Optimal Control of Hybrid Solar Heat and Power Systems based on MPC and AI methods

Industrial processes use heat in the 50-1500°C temperature range, and heat accounts for around 70% of industrial energy consumption. Heat consumption in industry is generally classified into three temperature ranges: low (400°C), which can be addressed by different solar collector technologies. Concentrating solar technologies are needed to produce solar heat at T>150°C. The central issue of integrating solar heat into industrial processes is addressed in the SHIP4D project (PEPR SPLEEN programme).In this thesis, the work will focus on the high-level optimal control of hybrid solar systems for producing heat and electricity for industrial processes. The control tools will be developed in PEGASE, and applied to a simulator of the LACTOSOL power plant supplied by NEWHEAT. The thesis work will also serve as a basis for the European INDHEAP project (Optimal Solar Systems for Industrial Heat and Power), coordinated by the CEA, and starting in January 2024.

Definition of an asynchronous on-the-fly data compression model on accelerators for HPC

This thesis is related to high-performance computing for numerical simulation of complex physical phenomena.
The CEA provides hardware and software resources to achieve the required computing power.
We have witnessed the advent of accelerators, leading to new problems. In particular, memory management becomes critical for achieving exascale performance as the memory ratio per number of computing units is reducing.
This problem affects all areas requiring a large volume of data. Thus, many aspects of this thesis will be general and of global interest.

This thesis will aim to propose an asynchronous model for making data available through compression/decompression techniques. It should be efficient enough to be used "on the fly" (during computations without slowing them down), allowing memory constraints to be relaxed.
Targeted codes are iterative and sequence different phases. Ideally, all computations will be performed on accelerators, leaving CPU resources unoccupied. The proposed model should take advantage of these specificities. The final goal will be to integrate the work into a representative code to evaluate gains in an industrial context.

Antenna Array In-Situ Calibration through Source Reconstruction

Take the opportunity to develop a motivating career path in a multidisciplinary scientific community at the cutting edge of technological research at CEA Grenoble, as part of an internationally renowned R&D team in the field of antennas.

PhD Subject:
In many advanced applications (radar, direction finding, electromagnetic -EM- context monitoring), precise knowledge of antenna radiation rules the accuracy of processing (angular direction, polarization of received signals). The integration of miniature antennas on objects or vehicles of a few wavelengths largely impacts their radiation pattern. Particularly in low-frequency bands, antenna calibration is not sufficient to achieve the best levels of performance, let alone robustness over time.
The challenge of the proposed PhD is to be able to update the antenna array far-field calibration table in situ (i.e., in near-real time). To do this, the first part on EM analysis will be based on an exhaustive analysis of the equivalent modes/sources induced on the carrier structure via EM simulations, with the aim of extracting the modes present and their radiation. A second part dealing more with RF instrumentation will size and develop an array of spatial sampling probes installed on the structure of the carrier, which will measure the weightings of these modes in situ. Finally, the last part will hybridize the two previous parts in order to reconstruct the far-field radiation by weighting the simulated modes by the measured points.
During the final year, an experimental implementation will be used to validate the methodology and to analyze its performance.
This subject (EM simulation of antennas, EM analyses, RF measurements) will be supervised by an experienced team relying on exceptional tools and instruments (http://www.leti-cea.fr/cea-tech/leti/english/Pages/Applied-Research/Facilities/telecommunications-platform.aspx).

Applicant Profile: Engineer School or Master with major on Antenna, Electromagnetism, RF instrumentation

Laboratory: CEA Grenoble, heart of the French Alps
The CEA is a major research organization working in the best interests of the French State, its economy and citizens. Thanks to its strong roots in fundamental research, it is able to provide tangible solutions to meet their needs in four key fields: Low-carbon energy (nuclear and renewable), Digital technology, Technology for medicine of the future, Defense and national security.
CEA Tech leverages a unique innovation-driven culture and unrivalled expertise to develop and disseminate new technologies for industry, effectively bridging the gap between the worlds of research and business.

