THE ROLE OF RENEWABLE ENERGY IN THE GLOBAL ENERGY TRANSFORMATION
Leak Monitoring in Water Distribution Networks
This talk will present different approaches to detect, estimate and locate water leaks (with the main focus in the localization problem) in water distribution networks (WDNs) using hydraulic models and machine learning techniques. First, the actual state of the art will be shortly revised. Then, the theoretical basis of the machine learning techniques considered will be introduced as well as WDN hydraulic models usually used by WDN management companies. Different real WDNs and district metered areas (DMA) will be used to illustrate the presented approaches using simulated and real scenarios. The talk will finalize with a discussion about the comparison results of the different presented approaches and with the current and future research trends in the field of leak monitoring in WDNs.
Characterization of Learning Models via Contrastive Reasoning and Uncertainty
Abstract: In this talk, I will share our most recent work on providing robust representations of visual data while providing explainability to the neural network models. Model-based characterization of neural networks and visual explanations are logical arguments based on visual features that justify the predictions made by neural networks. Our work has been based on quantifying the model changes per input data. We believe that model responses to data and the interaction between the model and the data gives us a window into the model. This can be utilized to provide a number of capabilities such as explainability, uncertainty quantification, and behavior prediction.
Computational Visual Perception for Image and Video Processing and Analysis in the Era of Deep Learning.
Low rotational Inertia Systems and Grid Friendly Power Electronic Converters
The total system inertia (H) is the primary source of electricity system robustness to frequency disturbances. The traditional large synchronous generators directly connected to the grid are the primary sources of inertia, and they play a crucial role in limiting the rate of change of frequency (ROCOF) and provide a natural response to the system frequency changes following an unscheduled loss of generation or demand from the power system. The transition to a low carbon society is the driving force pushing the traditional power system to increase the volume of non-synchronous technologies, which mainly use power converters (PCs) as an interface to the power network. The PCs decoupled the primary source from the power network; therefore, they are not able to contribute with “natural” inertia in the same way as classical synchronous generators. During a system frequency disturbance (SFD), the system frequency will change at a rate initially determined by the total system inertia (H). As a result, the inertial response of the system might be negatively affected, with devastating consequences for system security and reliability. Power Electronic converters (PECs) have been in the power system for many decades; however, it has been just in recent times when the PECs start to be a considerable share of the generation, transmission, and demand. PEC has dramatically changed from a minimal support role to the power system operation to a critical element to the transition to a zero-carbon society. The early developments on high voltage direct current (HVDC) based on Thyristor were a formidable step forward for bulk power transmission. However, the development of more flexible commutation devices and sophisticated control mechanisms, together with appropriate practices and grid codes, are making the voltage source converter (VSC) interfaced technologies a precious component of the operation of modern and future power systems and pavement the secure transition to a zero-carbon society. This seminar presents fundamental aspects of system frequency control in low inertia systems. The seminar initiates with a general introduction of power electronic converter and its transition from the concept of grid-following to grid forming converter, including some practical discussion about the importance of several elements and control philosophy. The seminar includes (but is not limited) to discuss the benefits of an intelligent grid-friendly converter.
New horizons in deep learning-assisted multimodality medical image analysis
Positron emission tomography (PET), x-ray computed tomography (CT) and magnetic resonance imaging (MRI) and their combinations (PET/CT and PET/MRI) provide powerful multimodality techniques for in vivo imaging. This talk presents the fundamental principles of multimodality imaging and reviews themajor applications of artificial intelligence (AI), in particular deep learning approaches, in multimodality medical image analysis. It will inform the audience about a series of advanced development recently carried out at the PET instrumentation & Neuroimaging Lab of Geneva University Hospital and other active research groups. To this end, the applications of deep learning in five generic fields of multimodality medical imaging, including imaging instrumentation design, image denoising (low-dose imaging), image reconstruction quantification and segmentation, radiation dosimetry and computer-aided diagnosis and outcome prediction are discussed.Deep learning algorithms have been widely utilized in various medical image analysis problems owing to the promising results achieved in image reconstruction, segmentation, regression, denoising (low-dose scanning) and radiomics analysis. This talk reflects the tremendous increase in interest in quantitative molecular imaging using deep learning techniques in the past decade to improve image quality and to obtain quantitatively accurate data from dedicated combined PET/CT and PET/MR systems.The deployment of AI-based methods when exposed to a different test dataset requires ensuring that the developed model has sufficient generalizability. This is an important part of quality control measures prior to implementation in the clinic.Novel deep learning techniques are revolutionizing clinical practice and are now offering unique capabilities to the clinical medical imaging community. Future opportunities and the challenges facing the adoption of deep learning approaches and their role in molecular imaging research are also addressed.
