Search

This paper proposes an integrated actuator and sensor active fault-tolerant model predictive control scheme. In this scheme, fault detection is implemented by using a set-valued observer, fault isolation (FI) is performed by set manipulations, and fault-tolerant control is carried out through the design of a robust model predictive control law. In this paper, a set-valued observer is used to passively complete the fault detection task, while FI is actively performed by making use of the constraint-handling capability of robust model predictive control. The set-valued observer is chosen to implement fault detection and isolation (FDI) because of its simple mathematical structure that is not affected by the type of faults such as sensor, actuator, and system-structural faults. This means that only one set-valued observer is needed to monitor all considered actuator and sensor statuses (health and fault) and to carry out the fault detection and isolation task instead of using a bank of observers (each observer matching a health/fault status). Furthermore, in the proposed scheme, the advantage of robust model predictive control is that it can effectively deal with system constraints, disturbances, and noises and allow to implement an active FI strategy, which can improve FI sensitivity when compared with the passive FI methods. Finally, a case study based on the well-known two-tank system is used to illustrate the effectiveness of the proposed fault-tolerant model predictive control scheme., This work has been partially funded by the Spanish Government (MINECO) through the project CICYT ECOCIS (ref. DPI2013-48243-C2-1-R), MINECO and FEDER through the project CICYT HARCRICS (ref. DPI2014- 58104-R), the Natural Science Foundation of Guangdong (No.2015A030313881), the National Natural Science Foundation of China (No.61673239), and the Shenzhen Key Lab of Space Robotic Technology and Telescience.

With the aim of distributing the total power required at the wind farm control level for each wind turbine, we proposed a multi-objective model predictive control scheme able to (i) ensure the tracking of the power demanded by the electrical grid; (ii) provide primary frequency control by de-loading the wind turbines; and (iii) maximize the total power reserve when the available power is higher than the power demand.

© 2017 IEEE. Personal use of this ma terial is permitted. Permission from IEEE must be obtained for al l other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, f or resale or redistribution to se rvers or lists, or reuse of any copyrighted compone nt of this work in other works The problem of energy dispatch in heterogeneous complex systems such as smart grids cannot be efficiently solved using classical control or ad-hoc methods. This paper proposes the application of Economic Model Predictive Control (EMPC) for the management of a smart micro-grid system connected to an electrical power grid. The system comprises several subsystems, namely some photovoltaic (PV) panels, a wind generator, a hydroelectric generator, a diesel generator, and some storage devices (batteries). The batteries are charged with the energy from the PV panels, wind and hydroelectric generators, and they are discharged whenever the generators produce less energy than needed. The subsystems are interconnected via a DC Bus, from which load demands are satisfied. Assuming the load demand and the energy prices to be known, this study shows that EMPC is economically superior to other Model Predictive Control (MPC) based strategies (a standard tracking MPC, and their cascaded version in form of hierarchical two-layer approach).

This chapter proposes a reliable fault-tolerant model predictive control applied to drinking-water transport networks. After a fault has occurred, the predictive controller should be redesigned to cope with the fault effect. Before starting to apply the fault-tolerant control strategy, it should be evaluated whether the predictive controller will be able to continue operating after the fault appearance. This is done by means of a structural analysis to determine loss of controllability after the fault complemented with feasibility analysis of the optimization problem related to the predictive controller design, so as to consider the fault effect in actuator constraints. Moreover, by evaluating the admissibility of the different actuator-fault configurations, critical actuators regarding fault tolerance can be identified considering structural, feasibility, performance and reliability analyses. On the other hand, the proposed approach allows a degradation analysis of the system to be performed. As a result of these analyses, the predictive controller design can be modified by adapting constraints such that the best achievable performance with some pre-established level of reliability will be achieved. The proposed approach is tested on the Barcelona WTN.

Control of drinking water networks is an arduous task, given their size and the presence of uncertainty in water demand. It is necessary to impose different constraints for ensuring a reliable water supply in the most economic and safe ways. To cope with uncertainty in system disturbances due to the stochastic water demand/consumption and optimize operational costs, this paper proposes three stochastic model predictive control (MPC) approaches, namely, chance-constrained MPC, tree-based MPC, and multiple-scenario MPC. A comparative assessment of these approaches is performed when they are applied to real case studies, specifically, a sector and an aggregate version of the Barcelona drinking water network in Spain., This research has been supported by the Spanish Project ECOCIS (Ref. DPI2013-48243-C2-1-R), the EU Network of Excellence Highly Complex and Networked Control Systems (HYCON2) under grant agreement no. 257462, and the DGR of Generalitat de Catalunya (SAC group Ref. 2014/SGR/0375).