Multi-architecture Adaptive Mesh Refinement for multi-material compressible hydrodynamics simulation

CEA DAM is actively developing scientific software in computational fluid mechanics (CFD) for the numerical simulation of compressible and multi-material flows. Such numerical tools requires the use of parallel programming models designed for efficient use of large supercomputers. From the algorithmic point view, the fluid dynamics equations must be discretized and solved using the adaptive mesh refinement (AMR) strategy which allows to reduce the computational cost of such simulations, in particular the number of cells (therefore the memory footprint) and to concentrate the computational work load on the areas of interest (discontinuities, shocks, multi-fluid interfaces, etc. ).

Over the past fifteen years, with the appearance of graphics processors (GPUs), the hardware architectures used in the field of high-performance computing (HPC) have evolved profoundly. This PhD thesis is about designing a parallel implementation of the AMR techniques for the case of multi-material flows with the aim of using as efficiently as possible a GPU-based supercomputer. After required numerical verification and validation process, the developed code will be used to perform numerical simulation of a blast wave and its interaction with surrounding structures.

Design of a misalignment-robust, high-frequency GaN-based inductive power transmission system

The LAIC laboratory of CEA-LETI's Systems Department in Grenoble is specialized in the development of innovative electronic and mechatronic systems, taking into account challenges linked to energy recovery / management / transmission and sensor integration in a variety of environments. As part of the development of its R&D activities, the LAIC is offering a PhD thesis on wireless power transmission using GaN-based resonant inductive coupling.

Wireless power transmission technologies are booming, with applications in space, consumer electronics, medical, automotive and defense sectors. Power transmission technology using resonant inductive coupling appears to be the most promising in terms of near-field efficiency.

The proposed thesis will follow the development of a system including a fixed-coupling electromagnetic coupler and HF electronics based on a GaN transistor-based class-E topology. In this context, the aim of the thesis is to develop a system robust to coupler coil misalignment. The aim is to study, develop and test the performance of a new coupler and an adaptive drive electronics. The candidate will be required to develop analytical and numerical models to optimize the electronics, compare the performance of existing systems in the literature, and propose, develop and test the performance of innovative GaN-based topologies ensuring good robustness to electromagnetic coupling variation.

A multi-disciplinary profile with a focus on power electronics and physics is required for this thesis. In addition to a solid theoretical ground and strong simulation skills, the PhD student will need to be able to work as part of a team, with an aptitude for experimentation and an attraction for practical applications.

3D assembly of GaN power devices

The increase in electrical power density in everyday uses is the result of technological developments including materials and components. The first element to address is the use of a semiconductor material suitable for strong integration and capable of managing high power densities. Since the 2010s, wide bandgap semiconductors such as SiC or GaN have emerged in several applications and are causing a revolution in power electronics design, notably with an increase in the operating frequency and specific power of converters. Concerning Galium nitride (GaN), the increase in switching frequency was made possible thanks to the HEMT (High Electron Mobility Transistor).
The idea of ??this PhD topic is to work on a HEMT GaN cell assembly. The work will involve the an assembly of two components through a common electrode on their backsides in order, making it possible to reduce parasitic inductances and increase the operating frequency. The work will be based on simulation tools such as COMSOL and Synopsys. The thesis will be in collaboration with the GEEPS laboratory at CentraleSupelec and the University of Paris-Saclay.

Privacy-preserving federated learning over vertically partitioned data from heterogeneous participants

Federated learning enables multiple participants to collaboratively train a global model, without sharing their data, but only model parameters are exchanged between the participants and the server. In vertical federated learning (VFL), datasets of the participants share similar samples, but have different features. For instance, companies and institutions from different fields own data with different features of overlapping samples collaborate to solve a machine learning task. Though data are private, VFL remains vulnerable to attacks such as label and feature inference attacks. Various privacy measures (e.g., differential privacy, homomorphic encryption) have been investigated to prevent privacy leakage. Choosing the appropriate measures is a challenging task as it depends on the VLF architecture and the desired level of privacy (e.g., local models, intermediate results, learned models). The variability of each participant’s system can also result in high latency and asynchronous updates, affecting training efficiency and model effectiveness.