State of the art of Energy & Artificial Intelligence and New Challenges
To-date, most of the energy sector’s transition efforts have focused on hardware: new low-carbon infrastructure that will replace legacy carbon-intensive systems. Relatively little effort and investment has focused on another critical tool for the transition: next-generation digital technologies, in particular artificial intelligence (AI). These powerful technologies can be adopted more quickly at larger scales than new hardware solutions, and can become an essential enabler for the energy transition.
AI is already proving its value to the energy transition in multiple domains, driving measurable improvements in renewable energy forecasting, grid operations and optimization, coordination of distributed energy assets and demand-side management, and materials innovation and discovery. AI holds far greater potential to accelerate the global energy transition, but it will only be realized if there is greater AI innovation, adoption and collaboration across the industry.
The principles define the actions that are needed to unlock AI’s potential in the energy sector across three critical domains:
- Governing the use of AI:
- Standards – implement compatible software standards and interoperable interfaces.
- Risk management – agree upon a common technology and education approach to managing the risks presented by AI.
- Responsibility – ensure that AI ethics and responsible use are at the core of AI development and deployment.
- Designing AI that’s fit for purpose:
- Automation – design generation equipment and grid operations for automation and increased autonomy of AI.
- Sustainability – adopt the most energy-efficient infrastructure as well as best practices around sustainable computing to limit the carbon footprint of AI.
- Design – focus AI development on usability and interpretability.
- Enabling the deployment of AI at scale:
- Data – establish data standards, data-sharing mechanisms and platforms to increase the availability and quality of data.
- Education – empower consumers and the energy workforce with a human-centred AI approach and invest in education to match technology and skill development.
- Incentives – create market designs and regulatory frameworks that allow AI use cases to capture the value that they create.
Hydrogel-Based Chemical and Biochemical Sensors
Hydrogels are cross-linked polymer networks able to absorb or to release large amounts of water. The water uptake is associated with a considerable volume change but also with changes of optical properties like the refractive index. The swelling can be excited by a large spectrum of different physical (e.g. temperature, electrical voltage, magnetic field) and chemical factors (e.g. pH value, concentrations of chemical or biochemical species). The particular sensitivity can be adjusted by tailoring the composition of the hydrogel or via its functionalization. If the interaction between hydrogel and analyte to be measured is reversible then such hydrogels are becoming a promising candidate for miniaturized, cost-effective and inline-capable sensors.
In the talk the most important sensor approaches are presented, in particular mechanoelectrical and optical transducers that enable the creation of sensor platforms for a large variety of measurands. Besides sensitivity, also selectivity, long-term-stability and fast sensor response are crucial points which are in the focus of current research. Advantageous approaches to advance the properties of current hydrogel-based sensors, e.g., by force compensation, porous hydrogels, and novel interrogation techniques, will be introduced. Most recent progress in research has already led to first commercial products.
Artificial Intelligence: Concepts, challenges, opportunities and Ethics
If the United States dominates the development and use of artificial intelligence (AI), China intends to challenge this supremacy…Algeria needs to pay attention to and respond to this new AI craze because a country that develops and uses AI will shape its future and significantly improve its economic competitiveness, while a country that falls behind risks losing out its competitiveness in key industries, and even its national sovereignty will be threatened.
Of course, we must not lose sight of the fact that AI arouses both enthusiasm and skepticism, albeit in different measures.
In this plenary session, we recall the concept of Artificial Intelligence and Deep Learning and the importance of Data. We show interest in some key applications in the fields of medicine, telecommunications, intelligent transportation systems, and security.
Ethical issues arise such as surveillance by AI, the role of AI in promoting misinformation and disinformation, the role of AI in politics and international relations, governance of AI, etc.
New technologies are always created in the interest of something good, and AI offers us amazing new capabilities to help people and make the world a better place. But to make the world a better place, we must choose to do so, ethically.
With the concerted effort of many people and organizations, we can hope that AI technology will help us create a better world.