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License., This paper proposes a reliability-based economic model predictive control (MPC) strategy for the management of generalised flow-based networks, integrating some ideas on network service reliability, dynamic safety stock planning, and degradation of equipment health. The proposed strategy is based on a single-layer economic optimisation problem with dynamic constraints, which includes two enhancements with respect to existing approaches. The first enhancement considers chance-constraint programming to compute an optimal inventory replenishment policy based on a desired risk acceptability level, leading to dynamically allocate safety stocks in flow-based networks to satisfy non-stationary flow demands. The second enhancement computes a smart distribution of the control effort and maximises actuators’ availability by estimating their degradation and reliability. The proposed approach is illustrated with an application of water transport networks using the Barcelona network as the considered case study., The authors wish to thank the projects that funded this research, i.e., the EU project EFFINET (FP7-ICT-2011-8-31855) and the Spanish projects ECOCIS (CICYT DPI2013-48243-C2-1-R) and SHERECS (CICYT DPI2011-26243) of the Spanish Ministry of Economy and Competitiveness.

Trabajo presentado a la 2nd KES International Conference on Innovation in Medicine and Healthcare (InMed-14), celebrada en San Sebastian (España) del 9 al 11 de julio de 2014., This extended abstract addresses the preliminary results of applying uncertainty handling strategies and advanced control techniques to the inventary management of hospitality pharmacy. Inventory management is one of the main tasks that a pharmacy department has to carry out in a hospital. It is a complex problem because it requires to establish a tradeoff between contradictory optimization criteria. The final goal of the proposed research is to update the inventory management system of hospitals such that it is possible to reduce the average inventory while maintaining preestablished clinical guarantees., The authors would like to acknowledge Junta de Andalucía (Pharmacontrol Project, P12- TIC-2400), for funding this work.

Presentado al 20th MED celebrado en Barcelona del 3 al 6 de julio de 2012., In this paper, a hierarchical temporal multi- layer decentralised model predictive control (HTML-DMPC) approach for drinking water networks (DWN) is proposed. The upper temporal layer works with a daily scale and is in charge of achieving the global objectives, which correspond with the optimal selection of the sources and the path to the reservoirs. On the other hand, the lower temporal layer is in charge of manipulating the set-point of the actuators to satisfy the local objectives, i.e., the minimisation of the energy needed for pumping water to the reservoirs. The system decomposition is based on graph partitioning theory, which considers the graph representation of the DWN topology. The obtained system decomposition allows to establish a hierarchical flow of information between the MPC controllers. Hence, the proposed DMPC strategy results in a hierarchical-like scheme. Results obtained when used selected simulation scenarios show the effectiveness of the proposed control strategy in terms of system modularity, reduced computational burden and, at the same time, the admissible loss of performance in contrast to a centralised MPC (CMPC) strategy., This work has been funded by the Spanish Ministry of Science and Technology through the WATMAN project, Ref. DPI2009-13744.

Trabajo presentado al ACC 2010 celebrado en Baltimore (EE.UU.) del 30 de junio al 2 de julio. In this paper, a decentralized model predictive control (DMPC) strategy for drinking water networks (DWN) is proposed. The DWN is partitioned in a set of subnetworks using a partitioning algorithm that makes use of the topology of the network, the information about the actuator usage and heuristics. A suboptimal DMPC strategy was derived that allows the hierarchical solution of the set of MPC controllers used to control each partition. A comparative study between the centralized MPC (CMPC) and DMPC approaches is developed on the case study, which consists in an aggregate version of the Barcelona DWN. Results have shown the effectiveness of the proposed DMPC approach in terms of the computation time while an admissible level of suboptimality is obtained in all the considered scenarios. This work was supported by projects: 'Contribució a la promoció i supervisió de projectes de recerca de l'àmbit de les xarxes de distribució d'aigua' (C-7006), 'Analysis and Design of distributed optimal control strategies applied to large-scale WATer systems MANagement' (4808). This work has been supported by the CICYT DPI 2009 -13744 of the Spanish Science Ministry, the Juan de la Cierva Research Programme (ref. JCI-2008-2438), the DGR of Generalitat de Catalunya (SAC group Ref. 2009/SGR/1491) and the EU project WIDE (FP7-IST-224168).

