SA/BA/MA projects

Have you heard of industry 4.0, smart grids, smart buildings, smart cities, intelligent traffic systems and intelligent self-driving cars? Did you ever wonder what makes them smart and intelligent?

The answer is control and automation and that’s what we do at the Automatic Control Laboratory (IfA). We use control theory, optimization, machine learning, and game theory to develop controllers and algorithms that are the backbone of nearly all modern technology. At our lab we span the whole area from pure theory to real-world applications and we are looking for you to help us push forward the state of the art. If you want to learn techniques and gain knowledge that enable you to work in any field from medical applications to spacecraft and from electrical grids to finance then the Automatic Control Laboratory is the place for you!

Lists of currently running and recently completed projects can be found in the sub-navigation menu above. Reports for several projects are available through the ETHZ Research Collection:

Open projects

Runtime Monitoring with Formal Specification-Guided LLMs

Autonomous systems increasingly rely on complex software stacks and data-driven components whose behavior is difficult to fully verify before deployment. Runtime verification provides a lightweight mechanism for ensuring that a system execution satisfies formally specified properties \cite{lindemann2023conformal,bauer2011runtime,lukina2021into}. Classical runtime monitors are typically symbolic and algorithmically constructed from temporal logic specifications. These properties have to be specified a-priori by a domain expert, posing a practical bottleneck. This thesis investigates a novel paradigm in which the runtime monitor itself is implemented as a Large Language Model (LLM) so that the system specification can be provided via a natural language interface. While this has the advantage of not requiring expert knowledge and being able to change specifications on-the-fly, it is unclear how reliable such a monitoring approach would be, which necessitates additional formal structure on the problem formulation and implementation.

Labels

Master Thesis

More information

Open this project...

Published since: 2026-03-09 , Earliest start: 2026-03-10 , Latest end: 2026-12-31

Organization Automatic Control Laboratory

Hosts Lindemann Lars , Balta Efe

Topics Engineering and Technology

De-bugging and Tuning of Learning Based Controllers

This project focuses on developing an autonomous debugging and tuning system for self-commissioning controllers in HVAC applications. These controllers automatically learn system dynamics and configure PI gains, but their performance depends on precise hyperparameter settings and adaptive tuning to address issues like misconfiguration or changing conditions. The goal is to create a program that analyzes controller behavior in real time, detects performance issues, and autonomously adjusts parameters for optimal operation. Key tasks include identifying critical hyperparameters, defining performance metrics, and designing a robust tuning algorithm. Numerical simulations will validate the approach across diverse scenarios, aiming for true plug-and-play functionality—enabling controllers to self-monitor and adapt like an expert engineer. Ideal for students with a background in control systems or automation, the project offers hands-on experience in adaptive control, system identification, and smart building technologies. Proficiency in MATLAB/Simulink and basic machine learning knowledge are beneficial. The project will be co-supervised with Belimo Automation AG.

Keywords

Self-learning controllers, Control systems, PI control, Hyperparameters, System identification, Plug-and-play control, Debugging, Tuning, Smart buildings, HVAC, Adaptive control, Simulation and validation.

Labels

Master Thesis , Theory (IfA) , Computation (IfA) , Energy (IfA) , Applications (IfA)

Description

Self-commissioning controllers for simple control devices regulate a single output variable using one actuator. When first connected, the controller learns the system’s dynamic behaviour and automatically designs suitable proportional–integral (PI) gains to ensure smooth and stable performance. For this learning-based approach to work effectively, key hyperparameters of the underlying learning and identification algorithms must be correctly selected during commissioning. In practice, these hyperparameters strongly influence convergence speed, robustness, and closed-loop performance. Misconfigured parameters, overly aggressive or slow responses, or changing operating conditions can therefore require systematic analysis and re-tuning. The goal of this project is to develop a program that can analyse a controller’s behaviour, identify the causes of undesired performance, and automatically adjust its hyperparameters to improve it. In particular, learning-based and system identification methods will be used to support structured hyperparameter identification and performance assessment. The system should reliably diagnose performance issues, tune hyperparameters of the learning algorithm, and ensure robust adaptation in real-world applications. Ultimately, the aim is to move closer to true plug-and-play functionality by enabling the controller to autonomously monitor its behaviour and take corrective actions. This project will be co-supervised with Belimo Automation AG. In addition to simulation-based development, the work will include hands-on experimentation in an industrial setting using real HVAC devices and embedded controllers. The student will gain practical experience with real hardware, data acquisition, controller implementation, and performance evaluation under realistic operating conditions. This provides direct exposure to industrial product development and validation processes. We are looking for a talented, motivated student interested in self-learning control systems and smart building technologies and HVAC systems with • Background/interest in control systems, automation, or mechatronics engineering. • Familiarity with PI control, system identification, and dynamic modelling. • Basic knowledge in data-based and machine learning methods. • Experience with MATLAB/Simulink or a similar simulation environment. • Strong analytical and problem-solving skills.

