compact thermal energy storage technologies; Line B – Liquid Sorption Heat Storage
The three heat storage technologies that are addressed in the COMTES project are ‘compact thermal storage technologies’, that need less volume and have much lower heat losses over the time (a season) than conventional water based storage systems. Moreover, the new systems demand less collector area compared to water-based seasonal storage systems.
This makes these technologies promising candidates for the next generation of seasonal solar heat storage technologies that will enable the step to nearly fully renewable heat supply of buildings in Europe.
The overall goal of the COMTES project is the technological development and demonstration of three compact thermal energy storage technologies, in three parallel development lines.
Empa was involved in line B: Liquid sorption heat storage; heat storage by absorption of water vapor in a liquid. Sodium hydroxide (NaOH) is used as the liquid sorption material.
Funding: An European Union Seventh Framework Program Project, EU grant number 295568
Partners: SPF Institute for Solar Technologies (Research), Kingspan Environmental Ltd (Industry)
Contact: Robert Weber

NaOH-Speicher für saisonale Wärmespeicherung
In the scope of this project, a NaOH sorption storage for seasonal solar thermal storage has been developed, built and tested. Rather than sensible thermal storage, this novel concept is based on the principle of the a sorption heat pump operation. In this manner a theoretical energy density increase of 3 to 6 times compared to that of the sensible thermal water storage is achieved and the system is not susceptible to thermal loss during storage time. This system was built in two steps, initially with one heat pump stage and then extended to a two stage system to increase the output temperature. Computer simulations accompanied this deve opment in order to improve the design process.
Funding: Swiss Federal Office of Energy, Pilot and Demonstration Project
Contact: Robert Weber



Nanoterra heatreserves
The project aims to make recommendations about the implementation of control, communications, and business schemes for enabling thermal loads to provide ancillary services in the form of control reserves for the Swiss power grid. Ancillary services provide a fast-reacting compensation for a power shortage or surplus in the transmission grid.
Thermal loads such as building HVAC systems and household appliances have an inherent thermal storage capacity, which provides flexibility in their power consumption without compromising their original purpose. Hence, one can envision effective demand response schemes exploiting these thermal loads to balance the power grid locally, reducing transmission congestions, improving ancillary service market operations, and reducing power peaks. Most importantly, this facilitates the integration of renewable energy sources, which critically rely on ancillary services today. heatreserves is the first external project using the ehub platform after its launch

CCEM ideas4cities

In Switzerland and many other European countries, the future energy system will rely heavily on renewable energy. This will cause an important reengineering of this part of the electrical infrastructure. Therefore, a massive penetration of distributed power sources and distributed storage devices calls for a new layout and system design of the urban energy system.
The results of our studies will form the basis for the planning such systems and grids. Both the design (planning) of the energy system and the operation will be considered.

The developed microgrid framework consists of independent resource and grid agents communicating with each other. The goal is that the grid operates in a safe state as it can determine the load on its lines and the resources can operate flexibly and independently. This cooperation based, distributed control scheme has a cycle time of 100ms leading to fast corrections form optimal trajectories. With this scheme one can also operate in islanding mode with the ability to connect and disconnect whole districts to the distribution grid.

Optimized local control scheme for active distribution grids using machine learning techniques

Description: The decentralized control scheme based on Machine Learning (ML) technique proposed in [1] is applied to the NEST microgrid, allocated in the EMPA Laboratory in Dübendorf. However, due to the low voltage magnitudes over the grid, in order to have consistent results, the battery is forced to operate into a high injection regime, increasing the local voltages and allowing to adopt the Reactive Power Control (RPC) as active measure to achieve network-wide optimal operation. In this approach, the online optimization problem is performed in two stages. First, a day-ahead optimization problem for the BESS is implemented with the objective to impose high injections during noon hours and guarantee a secure operation. Second, a centralized, OPF-based scheme is used to generate a sequence of optimal DER setpoints accounting for BESS injections. Finally, the local DER controllers are developed as explained in [1] and applied in the real-time operation.
[1] F. Bellizio, S. Karagiannopoulos, P. Aristidou, and G. Hug, "Optimized local control schemes for distribution grids using machine learning techniques," IEEE Power and Energy Society General Meeting, 2018.
Partner: ETH Zürich
Contact: Ralf Knechtle


Google Summer of code: Visualization Dashboard for Empa-NEST
Contact: Reto Fricker


Completed PhD theses


Marc Hohmann - Predictive optimal operational strategies for urban energy systems
Distributed energy systems require active control strategies in order to balance multiple energy streams and to ensure a reliable, economic and environmentally friendly operation. Multi-carrier energy systems offer additional degrees of freedoms compared to single carrier operational schemes managing only electricity, gas or heat. The objective is to increase energy efficiency, decrease carbon emissions and handle the uncertainty and intermittency of renewable energy sources. Active control strategies are therefore crucial to the transformation of energy systems, a central element of climate change mitigation measures.
The objective of this work is to develop improved central local energy management systems for urban districts. These must be able to include non-convex characteristics of energy conversion processes, decision-making for short- and long-term storage and avoid suboptimal results due to the self-interest of the energy system's agents. Using mixed-integer linear programming, energy conversion processes can be modeled with high accuracy. Aggregation methods can reduce the computation load when making long-term decisions. Mechanism design and locational marginal pricing can ensure that optimal operation modes are achieved in complex systems.
A prototype controller is implemented and tested on the NEST platform, located at Empa in Dübendorf, Switzerland. Distributed controller architectures are compared to the central control scheme proposed by this work.