Control of drinking water networks is an arduous task, given their size and the presence of uncertainty in water demand. It is necessary to impose different constraints for ensuring a reliable water supply in the most economic and safe ways. To cope with uncertainty in system disturbances due to the stochastic water demand/consumption and optimize operational costs, this paper proposes three stochastic model predictive control (MPC) approaches, namely, chance-constrained MPC, tree-based MPC, and multiple-scenario MPC. A comparative assessment of these approaches is performed when they are applied to real case studies, specifically, a sector and an aggregate version of the Barcelona drinking water network in Spain.

Universitat Politècnica de Catalunya (UPC). Departament d’Enginyeria de Sistemes, Automàtica i Informàtica Industrial. Doctoral Program: Automatic Control, Robotics and Computer Vision. Universidad de los Andes (UniAndes). Departamento de Ingeniería Eléctrica y Electrónica. Doctoral Program: Engineering. This PhD thesis was completed at: Institut de Robòtica i Informàtica Industrial, CSIC-UPC, and Universidad de los Andes., [EN]: This thesis is mainly devoted to the study of the role of evolutionary-game theory in the design of distributed optimization-based controllers. Game theoretical approaches have been used in several engineering fields, e.g., drainage wastewater systems, bandwidth allocation, wireless networks, cyber security, congestion games, wind turbines, temperature control, among others. On the other hand, a specific class of games, known as population games, have been mainly used in the design of controllers to manage a limited resource. This game approach is suitable for resource allocation problems since, under the framework of full-potential games, the population games can satisfy a unique coupled constraint while maximizing a potential function. First, this thesis discusses how the classical approach of the population games can contribute and complement the design of optimization-based controllers. Therefore, this dissertation assigns special interest on how the features of the population-game approach can be exploited extending their capabilities in the solution of distributed optimization problems. In addition, density games are studied in order to consider multiple coupled constraints and preserving the non-centralized information requirements. Furthermore, it is established a close relationship between the possible interactions among agents in a population with the constrained information sharing among different local controllers. On the other hand, coalitional games are discussed focusing on the Shapley power index. This power index has been used to assign an appropriate rewarding to players in function of their contributions to all possible coalitions. Even though this power index is quite useful in the engineering context, since it involves notions of fairness and/or relevance (how important players are), the main difficulty of the implementation of the Shapley value in engineering applications is related to the high computational burden. Therefore, this dissertation studies the Shapley value in order to propose an alternative manner to compute it reducing computational time, and a different way to find it by using distributed communication structures is presented. The studied game theoretical approaches are suitable for the modeling of rational agents involved in a strategic constrained interaction, following local rules and making local decisions in order to achieve a global objective. Making an analogy, distributed optimization-based controllers are composed of local controllers that compute optimal inputs based on local information (constrained interactions with other local controllers) in order to achieve a global control objective. In addition to this analogy, the features that relate the Nash equilibrium with the Karush-Kuhn-Tucker conditions for a constrained optimization problem are exploited for the design of optimization-based controllers, more specifically, for the design of model predictive controller. Moreover, the design of non-centralized controllers is directly related to the partitioning of a system, i.e., it is necessary to represent the whole system as the composition of multiple sub-systems. This task is not a, trivial procedure since several considerations should be taken into account, e.g., availability of information, dynamical coupling in the system, regularity in the amount of variables for each sub-system, among others. Then, this doctoral dissertation also discusses the partitioning problem for large-scale systems and the role that this procedure plays in the design of distributed optimization-based controllers. Finally, dynamical partitioning strategies are presented with distributed population-games-based controllers. Some engineering applications are presented to illustrate and test the performance of all the proposed control strategies, e.g., the Barcelona water supply network, multiple continuous stirred tank reactors, system of multiple unmanned aerial vehicles., [ES]: Esta tesis doctoral consiste principalmente en el estudio del rol que desempeña la teoría de juegos evolutiva en el diseño de controladores distribuidos basados en optimización. Diversos enfoques de la teoría de juegos han sido usados en múltiples campos de la ingeniera, por ejemplo, en sistemas de drenaje urbano, para la asignación de anchos de banda, en redes inalámbricas, en ciber-seguridad, en juegos de congestión, turbinas eólicas, control de temperatura, entre otros. Por otra parte, una clase especifica de juegos, conocidos como juegos poblacionales, se han usado principalmente en el diseño de controladores encargados de determinar la apropiada asignación de recursos. Esta clase de juegos es apropiada para problemas de distribución dinámica de recursos dado que, en el contexto de juegos poblacionales, los juegos poblacionales pueden ser usados para maximizar una función potencial mientras se satisface una restricción acoplada. Primero, esta tesis doctoral presenta como el enfoque clásico de los juegos poblacionales pueden contribuir y complementar en el diseño de controladores basados en optimización. Posteriormente, esta disertación concentra su atención en cómo las características de los juegos poblacionales pueden ser aprovechadas y extendidas para dar solución a problemas de optimización de forma distribuida. Adicionalmente, los juegos con dependencia de densidad son estudiados con el fin de considerar múltiples restricciones mientras se preservan las características no centralizadas de los requerimientos de información. Finalmente, se establece una estrecha relación entre las posibles interacciones de los agentes en una población y las restricciones de intercambio de información entre diversos controladores locales. También, se desarrolla una discusión sobre los juegos cooperativos y el índice de poder conocido como el valor de Shapley. Este índice de poder ha sido usado para la apropiada asignación de beneficios para un jugador en función de sus contribuciones a todas las posibles coaliciones que pueden formarse. Aunque este índice de poder es de gran utilidad en el contexto ingenieril, ya que involucra nociones de justicia y/o relevancia, la principal dificultad para implementar el valor de Shapley en aplicaciones de ingeniería está asociado a los altos costos computacionales para encontrarlo. En consecuencia, esta disertación doctoral estudia el valor de Shapley con el fin de ofrecer una