Goal

Contact Details

More information

Open this project...

Published since: 2026-03-02 , Earliest start: 2026-03-01 , Latest end: 2026-08-31

Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne , Empa , Eawag , Lucerne University of Applied Sciences and Arts , University of Basel , University of Berne , University of Fribourg , University of Geneva , University of Lausanne , University of Lucerne , University of St. Gallen , University of Zurich , Zurich University of Applied Sciences

Organization Automatic Control Laboratory

Hosts Balta Efe

Topics Mathematical Sciences , Information, Computing and Communication Sciences , Engineering and Technology

No-Regret Zero-Shot Meta Reinforcement Learning for Control-Affine Systems

Offline reinforcement learning typically assumes access to a simulator capable of generating training tasks from a fixed distribution. However, despite potentially broad training coverage, real-world deployment often presents environments that are unique and never encountered during training. This mismatch is especially pronounced when policies trained offline in simulation are deployed on physical systems, where online episodic training is impractical or unsafe. In such scenarios, agents must adapt rapidly to a new environment using a single trajectory, operating in a zero-shot or one-shot setting. This thesis aims to develop no-regret algorithms for zero-shot meta reinforcement learning using a grey- box modeling approach. We assume partial knowledge of the system dynamics—specifically, the functional structure of the model—while treating the system parameters as unknown. This abstraction enables principled adaptation at test time and facilitates the derivation of rigorous performance guarantees. Crucially, it also allows the incorporation of safety and stability constraints, which are essential for deployment in cyber-physical systems. The focus of the project is on control-affine systems, which arise frequently in practical applications. The thesis investigates how bilinear and control-affine quadratic control problems, subject to additive non- stochastic disturbances, can be formulated within a zero-shot reinforcement learning or online optimization framework. Building on this formulation, the goal is to derive state-of-the-art online controllers that achieve low dynamic regret while ensuring robustness and stability. The proposed methods will be evaluated both through simulation studies and rigorous theoretical analysis, with guarantees in terms of bounded-input bounded-state (BIBS) stability and dynamic regret.

Keywords

Reinforcement Learning, Meta Learning, Online Optimization, Artificial Intelligence, Fine Tuning, Adaptive Control

