ECC23 will host the following eight confirmed tutorial sessions:

1.  Data-driven models for control design. Organizers: Riccardo Scattolini, Melanie N. Zeilinger.
2.  Modern control problems in modern power systems. Organizers: Alessio Iovine, Patrick Pantiatici.
3.   Model reduction for control. Organizers: Jacquelien M.A. Scherpen, Tudor C. Ionescu, Alessio Moreschini, Joel David Simard.
4.  State of the art and new trends in explicit MPC . Organizer: Florin Stoican, Martin Mönnigmann, Sorin Olaru, Ionela Prodan, Mircea Lazar.
5.   Modeling and treatment of cancer. Organizers: Catherine Bonnet, Jean Clairambault, Walid Djema, Frederic Mazenc.
6. Control and operation of wastewater treatment systems. Organizers: Marian Barbu, Montse Meneses, Pastora Vega, Ramon Vilanova
7. Dynamic Virtual Power Plant for Powering System Flexibility Through Renewables. Organizers: Bogdan Marinescu, Horst Schulte, Florian Dörfler, Eduardo Prieto-Araujo, Oluwaseun Enoch Oladimeji.
8. SeaClear: search, identification and collection of marine litter with autonomous robots. Organizers: Lucian Bușoniu, Cosmin Delea, Bart De Schutter, Ivana Palunko, Stefan Sosnowski.

Data-Driven Models for Control Design

Date & Time:

Wednesday, June 14th 2023 @ 14:30-16:30

Location:

Library. Room L.2.1

View Map

Organizers:

Riccardo Scattolini, Melanie N. Zeilinger

Abstract:

This tutorial session aims to present some techniques based on a data-driven approach for modeling, control, and optimization of dynamic systems. Methods based on the use of Recurrent Neural Networks, non-parametric regression, and Set-Membership identification will be considered. Applications of these approaches for solving challenging control problems will be shown, such as those related to autonomous driving of vehicles. The goal is to provide a general overview of the approaches considered and their structural properties to enable their informed use in a wide class of problems.

The tutorial session will consist of the following talks:

• Recurrent Neural Networks for Learning-Based Control (Riccardo Scattolini)
• Model Learning for Control using Non-Parametric Techniques: Practical Results and Guarantees (Elena Arcari, Emilio Maddalena)
• Dealing with Uncertainty in Control and Optimization: the Set Membership Paradigm (Lorenzo Fagiano)
• Learning to Drive Efficiently Using Digital Twin (Son Tong)

Modern Control Problems in Modern Power Systems

Date & Time:

Thursday, June 15th 2023 @  13:30-15:30

Location:

Library. Room L.2.1
View Map

Organizers:

Alessio Iovine, Patrick Pantiatici

Abstract:

The future global transition to the paradigm of “Smart Grid” needs contributions from the automatic control community, as the problems to be solved in the short, medium and long term are numerous and impacting across the whole hierarchy of control levels. Recent work demonstrates the need for the integration of control systems in the fields of electrical generation/transmission/consumption, more and more based on power electronics, and a key element in transforming the current electrical grid into a new Smart Grid capable to handle massive integration of DER and ESS will revolve around the possibility of having more effective control methods.  Based on the new possibility to compose and decompose (pieces of) power systems in real-time, there is a need for control methods based on System of Systems (SoSs) approaches. Moreover, modern power systems need to combine a cyber side with the physical one, where the cyber side monitors and controls the physical processes, with feedback loops where physics affects computation and vice versa. To consider power systems as Cyber-Physical Systems (CPSs) is one of the means to facilitate the integration of new components. The physical layer comprising of power/energy infrastructures and the cyber one comprising control, communication, and computation need to be designed to achieve the overall low carbon energy and energy conservation goals. Thanks to their reconfigurable structure, CPSs can support the definition of the modern power systems as a flexible and resilient composition of (clusters of) agents, thus facilitating the possibility to decompose them in more manageable sub-systems.

The tutorial session will consist of the following talks:

• The Energy Transition: a Revolution for Electrical Systems more than Ever Cyber-Physical Systems of Systems (Patrick Panciatici)
• Robust Control for Dynamical Systems With Non-Gaussian Noise via Formal Abstraction (Alessandro Abate)
• Local Power Congestion Management in Subtransmission Areas (Alessio Iovine)
• Congestion Control in Transmission Grids via Hierarchical Feedback Optimization (Saverio Bolognani)
• Coordinated Blackstart Through Inverter-Based Generation (Adolfo Anta)

Model Reduction for Control

Date & Time:

Thursday, June 15th 2023 @  16:00-18:00

Location:

Library. Room L.2.1
View Map

Organizers:

Jacquelien M.A. Scherpen, Tudor C. Ionescu, Alessio Moreschini, Joel David Simard

Abstract:

Mathematical models are constantly increasing in dimension and complexity. The large number of differential equations as well as the connection between their variables make the models difficult to solve, simulate and control. Model reduction is called for to to yield simpler, lower-dimensional approximations, used for control design. The tutorial session contains the latest, most significant results in the model re- duction techniques tailored to the modern context of data- driven control. The session contains four talks on (nonlinear) balanced truncation and moment matching-based techniques, respectively.