alternativa para calcular este índice de poder reduciendo los costos computacionales e incluso contemplando estructuras distribuidas de comunicación. Los enfoques de la teoría de juegos estudiados son apropiados para el modelamiento de agentes racionales involucrados en una interacción estratégica con restricciones, siguiendo reglas locales y tomando decisiones locales para alcanzar un objetivo global. Realizando una, analogía, los controladores distribuidos basados en optimización están compuestos por controladores locales que calculan acciones óptimas basados en información local (considerando interacciones restringidas con otros controladores locales) con el fin de alcanzar un objetivo global. Adicional a esta analogía, las características que relacionan el equilibrio de Nash con las condiciones de Karush-Kuhn-Tucker en un problema de optimizaciones con restricciones son aprovechadas para el diseño de controladores basados en optimización, más específicamente, para el diseño de controladores predictivos. Por otra parte, el diseño de controladores no centralizados está directamente relacionado con el particionado de un sistema, es decir, es necesario representar el sistema en su totalidad por medio del conjunto de varios sub-sistemas. Esta tarea no es un procedimiento trivial puesto que es necesario tener en cuenta varias consideraciones, por ejemplo, la disponibilidad de información, el acople dinámico en el sistema, la regularidad en cuanto a la cantidad de variables en cada sub-sistema, entre otras. Por lo tanto, esta disertación doctoral también desarrolla una discusión alrededor del problema de particionado para sistemas de gran escala y respecto al rol que este procedimiento de particionado juega en el diseño de controladores distribuidos basados en optimización. Finalmente, se presentan estrategias de particionado dinámico junto con controladores basados en juegos poblacionales. Algunas aplicaciones en ingeniería son usadas para ilustrar y probar los controladores diseñados por medio de las contribuciones novedosas basadas en teoría de juegos, estas son, la red de agua potable de Barcelona, múltiples reactores, sistema compuesto por varios vehículos aéreos no tripulados y un sistema de distribución de agua., [CA]: Aquesta tesi doctoral consisteix principalment en l'estudi del paper que exerceix la teoria de jocs evolutiva en el disseny de controladors distribuïts basats en optimització. Diversos enfocaments de la teoria de jocs han estat usats en múltiples camps de l'enginyeria, per exemple, en sistemes de drenatge urbà, per a l’assignació d'amples de banda, en xarxes sense fils, a ciber-seguretat, en jocs de congestió, turbines eòliques, control de temperatura, entre altres. D'altra banda, una classe especifica de jocs, coneguts com jocs poblacionals, s'han fet servir principalment en el disseny de controladors encarregats de determinar l'apropiada assignació de recursos. Aquesta classe de jocs és apropiada per a problemes de distribució dinàmica de recursos atès que, en el context de jocs poblacionals, aquests poden ser usats per a maximitzar una funció potencial mentre es satisfà una restricció acoblada. Primer, aquesta tesi doctoral presenta com l'enfocament clàssic dels jocs poblacionals poden contribuir i complementar en el disseny de controladors basats en optimització. Posteriorment, aquesta dissertació concentra la seva atenció en com les característiques dels jocs poblacionals poden ser aprofitades i esteses per donar solució a problemes d’optimització de forma distribuïda. Addicionalment, els jocs amb dependència de densitat són estudiats amb la _finalitat de considerar múltiples restriccions mentre es preserven les característiques no centralitzades dels requeriments d’informació. Finalment, s'estableix una estreta relació entre les possibles interaccions dels agents en una població i les restriccions d'intercanvi d’informació entre diversos controladors locals. També, es desenvolupa una discussió sobre els jocs cooperatius i l’índex de poder conegut com el valor de Shapley. Aquest índex de poder ha estat usat per l'apropiada assignació de beneficis per a un jugador en funció de les seves contribucions a totes les possibles coalicions que poden formar-se. Encara que aquest índex de poder es de gran utilitat en el context de l'enginyeria, ja que involucra nocions de justícia i/o rellevància, la principal dificultat per implementar el valor de Shapley en aplicacions d'enginyeria està associat als alts costos computacionals per trobar-lo. En conseqüència, aquesta dissertació doctoral estudia el valor de Shapley per tal d'oferir una alternativa per calcular aquest índex de poder reduint els costos computacionals i fins i tot contemplant estructures distribuïdes de comunicació. Els enfocaments de la teoria de jocs estudiats són apropiats per al modelatge d'agents racionals involucrats en una interacció estratègica amb restriccions, seguint regles locals i prenent decisions locals per assolir un objectiu global., Realitzant una analogia, els controladors distribuïts basats en optimització estan compostos per controladors locals que calculen accions optimes basats en informació local (considerant interaccions restringides amb altres controladors locals) per tal d'assolir un objectiu global. Addicional a aquesta analogia, les característiques que relacionen l'equilibri de Nash amb les condicions de Karush-Kuhn-Tucker en un problema d’optimització amb restriccions són aprofitades per al disseny de controladors basats en optimització, més específicament, per al disseny de controladors predictius. D'altra banda, el disseny de controladors no centralitzats està directament relacionat amb la partició d'un sistema, és a dir, cal representar el sistema en la seva totalitat per mitjà del conjunt de diversos sub-sistemes. Aquesta tasca no és un procés trivial, ja que cal tenir en compte diverses consideracions, per exemple, la disponibilitat d’informació, l'acoblament dinàmic en el sistema, i la regularitat pel que fa a la quantitat de variables en cada sub-sistema, entre d'altres. Per tant, aquesta dissertació doctoral també desenvolupa una discussió al voltant del problema de partició per a sistemes de gran escala i respecte al paper que aquest procediment de partició juga en el disseny de controladors distribuïts basats en optimització. Finalment, es presenten estratègies de partició dinàmic juntament amb controladors basats en jocs poblacionals. Algunes aplicacions en enginyeria són usades per il·lustrar i provar els controladors dissenyats per mitjà de les contribucions noves basades en teoria de jocs, aquestes són: la xarxa d'aigua potable de Barcelona, múltiples reactors, sistema compost per diversos vehicles aeris no tripulats i un sistema de distribució d'aigua., I sincerely thank Agència de Gestió d’Ajust Universitaris i de Recerca AGAUR for having granted the FI scholarship for me to study the doctoral degree at UPC, and Colciencias for having granted the scholarship/condonable loan (Ref. 6172) for me to study the doctoral degree at UniAndes. I would also like to thank the projects ECOCIS: EConomic Operation of Critical Infrastructure Systems (Ref. DPI2013-48243- C2-1-R), EFFINET: EFFicient Integrated real-time monitoring and control of drinking water NETworks (Ref. FP7-ICT-2011-8-318556), ModCon: Modelado y control de redes de alcantarillado: aplicación en Bogotá (Colombia), and “Drenaje urbano y cambio climático: hacia los sistemas de alcantarillado del futuro. Fase II”, Mexichem, Colombia for partially supporting me.

