Proyectos

This project will consider the design of Stochastic MPC strategies based Computational Intelligence techniques such as fuzzy models and neural networks, but focusing on achieving theoretical properties such as stability and feasibility, increasing computation speed, and can use control policies. This is novel in that these properties are hardly ever obtained when using nonlinear models in SMPC. Additionally, we will aim to systematize the stability and convergence analyses. These developments will be validated on applications such as irrigation systems, climatization systems, microgrids. This is expected to improve the performance of existing control methods for systems with stochastic uncertainty, and enable the design of advanced control systems where the lack of guarantees or slow computation does not permit the implementation of advanced control systems. Additionally, the development of ad-hoc controllers based on SMPC with RL for Electric Vehicle Routing will be tackled in this research. This controller will be tackled following principles particularly selected for this application since stability is not a major concern here because of the finite horizon for every day of operation.

PROPOSAL ABSTRACT: Cyber-attacks in modular multilevel converters (MMCs) Cyber-attacks on electrical systems are a major concern to many countries as they can have significant impacts on social well-being and economic prosperity. Recent examples of such attacks are (i) the cyber-attacks that targeted the supervisory control data acquisition (SCADA) system of the Ukrainian power grid, causing a power outage that affected approximately 225,000 customers, (ii) the cyber-attack that occurred in 2019 in Venezuela, affecting the power supply of Caracas city, and (iii) the cyber-attacks that affected battery control systems in Korea, resulting in fire and damage. These examples show that cyber-attacks can significantly affect the normal operation of electrical systems and have motivated much research around the world. Much of the published research deals with cyber-attack issues in electrical systems such as microgrids, smart grids and modern power systems. However, cyber-attacks in modular multilevel converters (MMCs) have received limited attention: a recent letter (2021, see introduction section) is the only article published in this area so far. The MMC is deemed to be a prominent solution for medium to high-voltage and high-power applications, with several commercial converters adopting this approach and numerous projects worldwide using this topology. This project aims to investigate cyber-attack issues in the MMC control system and addresses the lack of investigation in this area. In particular, this research will consider the MMC in the context of high voltage direct current (HVDC) transmission systems since this topology is a promising solution to transfer power over long distances. It has been widely used in commercial projects (Trans Bay Cable, Dolwin2, Nano3-terminal DC grid, etc.). In this project, distributed control schemes for controlling the MMC are considered. For this control approach, low-level control tasks are distributed among local controllers placed in the MMC submodules, while high-level control tasks are undertaken by a central controller. The computational burden for the central controller is therefore reduced. The distributed architecture is chosen since the MMC for HVDC applications requires a high number of submodules (SMs). In this case, if the traditional centralised control scheme is used, where a central controller is in charge of processing all the information required for implementing the whole control system, the execution time may not be sufficient for each control cycle to perform all the control tasks, limiting the modularity, flexibility and expandability of this topology (in terms of software development). This project will investigate and quantify the impacts, in terms of operation and stability, of the type of cyber-attack named “false data injection attacks” (FDIAs ) on distributed control systems used for MMCs. Particular focus will be centred on designing strategies for detecting these cyber-attacks and locating the vulnerable sub-modules in the MMC. Also, novel cyber-attack-resilient distributed control schemes will be proposed. It must be pointed out that all the methods derived in this project will be validated through simulation and experimental results. To this end, the cyber-attack detection methods will be based on state observers (Kalman filter, particle filter, etc.) and artificial intelligence (AI) based observers. Note that these detection methods should operate in a distributed architecture, meaning the state estimation should be performed based on partial information of the system. They should be able to be implemented on the SMs local controllers. Finally, once the cyber-attacks are identified, control schemes to neutralise those attacks will be investigated, generating cyber-attack-resilient distributed control schemes for the MMC. The objectives of this proposal are (i) Analyse and quantify the effects of cyber-attacks on the standard distributed control strategies proposed for MMCs and the ones recently proposed based on the consensus theory, (ii) Design distributed methods for the detection of the cyber-attacks considered in this project, (iii) Design of cyber-attacks-resilient distributed control schemes for the MMC, and (iv) Implement an experimental rig for testing and validate the proposals derived from this research. Since there is no literature regarding cyber-attacks in MMCs, the literature review will be focused on work dealing with cyber-attacks on other electrical systems such as microgrids, smart grid and modern power systems. The aim will be to determine if the techniques proposed for these systems can be adapted for the MMC. A similar approach will be followed to address cyber-attack-resilient distributed control schemes. The main contributions of this project will be: 1. The project will provide the foundation for research into cyber-attacks in MMC, as so far, there is very little information in the literature on this topic. Note that cyber-attacks are a hot topic in other electrical systems. Thus it should be investigated for the MMC. 2. Novel distributed methods for detecting cyber-attacks in MMCs will be proposed and validated through simulations and experiment. 3. Novel cyber-attack-resilient distributed control schemes for the MMC will be proposed and validated through simulations and experiment. 4. An experimental rig to analyse and validate the points mentioned above will be designed and built. It will be composed of a central controller and local controllers placed in each SM of the MMC.

