Machine-learning (ML) holds significant promise in revolutionizing a wide range of applications, in particular in the domain of multi-scale and multi-physics problems. Success in realizing the promise of ML is predicated on the availability of training data, which are often obtained from scientific computations. Conventional approaches to solving the equations of physics require difficult and specialized software development, grid generation and adaptation, and the use of specialized data and software pipelines that differ from those adopted in ML. A disruptive new approach that was recently proposed by the US team is Evolutional Deep Neural Networks (EDNN, pronounced ``Eden") which leverages the software and hardware infrastructure used in ML to replace conventional computational methods, and to tackle their shortcomings. EDNN is unique because it does not rely on training to express known solutions, but rather the network parameters evolve using the governing physical laws such that the network can predict the evolution of the physical system. In the proposed effort, the EDNN framework will be extended to solve high-dimensional partial differential equations, used to model a vast range of phenomena in economics, finance, operational research, and multi-phase fluid dynamics, where population balance equations govern phenomena as diverse as aerosol transmission of airborne pathogens or mixing enhancement in energy conversion devices. The simulation of such flows is an open issue of particular interest to the US and French teams, a strong motivation for the proposed collaboration. We will demonstrate the ease of software development using automatic differentiation tools and the capacity of EDNN to eliminate the curse of dimensionality and the tyranny of moment closure. Success stands to disrupt and transform the decades-old computational approach to solving nonlinear differential equations and to remove the barriers to generation of training data required for ML.

In this research, we will investigate and develop applications of multistable composite panels and shells as morphing surfaces capable of large shape changes under external loading and embedded actuation. This research topic implies the study of complex non linear phenomena that take place during large displacements of slender structures (beams, arches, plates or shells) in order to develop models and tools for the prediction, design and control of such phenomena. Large non linearities are the fundamental keys to understand stability or loss of stability of slender structures, both in the static and dynamic domains. Our aim is to tackle two relevant subjects in the very large domain of non linear structural behaviour, which will represent the two main axes of the present research project: i. a first goal is to develop an integrated theoretical and numerical approach to design composite laminates which may hold several equilibrium configurations, exploiting the effects of geometric non-linearities, anisotropy, and pre-stresses induced by the through-the-thickness variation of the material properties. The idea is to start from reasonably simple analytical models for the study of complex non linearities, in order to capture the fundamental phenomena involved, and to validate the captured trends through numerical and experimental tools. These results will open the way to the modelling, design and optimisation of more complex multistable structures; ii. a second aspect is the theoretical study and experimental test of the use of embedded active materials, such as Shape Memory Alloys (SMA) wires/stripes, to control the shape of the structure without the need for joints or conventional actuators. Synthetically, our topic is the modelling, optimal design and shape control of slender multistable structures, i.e. structures that can hold several equilibrium positions without the application of external forces. Domains of application are quite wide and they stand at the border of very innovative and recent research topics: such structures are present in Nature, but also are constitutive parts of modern devices, such as morphing or deployable structures for aeronautical and aerospace applications, mirrors, thin screens, and so on. The developments proposed in our project embrace several aspects: modelling, numerical and experimental validation, optimal design of non linear slender structures. Considering the field of application and the envisaged research, the present project can be evaluated in the framework of the FRAE-ANR protocol (strong links to themes I, VI, VIII, IX; also possible link to theme III). This project aims at gathering a group of young scientists who are experts in different fields of structural mechanics (modelling, optimisation and design, elasticity, smart materials, singularity formation, stability, damage,…) in order to build up and consolidate a research team based at Institut Jean Le Rond d’Alembert (IJLRdA), Université Pierre et Marie Curie (UPMC) with external collaborations from Ecole Nationale Supérieure de Cachan and Università Roma 1 La Sapienza. This project will also enhance exchanges among the different fields of structural mechanics which are the background of the team members. Indeed, this project will be also an opportunity to give a contribution to the nascent experimental activity at IJLRdA by the development of simple experimental setups. This experimental activity was initiated few years ago (in the field of fluid mechanics) and it is highly encouraged by the direction of the institute IJLRdA and CNRS.

The PLASINTER project main objective is to expand the knowledge on catalysts behavior, both in structure and chemical nature, under non-thermal plasma CO2 hydrogenation. There is consensus, in the plasma community, that upgrading of this process can be achieved by untangling plasma-chemical processes. The design of catalysts specifically tailored for activation by plasma combined to in-situ / operando characterization with the support of modelling / simulation would yield in-depth information on the underlying processes and allow the development of optimum catalytic phases and supports. This fundamental project aims at understanding how the micro-meso porosity of monolithic channels modifies the plasma discharge propagation, thus activation barrier on the catalyst and subsequent CO2 activation. The analysis of the catalyst surface and gas phase by in-situ infrared probing techniques will be performed in order to establish the properties of atmospheric pressure plasma developing within the monoliths and identify the plasma chemistry induced on surfaces. The design of well-defined catalysts with controlled-structure will help to link the effects of the monolith’s micro-channels on the hydrogenation reaction. Ni or Cu supported mesoporous alumina and silica will be synthesized and effects of promoters on catalyst physico-chemical properties evaluated. The investigation by in-situ combined to ex-situ experimentations will provide key knowledge on catalyst design and reaction mechanisms under plasma conditions.