Master’s Thesis in Automatic Control and Robotics made at: Universitat Politècnica de Catalunya - Escola Tècnica Superior d’Enginyeria Indistrial de Barcelona., [EN]: Urban drainage systems (UDSs) are complex large-scale systems that carry stormwater and wastewater throughout urban areas. During heavy rain scenarios, UDSs are not able to handle the amount of extra water that enters the network and flooding occurs. Usually, this might happen because the network is not being used efficiently, i.e., some structures remain underused while many others are overused. This thesis proposes a control methology based on mean field game theory and model predictive control that aims to efficiently use the existing network elements in order to minimize overflows and properly manage the water resource. The proposed controller is tested on a UDS located in the city of Barcelona, Spain, and is compared with a centralized MPC achieving similar results in terms of flooding minimization and wastewater treatement plant usage, but only using local information on non-centralized controllers and using less computation times., [ES]: Los sistemas de drenaje urbano son sistemas complejos y de gran escala que transportan tanto aguas negras como aguas de lluvia a través de las zonas urbanas. Durante eventos de precipitación muy fuertes, estos sistemas no son capaces de manejar la cantidad de agua adicional que entra y ocurren inundaciones indeseadas. Normalmente, esto pasa porque la red no está siendo utilizada eficientemente, i.e., algunas estructuras permanecen sub-utilizadas y otras se encuentran sobre-utilizadas. Esta tesis propone un esquema de control basado en la teoría de mean field games y de control predictivo que busca utilizar eficientemente la red para minimizar las inundaciones en cualquier escenario, y garantizar una utilización adecuada del recurso hídrico. El esquema propuesto es probado en la red de drenaje de la Riera Blanca en la ciudad de Barcelona y es comparado con una estrategia de control basada puramente en control predictivo, obteniendo resultados similares en terminos de inundación y utilización de plantas de tratamiento, pero utlizando únicamente una fracción del costo computacional., [CA]: Sistemes de drenatge urbà (UDS) sòn sistemes complexos a gran escala que transporten aigües pluvials i residuals en les zones urbanes. Durant escenaris de fortes pluges, UDSs no són capaços de manejar la quantitat d’aigua addicional que entra a la xarxa i es produeix la inundació. En general, això pot succeir perquè la xarxa no està sent utilitzat de manera eficient, ès a dir, algunes estructures segueixen sent infrautilitzades mentre que molts altres són usats en excés. En aquest treball es proposa una metodologia de control basat en la teoria de mean field games i control predictiu que pretén utilitzar de forma eficient els elements de xarxa existents per tal de minimitzar els desbordaments i gestionar adequadament els recursos hídrics. El controlador proposat es prova en un UDS situada a la ciutat de Barcelona, Espanya, i es compara amb un MPC centralitzat aconsegint resultats similars en termes de minimització de les inundacions i la utilització de la planta de tractament d’aigües residuals, però només utilitzant informació local en els controladors no centralitzats i l’ús de menys temps de càlcul.