The aim of this thesis is to propose methods to enable privacy-preserving VFL, taking into account the heterogeneity of the participants. First, the candidate will study the architectures of VFL models and the privacy measures to propose privacy-preserving protocols for VFL. Second, the candidate will investigate the impacts of the heterogeneity of the participants’ system such as computation and communication resources to devise solutions to render the VFL protocols robust to such heterogeneity. Third, the trade-offs among effectiveness, privacy, and efficiency in VFL will be explored to propose a practical framework for adjusting the protocols according to the requirements of a given machine learning problem.

Advanced fully-depleted Silicon-on-insulator devices for Radio-Frequency applications

The PhD will be performed in the NEXTGEN project aimed at developing the next generation of Silicon-on-insulator devices. Our laboratory is driving the development of the RF active devices: this is a great opportunity to carry out fundamental research using state-of-the art processing equipment and characterization instruments while working in close collaboration with our industrial partners.

you will expected to engage in tasks encompassing:
- perform back-of the-envelope estimation of device properties and assess performace impact of technological choices
- Perform and/or analyze TCAD simulations to gain insight in the RF device behaviour
- data-mining on engineering measurements: grasp the relevant information and identify trends or correlations
- perform extensive periods of time in the lab to conduct or participate in on-wafer RF characterization champaign.
Based on you profile or expectations, above tasks may be dynamically rebalanced during the thesis.

Development and characterization of low temperature Cu-dielectric hybrid bonding

Cu-dielectric hybrid bonding is a technology that enables the assembly of components with very fine interconnection pitch, opening the path to new integrations for advanced applications such as High Performance Computing, Smart Imagers,… Leti has been engaged for more than 10 years in the development of this technology, in partnership with various industries and academies, to master smaller and smaller connection pitches (< 1µm), or to evaluate new techniques such as ‘die-to-wafer’ self-assembly. In this context, low temperature hybrid bonding would allow new integration routes notably for heterogeneous systems (III-V on CMOS,…) or for thermally sensitive components (colored resins, non-volatile memories,…).

The objective of this thesis is to develop and characterize Cu-dielectric hybrid assemblies performed at low temperature, from ambient to 250°C. A first part of the thesis will aim at identifying the dielectric materials that are relevant for the hybrid bonding technology (SiN, SiON, SiCN, …). The critical properties of these materials (permittivity, hygroscopy,…) will be measured and compared to the reference high temperature SiO2. In a second part, the selected dielectrics will be integrated in the ‘wafer-to-wafer’ hybrid bonding technology and each process step (damascene level, surface preparation, direct bonding) will be adapted as needed. The third part of the thesis will be dedicated to the electrical characterization and reliability tests of the obtained low temperature hybrid bonding.

Stocastic integrated power supplies based on emerging components

The widespread utilization of connected devices that process sensitive information necessitates the creation of new secure systems. The prevalent attack, referred to as power side-channel, involves the retrieval of encryption key information by analyzing the power consumption of the system. Integrating the system with its power supply management blocks can conceal the consumption of sensitive blocks, especially by utilizing various techniques to introduce randomized variations during power transfer. The CEA has wide experience in the design and testing of secure integrated circuits and it is exploring a new approach to DC-DC conversion that uses emerging devices available at CEA-Léti.
The work of the PhD researcher will be the following:
- Specification of integrated power supplies using switched-capacitor architecture.
- Study the circuit using emerging components and evaluate the improvement of its robustness against side channel attacks.
- Design of the integrated power supply in silicon technology.
- Performance and security characterization of the designed blocks and security primitives in
their whole.
The division of labor is 10% advanced study, 20% system architecture, 50% circuit design, 20% experimental measurement.