Bachelor's and Master's Theses or Research Internship / Bachelor- und Masterarbeiten oder Forschungspraxis

Spatio-Temporal Forecasting of Wind Power

Research method: Modeling

Research question:

How can spatial information improve temporal-only forecasting methods?

Possible approach: 

• Literature research on temporal and spatio-temporal methods for wind power forecasting (e.g., copula models, flow-based generators, graph neural networks, ARIMA model, RNNs, ...)

• Definition of benchmark data sets and state-of-the-art temporal forecasting methods
• Implementation of spatio-temporal and banchmark temporal methods of choice
• Evaluation of its performance compared to the benchmark

Your background/interests:

• Passion for working with data
• Motivated to program in Python, ideally already some basic Python programming skills
• Interest in understanding statistical/ML methods conceptually

Probabilistic PV / Load Forecasts

Research method: Protoytping, Modelling

Research questions: 

• How can we quantify the uncertainty via probabilistic forecasts of, e.g., PV production and load?
• What constraints do we face on the algorithm side if we want to have embedded forecasts?
• How can we use conformal prediction to go from point forecasts to probabilistic forecasts?

Possible approach: 

• Literature Research of State of the Art of PV or Load Forecasting
• Find Research Gap 
• Implement own ideas
• Validate Results e.g. with PV on institute's roof

Suggested Readings:

• A guideline to solar forecasting research practice: Reproducible, operational, probabilistic or physically-based, ensemble, and skill (ROPES) | Journal of Renewable and Sustainable Energy | AIP Publishing
• Core Concepts and Methods in Load Forecasting: With Applications in Distribution Networks | SpringerLink
• Energy Forecasting: A Review and Outlook: pierrepinson.com/wp-content/uploads/2020/10/Hongetal2020.pdf

Concepts and Tools:

• Python/Julia: pandas, DataFrames.jl
• Time Series Forecasting and Machine Learning: sktime, functime, Darts, scikit learn, MAPIE

Real World Distribution Grid Topology Clustering

Research method: Prototyping

Research question:

How can real world low voltage distribution grids be clustered and and categorized? What is a collection of parameters describing a wide range of real world distribution grids?

Possible approach:

• Literature review on LV distribution grids, allocation methods and existing benchmarking networks

• Collection of existing LV distribution grids, generation from open accessible data (e.g. OpenStreetMaps) using ML tools for data aggregation

• Clustering and Analysis

• Derivation of parameter collection

Sebastian Eichhorn
sebastian.eichhorn@tum.de

EnergyOS: Operating System Internal Resource and Data Management

Research method: Literature Review, Prototyping

Research question:

How can we store and retrieve data most efficiently?
Can we integrate caching and parallelization of requests?

Possible approach:

• Literature review on data access and storage for time series applications, and data caching.

• Creation and comparison of data management schemes
  • e.g. resource data stored using: (1) python data structures (local dictionaries and lists), (2) json files, (3) csv files, (4) common database like Postgres or MongoDB, (5) specialized time series database like InfluxDB.

• Evaluation of schemes, their performance, how they scale (i.e. complexity in terms of time and storage).

• Design of data caching procedure and parallelization of OS tasks.
• Implementation of Resource and Data Management design in EnergyOS.

Your background/interests:

• Energy Management Systems
• Data Management / Databases
• Programming experience helpful (python)

Ehimare Okoyomon
e.okoyomon@tum.de

EnergyOS: Scheduling and OS-controlled Schedule Generation in a Multi-App EMS Platform

Research method: Literature Review, Prototyping

Research question:

What level of control should the EnergyOS have in creating and executing operation plan schedules for energy components?
How can this be used in practice and what are the advantages and challenges of such an approach?

• Literature review of common EMS scheduling / optimization problems and state of the art methods for solving.

• Implementation of one or more scheduling solutions in EnergyOS.
• Creation of an interface to provide EnergyOS with a set of constraints, objectives, and data/resources.
• Selection and implementation of use cases (EnergyOS apps) to demonstrate how OS-controlled scheduling works.

Your background/interests:

• Energy Management Systems
• Optimizations
• Programming experience helpful (python)

Ehimare Okoyomon
e.okoyomon@tum.de

Project EDL: Building a Digital Platform for Energy Data and ML Models - Web Development

Research method: Prototyping

Research question:

How can we implement the web interface of a tool to serve energy researchers with data and code for a specific use case?

Possible approach:

• Research on platform concepts in other research fields (e.g., How did Hugging Face evolve?)
• Implementation of required components of the platform (backend and/or frontend)
• Testing and evaluation of its performance

Your background/interests:

• Passion for programming
• Interest in web-development, ideally some background in one or more of the following: Python, JavaScript, TypeScript, PostgreSQL
• Curiosity about how energy research infrastructure can be set up efficiently

For more info see here!