A dissertation submitted for the degree of Master in Automatic Control and Robotics in Departament d’Enginyeria de Sistemes, Automatica i Informàtica Industrial (Escola Tècnica Superior d’Enginyeria Industrial de Barcelona.-- Master’s Thesis made at: Institut de Robòtica i Informàtica Industrial (CSIC-UPC)., This thesis firstly proposes a new approach of modelling an ethanol steam reformer (ESR) for producing pure hydrogen. Hydrogen has obvious benefits as an alternative for feeding the proton exchange membrane fuel cells (PEMFCs) to produce electricity. However, an important drawback is that the hydrogen distribution and storage have high cost. So the ESR is regarded as a way to overcome these difficulties. Ethanol is currently considered as a promising energy source under the research for its renewable and easy-access. The modelled reformer has been divided into two stages in order to miximize the output of pure hydrogen. Specifically, the hydrogen is generated in the first stage by steam reforming (SR) over the Cobalt-based catalysts. The second stage is used to purify the hydrogen from the mixture gas through a Pd-Ag membrane separation. Unlike the conventional way, the two stages are carried out in a same module, since only one heating device is required. The main idea of the modelling approach is to develop a one-dimensional, pseudo-homogeneous and non-steady-state model based on the mass balance. Finite difference method is applied to transform the partial differential equations (PDEs) into ordinary differential equations (ODEs). Furthermore, the control of the ESR system is also discussed in this thesis. Some Model predictive control (MPC) strategies have been designed for the linear and non-linear models of the ESR. The mathematical model of the ESR by using the proposed approach is used for the design of a MPC controller. The first part of this thesis has been devoted to obtain the non-linear mathematical model with the assumptions of isothermal and isobaric conditions and perform simulations. The numerical linearisation of the mathematical model towards a controloriented model has been performed as well as the method of model-order reduction with Hankel norm approximation and temporal discretization with zero-order hold, which will be used as the prediction model of the MPC. The control objectives of the ESR system are to maximize the hydrogen yield and minimize the ethanol flow rate, which result in a multi-objective problem. Open-loop and close-loop simulations are developed to highlight the proposed controller results. The analysis has taken into account relevant indices such as the ethanol conversion,input smoothness and output error.

