Skip to main content

NUEVAS ESTRATEGIAS DE OPTIMIZACION MULTIOBJETIVO PARA LA IDENTIFICACION, CONTROL Y DISEÑO DE PROCESOS Y SISTEMAS

DPI2008-02133

•
Funding agency name Ministerio de Ciencia e Innovación
Funding agency acronym MICINN
Program Programa Nacional de Proyectos de Investigación Fundamental
Subprogram Subprograma de Proyectos de Investigación Fundamental no Orientada
Call Proyectos I+D
Call year 2008
State Plan 2008-2011 Plan Nacional de Investigación Científica, Desarrollo e Innovación Tecnológica
Management unit Subdirección General de Proyectos de Investigación
Beneficiary centre UNIVERSITAT POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 10
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Estudio comparativo de algoritmos de auto-ajuste de controladores PID. Resultados del Benchmark 2010-2011 del Grupo de Ingeniería de Control de CEA, Comparative Study of Auto-tuning Algorithms for PID Controllers. Results of the 2010-2011 Match of the Control Engineering Group of CEA

  • Romero Pérez, J.A.
  • Arrieta, O.
  • Padula, F.
  • Reynoso Meza, Gilberto
  • Balaguer, P.
  • Garcia-Nieto, Sergio|||0000-0002-2722-742X
[EN] In this paper three PID auto-tuning algorithms which are mainly focused on the disturbance rejection problem are compared. The algorithms differ in both the experiments carried out to obtain information of the process dynamic and the methods for calculating the controller parameters. The algorithms were presented in the 2010-2011 Match of the Control Engineering Group of CEA. The comparison is based on an evaluation methodology that takes into account the experimental phase as well as the closed loop performance during the control phase., [ES] En este artículo se comparan tres métodos de auto-ajuste de controladores PID que consideran diferentes tipos de experimentos para obtener información de la dinámica del proceso y distintos métodos de cálculo de los parámetros del controlador para minimizar el efecto de las perturbaciones. Los métodos fueron presentados en el concurso anual organizado por el grupo de Ingeniería de Control de CEA-IFAC del curso 2010-2011. Para la comparación se aplica la metodología de evaluación de algoritmos de autoajuste que tiene en cuenta tanto la fase de experimento como las prestaciones que se consiguen en la fase de control., Este trabajo ha sido parcialmente financiado por el Gobierno Español mediante los proyectos con referencia DPI2008-02133 y DPI2010-15230, la Fundación Caixa Castellón-Bancaixa y la Universitat Jaume I a través de proyecto con referencia P1-1A2010-16, la Universidad Politécnica de Valencia mediante el programa de becas FPI-2010/19. Además, es muy apreciado el apoyo de la Universidad de Costa Rica, el MICIT, y el CONICIT del Gobierno de la República de Costa Rica.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Latent variable based model predictive control: Ensuring validity of predictions

  • Laurí Pla, David
  • Sanchís Saez, Javier|||0000-0001-9697-2696
  • Martínez Iranzo, Miguel Andrés|||0000-0002-1444-0651
  • Hilario Caballero, Adolfo|||0000-0002-3237-8652
This paper presents a methodology to constrain the optimisation problem in LV-MPC so that validity of predictions can be ascertained. LV-MPC is a model-based predictive control methodology implemented in the space of the latent variables and is based on a linear predictor. Provided real processes are non-linear, there is model-process mismatch, and under tight control, the predictor can be used for extrapolation. Extrapolation leads to bad predictions which deteriorates control performance, hence the interest in validity of predictions. In the proposed approach first two validity indicators on predictions are defined. The novelty in the two indicators proposed is they neglect past data, and so validity of predictions is ascertained in terms of future moves which are actually the degrees of freedom in the optimisation. Second, the indicators are introduced in the optimisation as constraints. Provided the indicators are quadratic, recursive optimisation with linearised constraints is implemented. A MIMO example shows how ensuring validity of predictions neglecting past data can improve closed-loop performance, specially under tight control outside the identification region. (C) 2012 Elsevier Ltd. All rights reserved., The first author is recipient of a fellowship from the Spanish Ministry of Science and Innovation (FPU AP2007-04549). This paper is partially funded by projects DPI2008-02133/DPI, TIN2011-28082 and PROMETEO/2012/028. The authors gratefully acknowledge reviewers' comments.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Contribución de congreso (ConferenceOutput)