Annika Schneider

annikakristina.schneider@tum.de 

Project EDL: Building a Digital Platform for Energy Data and ML Models - Data Curation

Research method: Prototyping

Research question:

How can we curate energy data such that researchers can quickly understand and use them? How can we automate this process?

Possible approach:

• Research for open-source, high-quality energy data sets
• Implementing of Jupyter Notebooks to document and visualize the data
• Evaluation of the helpfulness of the curation process
• Developing of ideas of automating this process

Your background/interests:

• Passion for data
• Interest in data analysis, ideally some background in Python
• Curiosity about how energy research can be accelerated

For more info see here!

Annika Schneider

a.k.schneider@tum.de 

Project EDL: Building a Digital Platform for Energy Data and ML Models - Code & Tool Curation

Research method: Prototyping

Research question:

How can we curate energy model code and open-source tools such that researchers can quickly understand and use them? How can we automate this process?

Possible approach:

• Research for open-source, high-quality energy model implementations and tools
• Implementing of Jupyter Notebooks to document the usage of the code
• Evaluation of the helpfulness of the curation process
• Developing of ideas of automating this process

Your background/interests:

• Passion for playing around with code and tools
• Interest in energy system models, ideally some background in Python
• Curiosity about how energy research can be accelerated

For more info see here!

Annika Schneider

a.k.schneider@tum.de 

Project EDL: Building a Digital Platform for Energy Data and ML Models - Database Infrastructure

Research method: Prototyping

Research question:

How can we store diverse energy data sets in an efficient and secure database infrastructure?

Possible approach:

• Developing of pipelines importing data in a database in a sensible, accesible format
• Implementing of Jupyter Notebooks to showcase how data can be requested from the database
• Evaluation of the effiency and security of this process

Your background/interests:

• Passion for data
• Interest in databases, ideally some background in databases (e.g. MongoDB)
• Curiosity about how energy research can be accelerated

For more info see here!

Annika Schneider

a.k.schneider@tum.de 

Development of Smart Grid Simulator (Project EDGE)

Research method: Prototyping

Research questions:

• How can models of energy system components be containerized and deployed on a compute cluster?
• How can performance, scalability, and reliability of this simulation infrastructure be measured?
• How can basic EMS scenarios be simulated and analyzed using this infrastructure? 

Possible approach:

• Setup local container development and cluster environment (Docker, Kubernetes, etc.)

• Develop several energy system components as Docker containers exposing web-based interfaces (e.g., distribution grid, HEMS, HH energy demand simulator, solar PV simulator, battery charge controller, etc.)
• Implement message-based communication between containers running on cluster
• Measure performance in different scenarios  

Resources:

• Docker
• Kubernetes
• lemlab
• RabbitMQ

Christoph Goebel

christoph.goebel@tum.de

Sebastian Eichhorn

sebastian.eichhorn@tum.de

Elgin Kollnig 

elgin.kollnig@tum.de

Deep Learning based Estimation of Distribution Grid State

Research method: Prototyping

Research question:

• How can state of distribution grids be estimated using deep learning?

Possible approach: 

• Understand state of the art based on literature
• Develop efficient Python code applying deep learning to estimate distribution grid states
• Benchmark with state of the art methods

Christoph Goebel

christoph.goebel@tum.de

Project EDGE: Benchmarking of Monolithic Simulation Tools against Cluster Based Simulations

Research method: Prototyping

Research question:

• In which scenarios is a cluster based simulation faster than monolithic simulation tools?
• How does it differ between time series simulations and single state simulations?
• How can those differences be quantified?

Possible approach: 

• Understand EDGE cluster concept and other monolithic tools
• Define Scenarios and Parameters for which to investigate
• Benchmark with state of the art methods

Sebastian Eichhorn

sebastian.eichhorn@tum.de

Project EDGE: Investigation of Scalable Methods for Load Flow Calculations

Research method: Prototyping

Research question:

• How can large power grids (distribution grids, optionally connected by transmission grid) be split up in smaller part grids without losing precision in load flow and transient simulations?
• For which scenarios does this make sense, trading off the increased parallelization of calculation with increased overhead
• How can those differences be quantified?

Possible approach: 

• Understand EDGE cluster concept and other monolithic approaches
• Define example power grids and investigate strategies of power grid calculation parallelization
• Setup prototype for example power grid, ensuring precision while clustering the simulation
• Benchmark with state of the art methods

Sebastian Eichhorn

sebastian.eichhorn@tum.de