Lectures and Courses

This master thesis aims to develop and apply a methodology for conducting an environmental Life Cycle Assessment (LCA) of at least one AI system. The project will include the following phases: 1. Literature review: Conduct an extensive review of LCA methodologies applied to digital and AI technologies, as well as case studies in the energy sector, for example in PV forecasting and demand side management. 2. System boundary definition: Define the scope and boundaries of the LCA, identifying all stages in the AI system's life cycle (data collection, model training, deployment, inference, maintenance). 3. Data collection and modelling: Collect primary and secondary data related to energy consumption, hardware utilization, and software development across all stages in the AI life cycle. Model the life cycle using appropriate LCA software (guided by a domain expert from PSI). 4. Impact assessment and analysis: Quantify environmental impacts using standard indicators (e.g., material and energy use, greenhouse gas emissions). Identify major contributors and potential areas for improvement. 5. Interpretation and reporting: Present findings with recommendations for reducing the environmental footprint of AI tools in energy systems.

This master thesis shall focus on conducting a full LCA of one or several AI-based system(s) operated in the energy sector. By analysing and comparing the environmental impacts across the full AI system lifecycle (development, training, deployment, and usage phases), the study will provide comprehensive insights into trade-offs, drawbacks, and potential for optimisation in AI systems used in energy systems and beyond.

For applying, please forward your CV and transcript to Dr. Fabian Heymann, fheymann@ethz.ch

This project aims to develop a methodology for forecasting the locations and, eventually, the impact of DCs on energy systems. The research project will follow a three-phase approach: 1. Literature review: Conduct an extensive review of how currently DCs are modelled in energy systems and on how locational factors can be derived using geospatial modelling and analysis. 2. Modelling: Building a forecasting model that uses available data on DC locations and other energy system-relevant criteria on municipal level. This step will make use of geodata, geospatial analysis and includes data pre-processing, scenario development, model development and analysis to identify key drivers that affect DC locations. 3. Impact analysis and visualization: Eventually, the project will involve an analysis of current and foreseeable DC locations and attempt to quantify their effects on local energy systems (using findings from literature review), for the case of Switzerland.

The outcomes will provide highly relevant, timely and quantitative insights for policy design and decision making to achieve effective decarbonization of energy systems under growing digitalization and energy use.

For applying, please forward your CV and transcript to Dr. Fabian Heymann, fheymann@ethz.ch

This project aims at developing a literature-based methodology (techno-economic framework) for analysing the effects of AI on energy systems, building on a thorough, structured literature review while extracting key techno-economic characteristics of the documented use of AI in energy systems. The research project will follow a four-stage approach: 1. Literature review: Conduct an extensive review of the intersection between AI and energy system modelling, including use cases in scenario analysis, optimization, and efficiency gains. 2. Development of techno-economic framework: Based on the literature review, the applicant will develop a techno-economic framework that allows to study the interaction and impact of a digital technology such as AI on the various components of energy systems (supply, demand, networks). 3. Modelling: Together with an expert from the Paul Scherrer Institute (PSI), the successful applicant will apply the techno-economic framework to the Swiss TIMES model. This may include data pre-processing, scenario development and identifying key scenario drivers. 4. Result analysis and interpretation: Analysing and comparing the outcomes of STEM (that includes a modelling framework of AI) while comparing it to a baseline scenario (without AI). Comparative analysis and visualization of results and achievable transition pathways.

The outcomes will provide highly relevant, timely and quantifiable insights into the potential contributions of AI for decarbonizing energy systems.