Multiobjective Tuning of Robust GPC Controllers using Evolutionary Algorithms

  • Herrero Durá, Juan Manuel|||0000-0003-1914-7494
  • Blasco, Xavier|||0000-0002-9737-2833
  • Martínez Iranzo, Miguel Andrés|||0000-0002-1444-0651
  • Sanchís Saez, Javier|||0000-0001-9697-2696
[EN] In this article a procedure to tune robust Generalized Predictive Controllers (GPC) is presented. To tune the controller parameters a multiobjective optimization problem is formulated so the designer can consider conflicting objectives simultaneously without establishing any prior preference. Moreover model uncertainty, represented by a list of possible models, is considered. The multiobjective problem is solved with a specifc Evolutionary Algorithm (ev-MOGA). Finally, an application to a non-linear thermal process is presented to illustrate the technique., Partially funded by GVPRE/2008/326 and DPI2008-02133/DPI.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Practice tool based on open source SCADA for experimentation in nonlinear control using the inverted pendulum

  • Garcia-Nieto, Sergio|||0000-0002-2722-742X
  • Martínez Iranzo, Miguel Andrés|||0000-0002-1444-0651
  • Salcedo-Romero-de-Ávila, José-Vicente|||0000-0003-1577-5039
  • Laurí Pla, David
[EN] This paper presents the potential of open source software for designing educational tools in the automatization field. In particular, this paper presents a complete tool that students can use for studying and testing nonlinear control algorithms. The system has three different parts that students can evaluate and modify. Firstly, a virtual model represents the physical model and, in this case, an inverted pendulum is used. Secondly, the controller is implemented by a real-time distributed control system. Finally, the system can be managed with a JAVA application. Therefore, students have all the necessary elements to practice using nonlinear and complex systems. The main tools applied in the design are open source software and the developed platform is Generalized Public License(GPL). © 2009 Wiley Periodicals, Inc., This work was partially funded by GVPRE/2008/326 and DPI2008-02133/DPI. We would like to thank John Rawlins for revising the English text.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Solución Explícita al Control Predictivo de Sistemas Lineales Sujetos a Restricciones No Convexas, Explicit Solution to Predictive Control of Linear Systems Subject to Non-convex Constraints

  • Pérez, Emilio
  • Ariño, Carlos
  • Blasco, Xavier|||0000-0002-9737-2833
  • Martínez Iranzo, Miguel Andrés|||0000-0002-1444-0651
[EN] This paper presents an explicit solution to predictive control of linear systems subject to non-convex polyhedral constraints. The constraints are modeled as the union of finite number of polyhedra. The algorithm is based on the computation of the explicit solution of the problems subject to convex constraints defined by those polyhedra. The regions of the partitions obtained this way are intersected such that the new set of regions has as many possible solutions as convex problems are solved. By means of sum-of-squares programming, the non-optimal solutions are removed. Then a union of regions which share the same set of solutions is performed. Furthermore a suboptimal solution which can be used when the original methodology is too expensive is presented. Finally, some results are shown in an example., [ES] Este trabajo propone una solución explícita para el control predictivo de sistemas lineales sujetos a restricciones poliédricas no convexas, modeladas como la unión de un número finito de poliedros. El algoritmo se basa en el cálculo de la solución explícita de los problemas sujetos a las restricciones convexas definidas por dichos poliedros. Las regiones de las particiones así obtenidas se intersectan de forma que el nuevo conjunto de regiones tiene tantas soluciones posibles como problemas convexos se han resuelto. Mediante programación de suma de cuadrados se eliminan aquellas soluciones de cada región que no son óptimas para ningún estado. Posteriormente se realiza la unión de las regiones que compartan el mismo conjunto de soluciones. Tras la descripción de la metodología descrita, se incluye una justificación de ésta. Además, se incluye una posible solución subóptima utilizable cuando la metodología original es demasiado costosa. Por último, se muestran los resultados obtenidos en un ejemplo., Este trabajo ha sido realizado parcialmente gracias al apoyo de los proyectos DPI2008-02133 y DPI2008-06731/DPI Ministerio de Ciencia e Innovación.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Multiobjective evolutionary algorithms for multivariable PI controller design

  • Reynoso Meza, Gilberto
  • Sanchís Saez, Javier|||0000-0001-9697-2696
  • Blasco, Xavier|||0000-0002-9737-2833
  • Herrero Durá, Juan Manuel|||0000-0003-1914-7494
A multiobjective optimisation engineering design (MOED) methodology for PI controller tuning in multivariable processes is presented. The MOED procedure is a natural approach for facing multiobjective problems where several requirements and specifications need to be fulfilled. An algorithm based on the differential evolution technique and spherical pruning is used for this purpose. To evaluate the methodology, a multivariable control benchmark is used. The obtained results validate the MOED procedure as a practical and useful technique for parametric controller tuning in multivariable processes., This work was partially supported by the FPI-2010/19 grant and the project PAID-06-11 from the Universitat Politecnica de Valencia and the projects DPI2008-02133, TIN2011-28082 and ENE2011-25900 from the Spanish Ministry of Science and Innovation.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Controller tuning by means of multi-objective optimization algorithms: a global tuning framework