Labels

Master Thesis

Description

Meta Reinforcement Learning (meta RL) studies machine learning algorithms that learn to reinforcement learn [1]. Rather than learning from scratch on a given task through a large number of repeated episodes, agents in meta RL first perform meta-learning on meta-learning data (outer loop) which is a collection of tasks/environmental realisations sampled from a given distribution. Then, during test-time or online play time, the algorithms fine-tune or adapt to the particular task at hand (inner loop) and ideally do so fast through a small number of episodes (few-shot learning), as visualized in the Figure [1]. Such a setting is especially relevant for robotics applications such as legged robotics [2] or quadrotors[3], where the meta-learning happens in simulation due to safety and cost considerations, and the adaptation to the specific instantiation of the environment happens during test-time or online. Moreover, for many such applications there is no availability to do more than a single episode for fine-tuning, and the agent needs to adapt to control online, during a single trajectory, aka in a zero-shot setting. RL2 [4] and VariBAD [5] are among the algorithms that operate in this zero-shot meta RL setting. In addition, there is significant literature on online learning methods for control that study the zero-shot RL setting often assuming no offline training availability[6, 7]. In contrast to meta-RL, these methods concentrate on linear dynamical systems and provide rigorous theoretical guarantees on stability and online reward performance in the form of regret. This thesis concentrates on the zero-shot meta RL setting and aims to bridge the gap between data-driven meta RL methods and theoretically grounded online/adaptive control. It leverages the theoretical foundations in online learning [8] and adaptive control theory [9], as well as the advancements in meta RL [10, 1, 4] to introduce no-regret (sublinear) algorithms for control-affine dynamical systems. Such systems arise frequently in practice and capture mild nonlinearities through a control-affine structure, where the system dynamics depend linearly on the control input and affinely on the state through a known functional form. A gray-box modeling approach is adopted: the structural form of the system dynamics is assumed known, while key parameters—including the system matrices and additive non-stochastic disturbances—remain unknown. This modeling choice balances expressiveness and tractability, enabling both practical applicability and rigorous theoretical analysis. In addition to reinforcement learning and online learning methods [11, 12], the thesis will draw on optimal control results for control-affine systems [13, 14]. The thesis will explore two fundamental directions of synthesizing RL algorithms: • Model-free policy gradient methods, where the control policy is updated directly online without explicitly identifying a system model. An example of this approach is given in [7], where an affine policy is adapted online via gradient- based updates. • Model-based indirect methods, where the system dynamics and uncertainty are first learned online, and the control policy is subsequently derived using a certainty-equivalence principle. A representative example is the PLOT algorithm [6], which employs a receding horizon controller based on recursively updated model estimates. **By unifying ideas from meta RL, online optimization, and adaptive control, this thesis aims to develop policies that adapt online in a zero-shot setting while providing strong theoretical guarantees on stability and regret.** [1] J. Beck, R. Vuorio, E. Zheran Liu, Z. Xiong, L. Zintgraf, C. Finn, and S. Whiteson, “A survey of meta-reinforcement learning,” arXiv e-prints, pp. arXiv–2301, 2023. [2] A. Belmonte-Baeza, J. Lee, G. Valsecchi, and M. Hutter, “Meta reinforcement learning for optimal design of legged robots,” IEEE Robotics and Automation Letters, vol. 7, no. 4, pp. 12134–12141, 2022. [3] S. Belkhale, R. Li, G. Kahn, R. McAllister, R. Calandra, and S. Levine, “Model-based meta-reinforcement learning for flight with suspended payloads,” IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 1471–1478, 2021. [4] Y. Duan, J. Schulman, X. Chen, P. L. Bartlett, I. Sutskever, and P. Abbeel, “Rl2: Fast reinforcement learning via slow reinforcement learning,” arXiv preprint arXiv:1611.02779, 2016. [5] L. Zintgraf, K. Shiarlis, M. Igl, S. Schulze, Y. Gal, K. Hofmann, and S. Whiteson, “Varibad: A very good method for bayes-adaptive deep rl via meta-learning,” arXiv preprint arXiv:1910.08348, 2019. [6] A. Tsiamis, A. Karapetyan, Y. Li, E. C. Balta, and J. Lygeros, “Predictive linear online tracking for unknown targets,” in International Conference on Machine Learning, pp. 48657–48694, PMLR, 2024. [7] A. Karapetyan, D. Bolliger, A. Tsiamis, E. C. Balta, and J. Lygeros, “Online linear quadratic tracking with regret guarantees,” IEEE Control Systems Letters, vol. 7, pp. 3950–3955, 2023. [8] N. Cesa-Bianchi and G. Lugosi, Prediction, learning, and games. Cambridge university press, 2006. [9] A. M. Annaswamy, “Adaptive control and intersections with reinforcement learning,” Annual Review of Control, Robotics, and Autonomous Systems, vol. 6, no. 1, pp. 65–93, 2023. [10] C. Finn, P. Abbeel, and S. Levine, “Model-agnostic meta-learning for fast adaptation of deep networks,” in Inter- national conference on machine learning, pp. 1126–1135, PMLR, 2017. [11] M. Simchowitz and D. Foster, “Naive exploration is optimal for online lqr,” in International Conference on Machine Learning, pp. 8937–8948, PMLR, 2020. [12] A. Wagenmaker and K. Jamieson, “Active learning for identification of linear dynamical systems,” in Conference on Learning Theory, pp. 3487–3582, PMLR, 2020. [13] Z. Aganovic and Z. Gajic, “The successive approximation procedure for finite-time optimal control of bilinear systems,” IEEE Transactions on Automatic Control, vol. 39, no. 9, pp. 1932–1935, 2002. [14] Z. Yuan and J. Cort´es, “Data-driven optimal control of bilinear systems,” IEEE Control Systems Letters, vol. 6, pp. 2479–2484, 2022

Goal

Contact Details

More information

Open this project...