For applying, please forward your CV and transcript to Dr. Fabian Heymann, fheymann@ethz.ch

Task description: 1. Literature review on the factors which may have a local influence on the development of wind farms. 2. Defining the geographical scope of the study (it is suggested to select the United Kingdom). 3. Data collection, preprocessing and GIS analysis of the factors. If required, clustering of the factors. 4. Modelling the visual impact of wind turbines using available local population data, landscape scenic value data and viewshed analysis tools. 5. Identification of relationships among the chosen factors: whether high resource availability geospatially coincides with high visual impact or high property prices or other considered aspects (e.g., existing infrastructure). 6. Interpretation of the results: what do the uncovered relationships mean for further resource assessment steps? (possible simplification due to inference of a certain factor from others) 7. If time allows, building scenarios to analyse potential trade-offs between the factors. The scenarios will be applied for a specific region defined as a geographical scope of the study. However, it will be ensured that the approach developed in this project can be transferred to other geographical regions.

The goal of this project is to create a model which will explore the relationships between different geospatially distributed factors that can be used in the feasibility assessment of the local development of wind farms. These factors include, but are not limited to, local social perception of RE technologies, land prices, visual impact of wind turbines including visibility and landscape scenic value, wind resource availability, land use limitations, and existing grid infrastructure. The model will consider the potential visibility of wind turbines to the local population through a viewshed analysis. Therefore, this interdisciplinary project will integrate Geographic Information Systems (GIS) analysis and machine learning techniques, building upon prior research conducted within the Chair. In contrast to previous research, the main novelty of the study is to add the property price and visual impact into the analysis of the feasibility of future wind technology development, which will be conducted in high spatial resolution.

Alena Lohrmann (alohrmann@ethz.ch) Ruihong Chen (ruchen@ethz.ch)

This project will apply MCDA to evaluate and compare different wind project location alternatives. The tasks include identification of objectives with associated attributes, data collection for each attribute if needed, weightings of different objectives, and assessment of alternatives. This project will be contributing to WIMBY and taking input data from consortium partners. The student is expected to formulate the MCDA based on the provided impacts [5], process the collected results from WIMBY, conduct expert interview if necessary, identify trade-offs among multiple criteria, and visualize the results. The Chair of Energy System Analysis (ESA) together with the Laboratory of Energy System Analysis (LEA) at PSI will provide support and guidance on MCDA, data processing & analysis, modelling, and GIS.

In MCDA, each objective is quantified / measured by an attribute. Some of the attributes of the abovementioned objectives have been identified and modelled in the European Horizon project Wind in My Backyard (WIMBY), while others are still to be explored. This project will utilize the results from WIMBY, such as stakeholder mapping and various impact indicators (shadow flicker frequency, landscape scenicness classification, birds & bats mortality, noise level, life cycle emissions, job creation, etc.), to conduct MCDA on case studies. Other impacts that are not yet considered in WIMBY (but within the scope of this project) such as mismatch of supply and demand due to lack of dispatchability, electricity market price impact, end-of-life recyclability, change of local weather, etc., would require further literature review and data collection. The final goal is to showcase spatial differences and stakeholder preferences regarding wind project decisions.

Please forward your CV and transcripts by email to Ruihong Chen (ruchen@ethz.ch)

The main tasks during this MSc project are: • Literature review on prior work performed on DAC and the water nexus. In addition, an analysis of the available lifecycle inventory (LCI) databases will be performed. • Knowledge from the literature review is used to determine the amount of water generated from DAC in a spatially explicit way. Further, water needs for producing different low-carbon fuels will be indicated depending on location-specific conditions. As such, a global geospatial analysis will be performed. • Considering the water production and availability, the possibility of producing low-carbon fuels is investigated with a special focus on water-scarce regions with substantial renewable power potential. • Finally, recommendations will be given on how to consider water (scarcity) towards a potential large-scale integration of CDR and solid sorbent DAC.

The main project goal is to determine the production of water from DAC that can be used to produce such low-carbon fuels. The assessment will be performed in a spatially explicit way using a global geospatial analysis, given that water availability is very location-specific.

If you are interested, please send your motivation and transcript of records to Tom Terlouw (tom.terlouw@psi.ch) and Alena Lohrmann (alohrmann@ethz.ch).