Julien Marquant - Facilitating Multi-Scale Urban Energy Systems Modelling
website | poster
Network-level energy optimisation approaches can determine the optimal location of generation technologies within a region and the optimal layout of energy distribution networks to link them. This is a multi-scale problem as it must encompass decentralisation at the building, district, city or regional scale. However, computational limitations arise at larger spatial scales if full resolution is re-tained, and difficulties emerge in the quantification of different urban agglomeration levels when attempting to model network be-haviour at multiple spatial scales. Nevertheless, a tractable multi-scale optimisation of a large urban area is possible using a clustered, aggregated energy hub repre-sentation.
The main objective of this research is to model and optimise the interaction of complex energy net-works in order to understand the trade-off between centralised and decentralised energy systems at different urban scales. A method for multi-scale urban energy systems modelling with a hierarchical approach will be developed to facilitate this. An evaluation of the approximations necessary to make this modelling computationally feasible will then be conducted. Finally the process developed will be applied to a case study, and the outcomes analysed in the context of the Swiss Energiewende 2050.

Georgios Mavromatidis - Modelling of decentralized energy systems under uncertainty
website | poster
“Distributed Energy Systems (DES) are expected to be a core component of future urban energy systems incorporating a multitude of generation and storage technologies to supply the energy needs of urban districts. The motivation for this paradigm shift stems from global challenges like climate change, the high degrees of urbanization, cities’ high energy demand density and the potential for building integrated renewables.
The challenge to optimally design and operate DES relies heavily on modelling; however, as with any numerical modelling effort, models for the optimal DES design are irrevocably affected by uncertainty. Human behavior and the uncertain future global economic and energy outlook are only a subset of factors that can introduce uncertainty to a DES model.
Therefore, the aim of this research is to create an integrated modelling framework that will in a systematic way incorporate uncertainty into the design process of distributed urban energy systems. Embarking from a design model based on the energy hub concept, techniques from Uncertainty and Sensitivity Analysis are employed to investigate uncertainty’s impacts and identify its main drivers. Subsequently, techniques for Optimization under Uncertainty are used in order to select single optimal designs that will be robust in the face of uncertainty.”

Somil Miglani – Optimal energy system and retrofit measures for residential buildings/districts: A detailed modeling perspective on ground source heat pumps
website | poster
Future cities, urban areas and buildings are expected to undergo a transformation towards more sustainable energy systems. This transformation involves a move towards increased use of renewable energy resources, decentralized forms of energy production, energy efficient buildings, thermal networks etc. The aim is to achieve such a transformation optimally, considering economic and environmental constraints. The current energy systems in buildings, especially heating systems, are based on fossil fuels and must give way to more energy efficient and environmentally benign alternatives. Each building can either be connected to a district heating system or it can be individually heated through a fully decentralized standalone systems such as solar thermal for instance. Since the potential for decentralized energy sources exhibits a high degree of spatial and temporal variability, the optimal integration of these technologies in existing buildings remains an open question.
This research aims at investigating the technological tradeoff between renewable energy based standalone systems and a small scale district heating system taking the total costs and carbon emissions savings into account. More specifically, methods will be developed to evaluate optimal configurations of energy systems for single buildings and clusters of buildings representing spatially and temporally differentiated energy demand and supply patterns. Finally, this analysis of the above mentioned technological tradeoff is carried out on multiple case studies representing urban areas, diverse with regard to technical parameters that differentiate them in their spatial characteristics.

Boran Morvaj - Holistic optimisation of distributed multi energy systems for sustainable urban areas
website | poster
The aim of this PhD project is to develop a holistic optimisation model of distributed multi energy sys-tems to explore how existing urban areas can be best transformed into sustainable ones. Models are based on the energy hub concept. It combines models of distributed energy resources, electrical grids, decentralised district heating networks and building energy systems into one integrated optimisation model in order to find synergistic effects between them. The model is used to determine the optimal design and operation of distributed multi energy systems under different objectives such as the multi-objective minimisation of carbon emissions and costs, energy import independency and optimal power flow. Furthermore, the impact and benefits of decentralised district heating networks as well as large-scale integration of renewables in the existing electrical network are analysed. Finally, key influencing factors affecting the optimal solutions are identified.

Christoph Waibel - Hyper-Heuristic Framework for Multi-Objective Optimization of Urban Systems
website | poster
A major share of global resource and energy consumption can be associated with cities. Considering the ongoing trends of urbanization and population growth it becomes evident that their evolution is a crucial keystone in tackling global environmental and economic challenges. It has been shown that cities and buildings are far from the ”efficiency possibility frontier” and could achieve much higher utility with less energy and resource input. One reason for inefficient designs can be found in the inherent complexity of the design process, which requires the expertise of multiple disciplines. The traditional approach to cope with this is to separate responsibilities and to exchange information in sequential steps. However, generating, processing and exchanging information is a costly practice and every change in design requires re-evaluation of other related disciplines. Thus, only a low degree of reciprocity is realized.
Holistic optimization methods may overcome this issue, as they can inform the design process by exploring vast numbers of design solutions across multiple performance criteria simultaneously. One of the practical challenges lies in the selection of appropriate optimization algorithms best suited to specific problem formulations. Hyper-heuristics deals with this by introducing a higher-level method for automatically selecting and tuning a tailored heuristic from a set of algorithmic operators. This research focuses on the development and application of a hyper-heuristic framework for multi-objective urban design, including building morphology and urban energy systems. Questions to be addressed using the hyper-heuristic optimization framework include the range and degree of multi-energy network connectivity within and across neighborhoods, the degree of densification for optimal demand and renewable energy generation, and the use of building standards (Passivhaus, nZEB, active house) in the context of a connected multi-energy-grid. Hyper-heuristic methods have the potential to change the overall design approach and enable holistic design and planning, where reciprocities between different disciplines and scales can be captured, thus leading to more efficient urban systems.