Published since: 2026-02-23 , Earliest start: 2026-03-01 , Latest end: 2026-10-01

Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne

Organization Automatic Control Laboratory

Hosts Karapetyan Aren

Topics Mathematical Sciences , Information, Computing and Communication Sciences

Robust Feedback Control for Robotic Carton Folding

This project aims to develop control strategies for robotic carton folding that explicitly regulate the evolution of the carton state during manipulation. Controlling articulated and partially deformable objects remains an open problem in robotic manipulation, particularly in contact-rich tasks such as carton folding, where success depends on precise coordination of motion, force, and compliance. The overall objective is to investigate control formulations that enable robust, generalizable folding behaviors beyond open-loop or purely trajectory-based execution.

Keywords

Robotics, Force Control, Vision, Hardware

Labels

Semester Project , Master Thesis

Description

Object manipulation with contact-rich interations involves switching between contact and no-contact modes. Open-loop execution of stiff position control fails under these hybrid dynamics by ignoring dynamic coupling, by being too slow to solve the combinatorial planning of optimal motion sequences, and by being inaccurate under occlusion and physical uncertainties. Thus, the motivation for robust, fast feedback controllers, ideally which address the impracticality of modeling changing environments. Several classes of feedback-based control strategies have been proposed. Compliance-based strategies, such as impedance control, avoid crashing by adjusting control magnitude based on resistance. Extending the idea, variable impedance control adapts online based on safety and dexterity (precision) requirements. Recent work [1] learns a general representation of variable impedance expert demonstrations through inverse reinforcement learning (IRL) to generate new policies with policy gradient reinforcement learning (RL), while [2] uses probabilistic ensembles as a variable impedance model within model predictive control (MPC). Beyond compliance based strategies, hybrid force-motion control explicitly jointly satisfies constraints needed to maintain contact during manipulation. [3] plans over a predefined skill set to specify force and velocity targets, which are then tracked using a low-level torque controller inconveniently requiring careful parameter tuning. [4] instead learn low-level force control parameters with RL, integrating RL into both hybrid force position control and admittance controllers. Optimization-based formulations such as MPC-based approaches offer the advantage of incorporating look-ahead planning over contact modes. [5] trains with supervised learning a classifier to select contact mode sequences offline, then solves a fast reactive controller online. [6] devises a mode-invariant, contact-implicit MPC formulation using linear complementarity contact formulation and bilevel trajectory planning. [1] X. Zhang, L. Sun, Z. Kuang, and M. Tomizuka, “Learning variable impedance control via inverse reinforcement learning for force-related tasks,” IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 2225–2232, 2021. [2] A. S. Anand, J. T. Gravdahl, and F. J. Abu-Dakka, “Model-based variable impedance learning control for robotic manipulation,” Robotics and Autonomous Systems, vol. 170, p. 104531, 2023. [3] J. Liang, X. Cheng, and O. Kroemer, “Learning preconditions of hybrid force-velocity controllers for contact-rich manipulation,” in Proceedings of The 6th Conference on Robot Learning (K. Liu, D. Kulic, and J. Ichnowski, eds.), vol. 205 of Proceedings of Machine Learning Research, pp. 679–689, PMLR, 14–18 Dec 2023. [4] C. C. Beltran-Hernandez, D. Petit, I. G. Ramirez-Alpizar, T. Nishi, S. Kikuchi, T. Matsubara, and K. Harada, “Learning force control for contact-rich manipulation tasks with rigid position- controlled robots,” IEEE Robotics and Automation Letters, vol. 5, no. 4, pp. 5709–5716, 2020. [5] F. R. Hogan, E. R. Grau, and A. Rodriguez, “Reactive planar manipulation with convex hybrid mpc,” in 2018 IEEE International Conference on Robotics and Automation (ICRA), p. 247–253, IEEE Press, 2018. [6] S. Le Cleac’h, T. A. Howell, S. Yang, C.-Y. Lee, J. Zhang, A. Bishop, M. Schwager, and Z. Manch- ester, “Fast contact-implicit model predictive control,” IEEE Transactions on Robotics, vol. 40, pp. 1617–1629, 2024.

Goal

Contact Details

More information

Open this project...

Published since: 2026-02-19 , Earliest start: 2026-02-23 , Latest end: 2027-02-23

Organization Automatic Control Laboratory

Hosts Stocco Paula

Topics Information, Computing and Communication Sciences , Engineering and Technology

Multi-Agent Grid Impedance Identification in Three-Phase Power Systems

This project investigates multi-agent grid impedance identification in three-phase power systems. It will develop and compare two approaches: a global multi-port identification framework and a locally simultaneous multi-agent single-port identification framework. The study will evaluate and compare the accuracy of the two approaches and explore potential downstream applications in stability analysis and control design.

