Description

  • Participants of the course are allocated individual research topics to work over the seminar course. Group works are possible.
  • A kickoff session will be conducted to introduce the course and topics. Afterwards, students can apply for topics via sending the top three topics and the current transcript of records to a.k.schneider@tum.de.
  • Regular follow-up meetings with the participants and supervisor are held to aid the participants with the seminar tasks.

The seminar tries to mimic the event flow of a conference publication. Therefore, the deliverables are the following:

  • a scientific paper (submitted as the seminar report).
  • a conference-style final presentation.

More details on ECTS, learning outcomes, etc is offered at the kickoff session and later on moodle.

Process and Time Line (Sommer semester 2024)

(Prospected) Registration Period: 01.03.24 - 19.04.24

Kickoff Meeting: 17.04.24, 3pm-4:30pm, room N3815.

Due date to choose top 3 topics: 19.04.24, 12 noon.

Topic allocation: 22.04.24, 7pm the latest.

Proposal Presentations: 08. & 15. of May, 2024, 3pm-4:30pm, N3815.

Final Presentations: 10. & 17. of July, 2024, 3pm-4:30pm, N3815.

Available Topics

If you are interested in working on a topic please feel free to contact the related colleague. 

AcronymTopic descriptionRequired skillsContact
SPATIO_TEMPORAL

Does Location Matter?

Research method: Literature Research

Wind power forecasting is a vast research field, where hundreds of physical, statistical and AI methods were developed over the last decades. Some methods are based on temporal information only, while other also take the location of wind power sites into account.

Research question:

  • Does spatial information actually add a benefit to forecasting performance?
  • What are other advantages and disadvantages to add spatial information to a forecasting model?

Possible approach

  • Thorough literature research on the best-performing (1) temporal and (2) spatio-temporal techniques for wind power forecasting.
  • Provide an overview and analysis of their performance, advantages and disadvantages.
  • Try to draw a global conclusion of whether location matters.
  • Passion for structured working
  • Interested in time series forecasting

Annika Schneider
annikakristina.schneider@tum.de


FLOW_NRG

Flow-Based Generators in the energy context

Research method: Literature Review

Research question:

  • For which tasks in the energy context can flow-based generators be applied?
  • How can the flow-based generator architecture be exploited/adopted for these specific tasks?

Possible approach

  • Learn about flow-based generators and identify their strengths, weaknesses and typical applications
  • Find existing studies of flow-based generators in the energy context
  • Identify appropriate tasks in the energy context for these networks
  • Identify a way to use/adopt these networks for these specific tasks


  • Interest in modern machine learning methods

Annika Schneider

annikakristina.schneider@tum.de


GFM

Global Forecasting Methods for Joint Wind Power Forecasts

Research method: Modeling

Research question

  • How can Global Forecasting Methods (e.g., lightGBM) be leveraged for wind power predictions of nearby wind power plants?

Possible approach

  • Identification of approved global forecasting methods
  • Identification of suitable data sets
  • Apply method to data sets
  • Evaluate results
  • Programming skills in Python
  • Interest in time series data and forecasting

Annika Schneider

annikakristina.schneider@tum.de


DEEP_NILM

Deep Learning for Energy Disaggregation

Research method: Prototyping

Research question:

  • How can Deep Learning be applied to Energy Disaggregation / NILM?
  • How does Deep Learning perform compared to other methods?

Possible approach

  • Understand state-of-the-art based on literature
  • Develop efficient Python code to train DL models on public NILM datasets
  • Evaluate performance by benchmarking with methods described in literature


  • Python programming skills

Christoph Goebel

christoph.goebel@tum.de

DEEP_HEAT

Deep Reinforcement Learning for HVAC Control

Research method: Prototyping

Research question:

  • How can Deep Reinforcement Learning be applied to HVAC Control problems?
  • How does Deep Reinforcement Learning perform compared to other methods?

Possible approach

  • Understand state-of-the-art based on literature
  • Develop efficient Python code to train DRL models using HVAC model
  • Evaluate performance by benchmarking with methods described in literature


  • Python programming skills

Christoph Goebel

christoph.goebel@tum.de

ES_EVAL

Techno-Economic Evaluation of Combined Energy Storage Technologies in Neighborhoods

Research method: Data Analysis & Modeling

Research question:

  • How can different types of energy storage be combined in a neighborhood setting?
  • How can different scenarios be compared technically and economically?