This thesis is devoted to developing a robust Model Predictive Control (MPC) strategy based on Gaussian Processes (GP), especially for Drinking Water Networks (DWN). Nowadays there are many different MPC strategies developed for DWN, such as certain-equivalent MPC (CEMPC) and chance-constrained MPC (CC-MPC). The general control objectives for DWN are economic by managing the water supply to minimise water production and transport costs, all the tanks running in safe ways with their limitations and reducing the undesired abrupt changes by minimising their slew rate and obtaining smooth signals. For the deterministic system model, the control objectives are elementary fulfilled. But the main challenge for DWN is to propagate and incorporate exogenous and endogenous uncertainties to MPC closed loop over the prediction horizon. Considering the control-oriented model of the DWN, the MPC controller design is hereby divided into two parts: system disturbances forecasting and the robust MPC controller design. Case studies based on Barcelona DWN have been executed to verify the performance of proposed methodologies. The first part of this thesis leads to forecast system disturbances by using GP. In a DWN system, system disturbances come mainly water demands associated to consumer sectors. Hence, it is necessary to model each water demand and forecast the water demand in a short term that covers the MPC prediction horizon. GP regression is regarded as one of state-of-the-art regression methods able to select model parameters by using Bayesian Inference theory with a collection of past data. Besides, it is believed that the GP regression method has a difficult for the multiple-step ahead forecasting. Hence, the Double-seasonal Holt-winters method is used for forecasting the expected disturbances while the stochastic disturbances are forecasted by using GP. Finally, the desired forecasting results are a set of Gaussian distributions over the MPC prediction horizon. The second part of this thesis is to incorporate the forecasting results from GP within MPC closed loop. This MPC strategy based on GP is named GP-MPC. Using the given system model, the deterministic state evolutions can be obtained while the uncertainty of state propagation over a given prediction horizon can be also achieved though the linear approximation of GP. Therefore, the worst-case state evolutions over the MPC prediction horizon can also be determined in the MPC cost function and constraints. The desired performance of applying GP-MPC in the closed-loop system is that the system has more safety than the CE-MPC and meanwhile it probably brings more expenses. Comparisons of GP-MPC and previous developed approaches are carried out by a case study of the three-tank system inspired in the Barcelona DWN. A set of key performance indicators are defined to compare performances of different MPC strategies. Finally, through the simulation results, the GP-MPC has the similar performance as the CC-MPC, both of which have much more expenses than the CE-MPC. As a result of considering the uncertainties inside the system, more expenses is necessary to maintain the safety of the whole system. Hence, the GP-MPC is more advanced. Moreover, the proposed GP-MPC is required to be tested with the whole DWN and using the real data from a DWN system. So the future works of this thesis have been outlined.