Keywords

grid impedance, system identification, multi-agent systems, three-phase power system dynamic modelling

Labels

Master Thesis

Description

Goal

Contact Details

More information

Open this project...

Published since: 2025-12-31 , Earliest start: 2026-01-01

Organization Automatic Control Laboratory

Hosts Häberle Verena

Topics Engineering and Technology

Safe Feedback Optimization for Power Grids via Switched Controllers

Feedback optimization is emerging as an important control method for modern power systems, thanks to its robustness and ability to steer the grid to an efficient operating point. Clearly, power systems are safety-critical infrastructures: failures can cause severe consequences, such as blackouts. In this project, we will design and evaluate novel feedback optimization schemes, based on switched systems, which can guarantee safety of the grid at all times.

Keywords

Computational control, feedback optimization, switched systems, smart grid, projected dynamics

Labels

Semester Project , Master Thesis

PLEASE LOG IN TO SEE DESCRIPTION

More information

Open this project...

Published since: 2025-12-04 , Earliest start: 2026-01-05

Applications limited to ETH Zurich , University of Zurich

Organization Automatic Control Laboratory

Hosts Bianchi Mattia

Topics Mathematical Sciences

Unsupervised anomaly detection in additive manufacturing using flows

This thesis advances in-situ process monitoring in Laser Powder Bed Fusion (PBF-LB/M) by integrating Eddy Current Testing (ECT) and computer vision (CV) for defect detection. The study focuses on signal interpretation across simple to complex geometries, using multi-sensor data fusion to improve defect identification with real world data.

Keywords

Sensor fusion, machine learning, anomaly detection, additive manufacturing, computer vision, Neural Flows

Labels

Master Thesis

Description

Goal

Contact Details

More information

Open this project...

Published since: 2025-11-18 , Earliest start: 2026-02-16 , Latest end: 2026-05-29

Organization Automatic Control Laboratory

Hosts Zakwan Muhammad

Topics Information, Computing and Communication Sciences , Engineering and Technology

AI for Grid Stability: Deep Learning the Small-Signal Dynamics of Power Converters

This project focuses on developing and validating a scalable machine learning framework to address the modeling challenges of small-signal stability assessment in grid-connected converters.

Keywords

Data-driven modelling, Machine Learning, Power Converters, Small-signal modelling, Sub-synchronous oscillations, Control, Stability

Labels

Master Thesis

Description

Modern power grids are undergoing a fundamental transformation driven by the massive integration of power-electronics-based resources (e.g., wind, solar). This high penetration profoundly changes the system's dynamic behavior, creating a complex and fast interplay among various subsystems of the grid. Consequently, accurate dynamic modeling is paramount to ensure stable and optimal grid operation. Impedance-based stability assessment is a critical tool for understanding the small-signal dynamics. However, its application faces several challenges. The frequency-domain impedance model of a converter-grid interconnection is linearized and only valid around a specific Operating Point (OP). The analytical derivation of this model is often infeasible in practice because the internal control algorithms of commercial converters are typically treated as black-box systems and are kept confidential by manufacturers. Therefore, ensuring reliable operation under a wide operation range requires repetitive identification of the impedance model at numerous OPs, which is time-consuming and computationally demanding. To overcome these practical limitations, recent advances in data-driven modeling have proven highly valuable. Machine learning models can be trained to interpolate complex impedance models obtained at only a small set of OPs to accurately predict the behavior at arbitrary OPs. This project focuses on developing and validating a scalable machine learning framework to address the modeling challenges of small-signal stability assessment in grid-connected converters. The core idea is to design and train a deep learning model to serve as a surrogate model of the converter. This surrogate model will be capable of accurately predicting the converter's frequency-domain response across a wide range of arbitrary OPs. The resulting trained models will then be used as an efficient tool to analyze and ensure stability of interconnected systems. **Prerequisites** Competitive candidates will have a background in power systems/electronics, deep learning methods, and systems \& control. A prior experience or coursework related to grid-connected converters would be highly beneficial. Proficiency in Python and MATLAB/Simulink (or similar power system simulation tools) is required to implement and test the methods. Depending on the student's interest, we can help to publish the results in a conference paper. **References** 1. Zhang, Mengfan, Yang Zhang, and Qianwen Xu. "Transfer learning based online impedance identification for modular multilevel converters." IEEE Transactions on Power Electronics 38.10 (2023): 12207-12218. 2. Liao, Yicheng, et al. "Neural network design for impedance modeling of power electronic systems based on latent features." IEEE Transactions on Neural Networks and Learning Systems 35.5 (2023): 5968-5980. 3. Wu, Yang, et al. "Impedance profile prediction for grid-connected VSCs with data-driven feature extraction." IEEE Transactions on Power Electronics (2024).