The tutorial session will consist of the following talks:

• Data-Based Model Reduction for Non-Linear Systems Based on Differential Balancing (Jacquelien M. A. Scherpen)
• Model Reduction with Pole-Zero Placement and High Order Moment Matching (Tudor C. Ionescu)
• The Role of Loewner Matrices for Structure-Preserving Model Reduction (A. Moreschini)
• A New Parameterization of Interpolants in the Loewner Framework with Applications to Stability and Structure Assignment (J. D. Simard)

State of the Art and New Trends in Explicit MPC

Date & Time:

Wednesday, June 14th 2023 @ 17:00-19:00

Location:

Library. Room L.2.1

View Map

Organizers:

Florin Stoican, Martin Mönnigmann, Sorin Olaru, Ionela Prodan, Mircea Lazar

Abstract:

Explicit MPC has been a mainstay of control theory for the last 20 years. It couples a deceptively simple description with fast (in the online part at least) implementa- tions. Recent advances in hardware capabilities and software libraries make it even more relevant for today’s applications.  We propose to: i) recapitulate the main results from the state of the art by emphasizing a common narrative centered around combinatorial and geometric interpretations; and ii) present new trends which improve computation time, offer new descriptions and open new areas of application.

The tutorial session will consist of the following talks:

• The Last 20 Years in Explicit MPC: Geometrical and Combinatorial Approaches (Florin Stoican)
• Constructing Explicit MPC Laws by Backward Dynamic Programming with Active Sets (Martin Mönnigmann)
• Convex Liftings in Control Design and Connections with Inverse Optimality (Sorin Olaru)
• On the Implementation of an Explicit MPC Via Flat Mapping for a Multicopter System (Ionela Prodan)
• On Neural Network Architectures for Nonlinear MPC (Mircea Lazar)

Modeling and Treatment of Cancer

Date & Time:

Friday, June 16th 2023 @ 09:20-11:20

Location:

Library. Room L.2.1

View Map

Organizers:

Catherine Bonnet*, Jean Clairambault, Walid Djema, Frederic Mazenc

Abstract:

This tutorial aims at gathering scientists interested in the questions of modeling and control of cancer (with a focus on blood cancer) through different points of view (taking into account or not mutations, treatment resistance, cell plasticity. . .) and tools (analysis of Ordinary Differential Equations, of Ordinary Differential Equations with delays, of Partial Differential Equations ; estimation of parameters ; planning tools ; deterministic and probabilistic frameworks. . .).The tutorial should be of interest for modelers, control theory theorists, applied control theory specialists, mathematicians and medical doctors (if any in the audience) in order to share knowledge on strategies available to best answer medical questions from an applied point of view (in link with hospital) with advanced control theory methods at large.

Note that two first speakers of this tutorial are medical doctors (in addition to being applied maths researchers) and that many studies are done in collaboration with renamed Hospitals (e.g. Hôpital Saint-Antoine, Paris, France; Institut Gustave Roussy, Villejuif, France). Having in mind the set of all proposed contributions, the instructors will see beforehand with each author how he/she can present his/her point of view by putting their results into perspective with regard to the points of view of other authors.

The tutorial session will consist of the following talks:

• Mathematical Modelling of Cancer Growth and Drug Treatments : Taking into Account Cell Population Heterogeneity and Plasticity (Jean Clairambault)
• The Impact of the Stem Cell Niche on Blood Cancer Dynamics (Thomas Stiehl)
• Early Screening for Blood Cancers (Gurvan Hermange)
• A Practical Cell Density Stabilization Technique Through Sub-Optimal Drug Infusions (Hitay Ozbay and Walid Djema)
• Regions of Attraction Estimation for Tumor Immune Dynamical Systems (Kaouther Moussa)

Control and Operation of Wastewater Treatment Systems

Date & Time:

Thursday, June 15th 2023 @ 09:20-11:20

Location:

Library. Room L.2.1

View Map

Organizers:

Marian Barbu, Montse Meneses, Pastora Vega, Ramon Vilanova

Abstract:

A Wastewater Treatment Plant (hereafter WWTP) is a complex structure developed to treat domestic, industrial, agricultural influent or often a combination of all three. The operation of a WWTP often has the primary objective of ensuring that discharged effluent complies with the local regulations in term of water quality, despite changing influent conditions. However, despite the advances made in recent decades, a large percentage of wastewater treatment plants are still being operated below the optimal performance achievable. This low performance becomes evident in the form of both treated water discharges that do not meet water quality standards and low efficiencies in terms of energy consumption. Additionally, many existing urban treatment plants face more stringent criteria for their wastewater effluent that needs to be treated.

Additionally, in recent years, WWTP operators have experienced increased pressure not only to meet effluent standards, but also to increase energy efficiency, perform resource recovery, and monitor and mitigate greenhouse gas (GHG) emissions. Recent studies have identified WWTPs as potential sources of anthropogenic GHG emissions that contribute to climate change and air pollution, such as methane (CH4) and nitrous oxide (N2O). WWTPs also emit carbon dioxide (CO2) during the production of the energy required for the plant operation.