Numerous real-life decision-making processes involve solving a task with uncertain input that can be estimated from historic data. There is a growing interest in decision-focused learning methods (a.k.a. smart predict-then-optimize) whose goal is to find models that fit the data while considering how the predicted input will perform in a particular task. For example, the task can be a shortest path problem that uses predictions on travel times in the objective function. Fitting the data and ignoring the task may lead to sub-optimal decisions. Sometimes, uncertainty is involved in the constraints of the model. In this case, ignoring the task would lead to infeasible decisions. The goal of this project is to develop efficient exact algorithms and new applications to train a Machine Learning models that perform well in one or several tasks using mathematical programming (MP) tools. In this context, the typical measure of a predictor is the regret: the excess cost incurred when making a suboptimal decision due to an imprecise predictor. This problem is bilevel in nature: the top-level decision consists in determining a predictor that minimizes a regret while considering that the predictions will affect a task, e.g. an optimization problem in a lower level. This structure is typically non-convex and non-continuous, making the problem difficult to solve for realistic instances. However, several recent advances in bilevel-tailored approaches exploit this structure and can solve large scale problems. There are two main ways of estimating task-oriented predictors: 1) stochastic gradient-based methods; and 2) MP reformulations of the problem. Stochastic gradient-based methods replace the non-differentiable regret for some differentiable surrogate loss function approximating the real loss. Due to advances in neural network implementations and stochastic gradient-descent this approach is the most applied among researchers and practitioners in the Machine Learning community. MP for data science has attracted the attention of researchers and practitioners in different areas as mathematics, operations research, computer science during the last years. It provides some degree of flexibility, being able to model desirable considerations for predictions models. For instance, MP has been successfully used to train sparse models yielding improved explainability and/or fairness. Moreover, MP models are in many cases solvable by any off-the-shelf solver. However, for the decision-focused learning problem there are still many gaps using MP formulations. To the best of our knowledge, MP formulations have been used only for surrogate loss functions. Behind the low usage of MP tools is the scalability. A typical ML setting involves data sets involving thousands of observations. In consequence, the training task becomes more difficult. To tackle this issue, decomposition methods such as cut and column generation can help to solve problems at scale. We aim to provide efficient exact methods that can return either optimal solutions, or optimality guarantees for large scale instances. During the last years, decision-focused learning has been widely applied to combinatorial optimization problems. However, this approach can also be used in many other applications such as Markov decision processes (MDPs) or game theory. In the first case, we can take algorithmic advantage of well-known existing algorithms for MDPs such as value iteration or policy iteration, or approximation techniques such as Q-learning. The game-theoretic setting is more challenging: depending on the notion of equilibrium the loss function would change. For instance, the definition of regret can be applied straightforwardly to Stackelberg equilibria (or leader-follower equilibria), a concept widely applied to energy markets, security and transportation. In the case of Nash equilibrium, the definition of regret is not direct anymore. We hope to develop models that adapt the decision-focused learning paradigm to this broader context. The study of this topic requires the use and development of MP of tools along with their algorithmic analysis. During the project, we expect to develop efficient of algorithms that can be used by decision making, as well as contribute in the understanding of the theoretical aspects of decision-focused learning.

El proyecto tiene como propósito comprender cómo las prácticas artísticas de mediación son un aporte para abordar problemáticas socioculturales y generar estrategias para implementar dichas prácticas en diversos contextos y comunidades del país, expandiendo el rol social de las artes. Su ejecución esta contemplada en cuatro regiones del país y en Barcelona, España, a través de diferentes casos de estudio y un acucioso trabajo de campo.

Proyecto que busca determinar mecanismos moleculares de como los ácidos grasos omega-3 y el ejercicio físico pueden mejorar la calidad de vida, funcionalidad y parámetros de inflamación de pacientes con artritis reumatoide.

El Centro de Modelamiento Matemático (CMM) es un centro científico líder en Chile para la investigación y aplicaciones de las matemáticas. Fue inaugurado en abril del 2000 y forma parte de la Facultad de Ciencias Físicas y Matemáticas (FCFM) de la Universidad de Chile, en la que se encuentra la principal y más antigua escuela de ingeniería del país. Su objetivo es crear nuevas matemáticas y utilizarlas para resolver problemas provenientes de otras ciencias, la industria y las políticas públicas.