Possible approach:

  • Understand state-of-the-art based on literature
  • Develop models to understand technical limitations and economics of energy storage
  • Define scenarios and investigate potential of different types of energy storage and their combinations


  • Python programming skills
  • Background in energy system modelling

Christoph Goebel

christoph.goebel@tum.de

DEEP_WIND

Deep Learning for Wind Power Prediction

Research method: Prototyping

Research question:

  • How can Deep Learning be applied to Wind Power Forecasting?
  • How does Deep Learning perform compared to other methods?

Possible approach

  • Understand state of the art based on literature
  • Develop efficient Python code to train DL models on public datasets
  • Evaluate performance by benchmarking with methods described in literature

  • Python programming skills
  • First experience using deep learning

Christoph Goebel

christoph.goebel@tum.de

LEM_GRID_CONSTRAINTS

Impact of Local Energy Markets on Distribution Grid Stability 

Research method: Prototyping

Research question:

  • How can the impact of Local Energy Markets on distribution grid stability be measured?
  • Which mechanisms could be introduced to prevent stability issues?

Possible approach

  • Understand state of the art based on literature
  • Develop efficient Python code to simulate simple LEM scenarios together with distribution grid effects
  • Develop basic mechanisms to prevent stability issues and evaluate using LEM/DG simulation

  • Python programming skills

Christoph Goebel

christoph.goebel@tum.de

DL_DG_ESTIMATION

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 applying deep learning to estimate distribution grid states
  • Benchmark with state of the art methods


  • Python programming skills
  • First experience using deep learning

Christoph Goebel

christoph.goebel@tum.de

DYN_BAT_AGE_PRED_CONTROL

Integration of a dynamic aging prediction for Li-ion batteries in EMS control

Research method: prototyping

Research question:

  • How can the aging process of Li-ion batteries be modeled for use in control?
  • How can such models be integrated into dynamic controllers?

Possible approach:

  • Understand the state of research based on literature
  • Development of an efficient Python program that uses a battery aging models for control
  • Benchmarking of the developed method


  • Python programming skills

Christoph Goebel

christoph.goebel@tum.de

BAT_AGE_PRED

Machine learning-based prediction of battery aging

Research method: prototyping

Research question:

  • How to predict lithium-ion battery aging using ML models?
  • How performant are ML models compared to other methods?

Possible approach:

  • Understand state-of-the-art based on literature
  • Develop an efficient Python program that fits various ML models on battery aging data
  • Benchmarking of the developed models
  • Python programming skills

Christoph Goebel

christoph.goebel@tum.de

RC_MODELS

R-C Models for Heat Transfer in Buildings

Research method: Prototyping

Research question:

  • How to design an easy-to-use Python library for fitting R-C-models of the heat transfer in buildings?
  • How to use fitted models to simulate buildings in control environments?

Possible approach

  • Understand state-of-the-art based on literature
  • Develop efficient Python code to fit R-C models
  • Evaluate implementation, e.g., by benchmarking results and evaluating computational complexity
  • Python programming skills

Christoph Goebel

christoph.goebel@tum.de

RC_MODEL_FITTING

Using EnergyPlus to fit R-C Models

Research method: Prototyping

Research question

  • How can EnergyPlus models be approximated by R-C models?
  • How effective and efficient is the approach?   

Possible approach

  • Learn to work with EnergyPlus
  • Develop efficient Python code fit R-C models to output data
  • Benchmark method
  • Python programming skills

Christoph Goebel

christoph.goebel@tum.de

 PV_AGING

Age-Conscious System Efficiency Modeling of a Photovoltaic System

Research method: Prototyping

Research question:

  • How can the efficiency of a PV system on the institute's roof be modeled based on measured data and weather conditions?
  • What conclusions about the age of the system can be drawn based on this modeling?