Goal

Contact Details

More information

Open this project...

Published since: 2025-11-18

Organization Automatic Control Laboratory

Hosts Abdalmoaty Mohamed

Topics Engineering and Technology

Deep reinforcement learning for job shop scheduling

Production scheduling is a critical event in the production management of smart manufacturing systems. Many job shop scheduling (JSP) problems are known as NP-hard. Meanwhile, in the real production plants, not only is the objective of scheduling usually multiple (e.g., makespan, tardiness, machine utilization), but also the environment is mostly dynamic with unexpected machine breakdowns, order arrivals and cancellations, etc. Because the JSP can be modelled as a Markov Decision Process (MDP), using deep reinforcement learning techniques for scheduling has gained significant attention in this domain. In this project, a deep reinforcement learning environment for JSP will be developed and enhanced to tackle the aforementioned challenges.

Keywords

Deep reinforcement learning, Job shop scheduling, Operational research, Multi-objective optimization

Labels

Semester Project , Master Thesis

Description

Contact Details

More information

Open this project...

Published since: 2025-11-15 , Earliest start: 2026-01-01 , Latest end: 2026-09-01

Applications limited to ETH Zurich

Organization Automatic Control Laboratory

Hosts Tang Yu, Mr.

Topics Engineering and Technology

Fair Demand-Side Flexibility Allocation with Karma

The rapid integration of distributed renewable energy sources into electric power grids introduces the challenges of balancing intermittent power production and guaranteeing grid stability. Demand-side energy flexibility, which involves the end-user shifting or adjusting energy consumption in line with the supply and capacity of the electricity grid, is receiving increasing attention as an important approach towards tackling these challenges. Buildings are significant energy consumers and hence present a promising source of flexibility. Leveraging the thermal energy storage capabilities of buildings allows to intentionally modify their energy consumption to support grid needs without compromising occupants’ thermal comfort. Previous work has focused on quantifying energy flexibility envelopes or designing temperature control strategies which are important to meet flexibility objectives. However, in practice, Distribution System Operators (DSOs) often prefer to provide flexibility targets to groups of buildings rather than to individual units. This leads to the question of how to fairly and efficiently distribute flexibility targets to individual households within a designated group. This project aims to develop a karma mechanism for this flexibility allocation problem. Karma economies are a promising recently developed non-monetary solution to the allocation of shared resources. These economies leverage the fact that shared resources, such as flexibility requests, do not occur once but are repeated frequently. It is thus envisioned that karma will enable consumers to express when they have high urgency to avoid shifting their consumption, meanwhile guaranteeing that everyone contributes fairly to providing flexibility over time.

Labels

Master Thesis

Description

Contact Details

More information

Open this project...

Published since: 2025-11-12 , Earliest start: 2025-12-01

Organization Automatic Control Laboratory

Hosts Elokda Ezzat , Nortmann Benita , Montazeri Mina

Topics Engineering and Technology

Online Feedback Optimization for Real Time Power Grid Control

Online Feedback Optimization (OFO) allows to drive a dynamical system to a steady state that satisfies given optimality criteria, such as efficiency and satisfactions of constraints. In the past years, we had great success in applying OFO to the real-time control of power grids, both for energy distribution and transmission. With RTE France we simulated how to solve grid congestion when there is an excess of wind generation. We deployed an OFO algorithm on the Swiss grid to control the voltage and the reactive power of a distribution grid. And we won the Watt D'Or award by the Swiss Federal Office of Energy!

Keywords

smart grid, optimization, control of power grids, feedback optimization, energy

Labels

Semester Project , Bachelor Thesis , Master Thesis

Description

More information

Open this project...

Published since: 2025-11-11 , Earliest start: 2026-01-01

Organization Automatic Control Laboratory

Hosts Bolognani Saverio

Topics Information, Computing and Communication Sciences

Contact

ETH Zurich
Automatic Control Laboratory
Physikstrasse 3
8092 Zurich
Switzerland