  • Reynoso Meza, Gilberto
  • Garcia-Nieto, Sergio|||0000-0002-2722-742X
  • Sanchís Saez, Javier|||0000-0001-9697-2696
  • Blasco, Xavier|||0000-0002-9737-2833
A holistic multi-objective optimization design technique for controller tuning is presented. This approach gives control engineers greater flexibility to select a controller that matches their specifications. Furthermore, for a given controller it is simple to analyze the tradeoff achieved between conflicting objectives. By using the multi-objective design technique it is also possible to perform a global comparison between different control strategies in a simple and robust way. This approach thereby enables an analysis to be made of whether a preference for a certain control technique is justified. This proposal is evaluated and validated in a nonlinear multiple-input multiple-output system using two control strategies: a classical proportional- integral-derivative control scheme and a feedback state controller., This work was supported in part by the FPI-2010/19 Grant and the Project PAID-06-11 from the Universitat Politecnica de Valencia and in part by the Projects DPI2008-02133, TIN2011-28082, and ENE2011-25900 from the Spanish Ministry of Science and Innovation.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Explicit predictive control with non-convex polyhedral constraints

  • Pérez Soler, Emilio
  • Ariño Latorre, Carlos Vicente
  • Blasco, Xavier|||0000-0002-9737-2833
  • Martínez Iranzo, Miguel Andrés|||0000-0002-1444-0651
This paper proposes an explicit solution to the model predictive control of linear systems subject to non-convex polyhedral constraints. These constraints are modeled as the union of a finite number of convex polyhedra. The algorithm is based on calculating the explicit solution to a modified problem with linear constraints defined as the convex hull of the original ones and classifying its regions by their relation with the regions of the explicit solution to the original problem. Some of the regions are divided, and a procedure based on sum-of-squares programming is designed to determine which of the possible solutions are in fact optimal. Finally, the online algorithm is shown to be better in terms of computational cost and memory requirements than an algorithm based on obtaining and comparing the solutions of the problem using as constraints the polyhedra whose union forms the non-convex regions, both theoretically and by the results of an example, This work was partially funded by projects DPI2008-02133 and DPI2008-06731-C02-01(and -02)/DPI Ministerio de Ciencia e Innovacion-Spanish Government.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Diseño Multiobjetivo de controladores PID para el Benchmark de Control 2008-2009, Multiobjective Design of PID controllers for the 2008-2009 Control Benchmark

  • Reynoso Meza, Gilberto
  • Blasco, Xavier|||0000-0002-9737-2833
  • Sanchís Saez, Javier|||0000-0001-9697-2696
[ES] En el presente trabajo se emplea una optimización multiobjetivo para proponer un conjunto de controladores digitales para el Benchmark de Ingeniería de Control del grupo Temático de la IFAC-CEA, de manera que cumpla con todos sus requerimientos y restricciones impuestas por el comité organizador. La estrategia multiobjetivo permite determinar el mejor conjunto de soluciones que caracteriza al frente de Pareto, compuesto por soluciones no-dominadas donde cada una es óptima y solo difiere del resto en el grado de compromiso y prestaciones que ofrece con respecto a los objetivos de optimización., [EN] In this work a multiobjective optimization approach is used to propose a set of digital controllers for the Benchmark presented by the IFAC-CEA Thematic Group of Control Engineering, fulfilling all its restrictions and requirements. The multiobjective strategy allows to determine the best set of solutions that represents which is known as the Pareto front, where all the solutions are non-nominated, optimal and each solution differs from others only by its degree of compromise with the given optimization objectives., Este trabajo es apoyado parcialmente por PDI2008-02133/PDI, Ministerio de Ciencia e Innovación.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publication Artículo científico (JournalArticle)

Maximal closed loop admissible set for linear systems with non-convex polyhedral constraints

  • Pérez Soler, Emilio
  • Ariño Latorre, Carlos Vicente
  • Blasco, Xavier|||0000-0002-9737-2833
  • Martínez Iranzo, Miguel Andrés|||0000-0002-1444-0651
[EN] This paper deals with the computation of the maximal closed-loop admissible set for linear systems with non-convex polyhedral constraints. These constraints are modeled as the union of a finite number of convex polyhedra. An efficient algorithm for the computation of this set, based on removing subsets of the maximal closed-loop invariant set for the convex hull of the original constraints, is proposed and compared with the existing generic algorithm. Next, stability conditions for a general predictive control scheme are applied to the particular problem proposed. Finally, two examples showing the performance of the maximal admissible set algorithm are given., This work was partially funded by projects DPI2008-02133 and DPI2008-06731/DPI Ministerio de Ciencia e Innovación – Spanish Government.