In the recent years armored materials have been intensively developed for various applications. The renaissance of gerrilla groups in several international and national conflicts around the world has increased the number of incidents involving light weapons. These weapons are increasingly harmful, now being able to perforate the armored vehicles and aircrafts that were secure before and making compulsory to reinforce the resistance of the materials used. In addition, there is an interest in increasing the action area of these vehicles being exposed to more severe conditions. The design of armored materials must find a compromise between performance and weight. This can only be achieved investing into new research lines focusing on the development of new materials with significant added value. Recent advances on the performance of armored materials have been based on materials able to quickly dissipate most of the kinetic energy generated during the impact of the projectile on the armored plate. For instance, ceramic materials have shown a good balance between weight and the amount of energy dissipated during the fracture process. Composite materials present similar densities and energy dissipation rates by resorting to the delamination process. Finally, porous materials can be even lighter being the process of compaction the main responsible for the energy dissipation. One energy dissipation process that remains still unexplored is the process of cavitation. The project CACHMAP (Laser shock induced cavitation in porous matrices) is a fundamental research project aiming at quantifying the total amount of energy dissipated by cavitation when a light solid porous matrix saturated with a liquid is exposed to short pressure pulses. Because this project proposes the development of new technologies to protect citizens against ballistic aggressions, it fits into the EU societal challenge “Secure societies- Protecting freedom and security of Europe and its citizens”. The project is proposed for a total duration of 3 years and it is based on the experimental characterization of various porous matrix saturated with various liquids. The use of shock waves induced by short laser pulses will allow characterizing micron size porous material samples with sizes in the range of few millimeters thickness and with surfaces of the order of square centimeters. The project will concentrate on the implementation and development of the experimental techniques required to characterize the processes. In a first stage, polymeric piezoelectric pressure sensors will be used to capture the averaged response of the material to shock waves which is expected to be influenced by the inception of cavitation. This technology will be further developed in order to filter out specific range of frequencies adapted to laser pulses. State of the art data treatment methods will be also explored in order to maximize the amount of information extracted from the experimental signals. Finally, the project will also include the development of models capable to predict the response of dry and saturated porous matrices which may be eventually used for future designs of new armored materials.

Human societies share the ability to develop cultural traits and to have them evolve in various geo-cultural environments, imposing different types of constraints on cultural changes. Given the geographic spread of human societies and their vast cultural diversity, these changes vary according to the populations considered and to the nature of the observed cultural domains. Mechanisms at stake in the transformation of a cultural object are still little-known. While one can easily describe the effects of diversity and hence the results of cultural evolution at one precise moment in time, the processes put in place by human societies to culturally distinguish themselves from one another are still rarely studied. Music and language are two of the cultural domains shared by all human communities on the planet. The ngombi project proposes to study the transformation processes of musical instruments in oral tradition populations while grounding its approach on inter-disciplinary research, combining methods from social sciences and natural sciences. The aim is to understand the specific mechanisms of instruments’ transformation processes, but also to understand the impact of socio-cultural contexts on these mechanisms. This study, which claims to be exploratory, will concern more specifically Central African harps. The choice of the instrument and of the study’s perspective are due to the geo-cultural anchoring of the team members (CAR, Gabon, Congo, Uganda), and to the research orientations developed i.e. the global comprehension of creation, transformation and diffusion processes of the musical patrimonies’ elements. Central African harps that are found nowadays, and which exist in the form of historical artifacts in museums’ collections bear witness of the great diversity of their morphologic and acoustic characteristics, of their repertoires and designations. Despite this diversity, it is nonetheless possible to recognize some similarities grounded for instance on the shape of the sound box, the symbolic representation of the instrument, the designation or the associated songs’ themes. Several studies have shown that these resemblances can transcend the identity distinctive features of the populations who use them (ethnonyms, linguistic groups), as well as their geographical dispersion. One of our first hypotheses is thus that, despite the diversity observed, it is possible to reveal proto-forms (which could be compared to hypothetical “ancestors”) on which would be grounded a categorization of Central African harps through the acknowledgment of common traits. Our aim is to determine if the processes leading to harps’ diversity are concurrents of the transformation of their socio-cultural contexts of performance and/or of identity strategies at different levels. The following hypothesis is to be verified: are the harps’ transformations due to adaptations to their performative context and/or to a multiplicity of markers of identity (ethnic, linguistic, technical, symbolic, etc.)? By using objects from museum collections as well as objects studied in the field, the novelty of our approach is not only to compare objects considering organological, acoustic and linguistic domains, but also to introduce anthropological data, allowing us to take into account the various factors having an impact of the harps’ variability and diversity.