In this tutorial we aim to provide a presentation of the state of the art regarding modelling and control at WWTP wide level, covering both, the sewer system and the WWTP itself as a complex water system. The role of benchmarking and control approaches scopes will be addressed. From nutrient removal, to joint economic approaches and recent advances that aim to take into account GHG emissions as one of the hot current challenges.

The tutorial session will consist of the following talks:

• Wastewater Treatment Plant: Process Modelling, Benchmarking and Control Problems (Ramon Vilanova, Marian Barbu)
•  Control Approaches for Nutrient Removal (Pastora Vega, Marian Barbu)
•  Environmental Impact Considerations (Montse Meneses, Marian Barbu)
•  Global Approaches: Plant Wide and Extended Performance Considerations (Marian Barbu, Montse Meneses, Pastora Vega)
•  Considerations on GHG Emissions (Ramon Vilanova, Marian Barbu)

Dynamic Virtual Power Plant for Powering System Flexibility Through Renewables

Date & Time:

Wednesday, June 14th 2023 @ 09:50-11:50

Location:

Library. Room L.2.1

View Map

Organizers:

Bogdan Marinescu, Horst Schulte, Florian Dörfler, Eduardo Prieto-Araujo, Oluwaseun Enoch Oladimeji

Abstract:

The concept of Virtual Power Plant (VPP) has arisen over a decade ago from the relatively low competitiveness of the back then emerging Renewable Energy Sources (RES). A set of RES generators imitates the behavior of large synchronous generators. So far, static aspects such as generation or slow dynamics have been of interest. However, considering dynamic aspects is of high importance, especially to further increase the current penetration level of RES. For that, we have proposed a new concept called Dynamic VPP (DVPP) which fully integrates the dynamic aspects at all levels: locally (for each RES generator), globally (for grid ancillary services and interaction with other close-by elements of the grid) and economically (for internal optimal dispatch and participation in electricity markets). A DVPP is a set of dispatchable and non-dispatchable RES along with a set of common control and operation procedures. We propose a top-down set of control structures and synthesis methodologies to deal in a coordinated way to all aspects of operation of the DVPP:  (a) Coordination of the control at all time scales of the DVPP: from very fast dynamics of the power converters which connect RES generators to the grid (few milliseconds) to internal DVPP redispatching to cope with volatility of natural resources (sun and wind) (few seconds) and DVPP participation to secondary (voltage & frequency) controls and power & energy markets (few minutes/hours); (b) Coordination of controls among space distributed actuators (the DVPP RES generators); (c) Solutions for centralized vs decentralized control synthesis and implementation as a trade-off between performances and resilience; (d) Modeling & simulation for stiff systems (very fast dynamics of power electronics of RES & slower dynamics of classic generators).

The tutorial session will consist of the following talks:

• The Dynamic Virtual Power Plant (DVPP): Concept and Coordinated Control of DVPP Renewable Generators (Bogdan Marinescu)
•  Pole Region and Model Matching Control Design for LPV systems with Application to Wind and Solar PV Power Plants for DVPP Integration (Horst Schulte)
• Decentralized Model Matching DVPP Control (Florian Dörfler)
• Modeling and Control of Modern Power Networks Incorporating DVPP (Eduardo Prieto-Araujo)
• Real-Time Operation of Dynamic Virtual Power Plants: Integrating Re- Dispatch Optimization with the Dynamic DVPP Model (Oluwaseun Enoch Oladimeji)

SeaClear: search, identification and collection of marine litter with autonomous robots

Date & Time:

Friday, June 16th 2023 @ 14:20-16:20

Location:

Library. Room L.2.1

View Map

Organizers:

Lucian Busoniu, Cosmin Delea, Bart De Schutter, Ivana Palunko, Stefan Sosnowski

Abstract:

The objective of this tutorial is to present the SeaClear European Horizon 2020 project, which is developing the first autonomous robotic team for search, identification, and collection of underwater robotic litter. The team consists of an observation unmanned underwater vehicle that searches for litter, a second underwater vehicle that collects the found litter using a combined gripper-suction device, an unmanned aerial vehicle to detect both litter and the locations of the underwater vehicles, and a surface vehicle that is the hub of the system. After introducing the overall system and approach, we detail communication, sensing, and control aspects that are relevant for the audience of the European Control Conference, paying special attention to practical challenges that arise in the operation of the system. Our tutorial will be useful to control engineering students, researchers, and practitioners with interests in marine and aerial robotics, nonlinear control and estimation, data-driven control, or artificial intelligence.

The tutorial session will consist of the following talks:

• SeaClear: From System Design to Sea Trials (Cosmin Delea)
• SeaClear: Litter Detection (Bart De Schutter)
• SeaClear: Pose Estimation and Mapping (Lucian Busoniu)
• SeaClear: the Control Perspective (Stefan Sosnowski)
• SeaClear: UAV Control and Sensing (Ivana Palunko)