Possible approach

  • Literature review on modeling efficiencies of real PV systems, with a focus on aging.
  • Calculation of efficiency trends based on measured time series and logged weather data from the region.
  • Analysis of efficiency trends with regard to system aging and nominal data of the components.
Basic Python programming skills

Sebastian Eichhorn
sebastian.eichhorn@tum.de

LAB_PSIM

Analysis of Stability Issues in Load Simulators

Research method: Literature Research

Research question

  • Are stability issues in load simulators a concern for power laboratories?
  • What methods exist for reducing the impact of these issues

Possible approach

  • Literature review on various load simulation methods in Power-Hardware-in-the-Loop (PHIL) laboratories.

  • Literature review on known instabilities in the technologies used OR analysis of the researched components for potential instability using minimal models.

  • Evaluation of the results.

Profound understanding of system stability concepts and power electronic systemsSebastian Eichhorn
sebastian.eichhorn@tum.de

FMI_RT

Real-time Simulation of Functional Mockup Units

Research method: Prototyping

Research question

  • How can Functional Mockup Units (FMU) be best integrated into a real-time simulation framework (HELICS)?
  • How can a scalable simulation environment be easily set up?

Possible approach

  • Research on FMU and HELICS as well as real-time simulation of energy systems

  • Definition of the programming language to be used and setup of the interface between FMU (which is available) and HELICS, as well as the scalable HELICS setup

  • Functional test and execution of exemplary real-time simulation

 

Basic Python or Julia programming skillsSebastian Eichhorn
sebastian.eichhorn@tum.de

HH_Control

Controllable Household Models

Research method: Prototyping

Research question

  • How can households, and in particular the control options that an energy management system (EMS) has, be modeled?
  • How can various intervention options be modeled within a model without needing to adapt the model for a new application?
  • How an the model be structured to incorporate individual detailed models later on(e.g. heat pump models)?

Possible approach

  • Research on smart households and intervention options of EMS

  • Development of a control concept using real-life EMS as an example

  • Programming of the operational model and representation of standard scenarios

 

Basic Python or Julia programming skillsSebastian Eichhorn
sebastian.eichhorn@tum.de

EMS_COMP

Comparison of Open-Source Energy Management System (EMS)

Research method: Literature Research

Research question

  • What are the (new) open-source Energy Management System (EMS) frameworks available?
  • How do they differ, and which frameworks are suitable for specific use cases?

Possible approach

  • Literature review on various EMS frameworks.

  • Analysis of the different advantages and disadvantages of the systems.

  • Compilation of a concise guide indicating which system is suitable for each specific application.

 

Profound understanding of energy management systemsSebastian Eichhorn
sebastian.eichhorn@tum.de

DIST_LEMLAB

Distributed Implementation of lemlab

Background: lemlab is a simulation tool for local energy markets, code if open-source 

Research method: Prototyping

Research question

  • How can lemlab be implemented in a scalable way using state of the art cluster technology?

Possible approach

  • Implement selected lemlab logic in Docker containers

  • Design and implement message-based communication between containers
  • Deploy containers on Kubernetes cluster (can be run locally)  

Good Python programming skills, knowledge of Docker and ideally Kubernetes

Christoph Goebel

christoph.goebel@tum.de


COMP_HH_ENERGY_SIM

Containerization and Comparison of Household Energy Demand Simulation Tools

Research method: Prototyping

Research question

  • How can different tools for household energy demand simulation be containerized for further use in distributed environments?
  • How can they be compared in terms of accuracy and computational performance?

Possible approach

  • Review different tools for household energy simulation
  • Implement selected tools as Docker containers with standardized interfaces

  • Write code for comparing accuracy and computational performance 
Python programming skills, knowledge of Docker

Christoph Goebel

christoph.goebel@tum.de

DATA_PLATFORM

Data Platforms an their Way to Success

Research method: Literature Research

In the energy domain, there are several open data platforms, but none of it seems to be highly successful and actually attract a large user base. In other domains, platforms like "Hugging Face" celebrate huge success and truly accelerate research in their field.

Research question:

  • Which data/code platforms are the most successful? 
  • How did these platforms evolve?
  • How did they become succesful? 