Sediment routing systems link the fate of sediment from source to sink in relation to the processes of sediment generation, transport and storage that take place at or near the surface. The transfer of sediment within the sediment routing system involves a cascade of sediment from erosional source areas to depositional sinks in which sediment connectivity between different compartments of the landscape modulate sediment pathways at different scales of space and time. Fluvial systems and transport of suspended sediment are key elements in the transfer of sediment across landscapes and their workings are being altered by climate change and human intervention. In central Chile (30º-37ºS) a decade-long drought is resulting in reduced water discharge, glacier retreat, and diminished sediment discharge to the ocean. The later reflects changing sediment dynamics within the fluvial basins of this region. In this project the temporal and spatial variability of sediment sources and pathways will be studied in the El Volcán River Basin (33ºS), a mountain catchment tributary to the Maipo River, during two consecutive high runoff periods (October-March; 2022-2023 and 2023-2024), with the goal to evaluate interannual and seasonal variations in sediment connectivity in the El Volcán River Basin by identifying the areas of the basin that feed sediment to the fluvial system and describe which pathways of sediment operate under changing flow conditions. Considering the hydrological and sedimentological regime of the mountain catchments in central Chile, the working hypothesis of this proposal is that sediment connectivity in the fluvial system varies throughout seasons as the flow regime, source of runoff, and sediment sources fluctuate from spring to late summer. The variability of the bedrock geology in this basin provides favorable conditions to use sediment provenance techniques to study sediment production and transfer from source to sink at the seasonal and interannual scale. The investigation will start with a geomorphological analysis of the basin that will allow the identification of potential sediment sources (alluvial fans and cones, scree slopes, gullies, fluvial terraces, landslides, etc), which will be consequently sampled. Suspended sediment will be sampled in different parts of this basin during two seasons of high runoff, from the upper tributaries to the catchment’s outlet. Sediment provenance will be analyzed in all suspended sediment samples in order to track sediment sources and pathways. Geochemistry (major and trace elements) is a widely used method to infer sediment provenance in continental environments. Fallout radionuclides (137Cs, 210Pb) are efficiently fixed in fine sediment particles and their activities are independent of lithology and soil type. Therefore, their activities are different in surficial and subsurface sources as well as in recently exposed land or in zones with variable erosion rates. Geochemistry and fallout radionuclides will be measured in the suspended sediment samples and will be compared to the same properties measured in all potential sediment sources within this basin. The compositional results will be analyzed using mixing models in order to establish the relative contribution of each of the potential sources over time. With the study of sediment provenance at seasonal and interannual scales in the El Volcán River Basin it is expected to i) determine temporal and spatial variation in the sources that supply sediment, and therefore the zones within the basin that produce sediment: low, medium, or high El Volcán River Basin, ii) evaluate the erosion processes of the surface (sheet or reel erosion) or subsurface (stream banks, gullies) that participate in the mobilization of sediment to the river, iii) establish variations in the transfer of the sediment provenance signal in this catchment (cascade of sediment from source to sink), iv) evaluate interannual variability of these processes, and v) build a conceptual model of sediment connectivity in this basin during changing flow conditions. The results of this investigation will provide insights into the processes that modulate sediment transport in the Maipo Basin, which is relevant considering the frequent episodes of high turbidity in this river. Moreover, the results could help to forecast the potential influence of projected hydroclimatic changes and anthropogenic activity in central Chile on particle fluxes across the Andes and resultant morphological and sedimentary adjustment of fluvial basins as the sediment is transferred from mountain source to ocean sink.

Chile se ha visto cada vez más afectado por múltiples eventos extremos climáticos que ocurren simultáneamente, como eventos compuestos, o consecutivamente, como eventos en cascada. Los eventos climáticos se consideran compuestos cuando ocurren al mismo tiempo. Por ejemplo, el centro de Chile (la región más poblada del país) se ha visto afectada por sequías frecuentes y severas, agravadas por el aumento de las olas de calor (HWs) que a su vez han favorecido persistentes incendios forestales. Estos eventos extremos han afectado la economía al dañar cultivos y provocar escasez de alimentos para el ganado. Los eventos en cascada actúan como una serie de fichas de dominó que se derrumban. Por ejemplo, en el sur de Chile, las fuertes lluvias orográficas asociadas con potentes ríos atmosféricos (ARs) han provocado graves inundaciones que, al arrastrar sedimentos ricos en nutrientes a lagos y fiordos, a menudo han favorecido floraciones de algas nocivas (HABs). En la misma región, el rápido derretimiento de los campos de hielo patagónicos no solo está canalizando hierro hacia lagos y fiordos (favoreciendo más floraciones de algas), sino que también ha formado cientos de nuevos lagos. El vaciamiento repentino de lagos glaciales (GLOFs) han provocado deslizamientos de tierra e inundaciones que han borrado del mapa pequeños poblados en la Patagonia.