Possible approach

  • Thorough literature research on the most successful platforms and their stories
  • Identification of "platform success strategies"
  • Curiosity about platform success storiess
  • Structured way of working

Annika Schneider

a.k.schneider@tum.de 

DATA_CURATION

The Curation Gold Standard

Research Method: Literature Review & Prototyping

Research Question:

  • What open-source data curation tools are available?
  • How can these be integrated into existing software?
  • How can the performance of these tools be evaluated?
  • How effective are these tools?

Possible Approach:

  • Thorough literature review on open-source data curation tools
  • Application and evaluation of the tools on open-source energy data

 

  • Passion for data
  • Structured way of working

Annika Schneider

a.k.schneider@tum.de 

LDM

TUM-EMT x Leibniz - Will the Leibniz Data Manager revolutionize our Research Data Management?

Research Method: Literature Review & Prototyping

Research Question:

  • What are the features of the Leibniz Data Manager?
  • What are its limitations?
  • How can the potential of the Leibniz Data Manager be measured and evaluated?

Possible Approach:

  • Literature review on research data management
  • Testing and evaluating the Leibniz Data Manager
  • Passion for data
  • Structured way of working

Annika Schneider

a.k.schneider@tum.de 

ML_Powerflow

Machine Learning based Load Flow Calculation

Research method:Literature Review

Research question

  • Which innovative solutions evolved recently to calculate load flow calculations using Machine Learning?

  • How can those solutions be compared, which strengths and weaknesses do they have?

Possible approach

  • Research on machine learning technologies for load flow calculation, including most recent advancements in Graph Neural Networks (GNN)

  • Identify strengths and weaknesses of different technologies

  • Compare them focusing on precision and computational efficiency

 

Passion for literature analysis and power system calculationSebastian Eichhorn
sebastian.eichhorn@tum.de

 PINN_Sim

Physics Informed Neural Networks (PINN) for Transient Power Grid Simulations

Research method: Literature Review/Prototyping

Research question

  • Can PINN be an option for future transient power grid simulations?

  • What are strengths and weaknesses compared to conventional solvers?

Possible approach

  • Research on PINNs and publications regarding challenges in transient simulations

  • Setup a little prototype using public repositories provided through publications

  • Compare the prototype with conventional solvers focusing on precision and computational efficiency

 


Profound understanding of Machine Learning Technologies, Power System Simulation and advanced Python codingSebastian Eichhorn
sebastian.eichhorn@tum.de

EMS_COMM

Evaluating the Next-Generation of EMS Communication

Research method: Literature Review 

Research question:

  • How do new communication methods such as EEBUS and Matter enable new opportunities for Energy Management System (EMS) and energy device connectivity?

  • How do these methods compare with one another? How are they built, what is the state of adoption, and what are the limitations?
  • Ultimately, is one clearly better than the other?

Possible approach:

  • Literature review on the issues related to older communication standards for EMS (Modbus, MQTT, HTTP, OCPP for electric vehicles, etc)
  • Literature review on the EEBUS and Matter standards
  • Provide summaries of their design, benefits, limitations, and use cases
  • Research products and organizations that already use each method
  • Develop guidelines for when to use which method


Interest in Energy Management Systems, IoT, or Distributed Energy Resources

Ehimare Okoyomon
e.okoyomon@tum.de

COMP_ENERGY

Evaluating the Energy Impact of Computation for Machine Learning

Research method: Literature Review / Prototyping

Research question:

  • How is energy estimation of computation done in research and how do the methods compare?

  • For ML applications, does faster training always mean more energy efficient, or are there other model or pipeline design considerations?
  • What are the popular libraries for measuring energy impact of computation for python and what do they provide?
  • (Stretch) Using one of these libraries, can we benchmark and compare the energy impact of one or more ML models in the EMT context (ex. solar or wind power prediction)? 

Possible approach:

  • Read and review papers on the environmental impacts of Machine Learning

  • Provide summaries of methods and develop guidelines on the types of ML model design, configurations, parameters, or algorithms that lead to lower energy consumption

  • Look for popular power estimation libraries from the literature and on GitHub
  • (Stretch) Use one of these libraries to benchmark state of the art ML techniques and document the results

Interest in Computational Energy

Interest in ML and computer architectures 

Basic Python skills

Ehimare Okoyomon
e.okoyomon@tum.de

  • No labels
Write a comment...