• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • 1
  • Tagged with
  • 2
  • 2
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

New Structure for Moving Horizon Estimators. Application to Space Debris Tracking during the Atmospheric Re-entries / Nouvelle Structure d’Estimateurs à Horizon Glissant. Application à l’Estimation de Trajectoires de Débris Spatiaux pendant la Rentrée Atmosphérique

Suwantong, Rata 02 December 2014 (has links)
L’estimation de trajectoires de débris spatiaux pendant la rentrée atmosphérique est un défi majeur pour les prochaines années, renforcé par plusieurs projets liés à l'enlèvement de débris établis par plusieurs agences spatiales. Cependant, ce problème s’avère complexe du fait des erreurs de modèle et des difficultés d’initialisation des algorithmes d’estimation induites par une mauvaise connaissance de la dynamique des débris suite à leur désintégration pendant la phase de rentrée atmosphérique. Tout estimateur choisi doit donc être robuste vis-à-vis de ces facteurs. L’estimateur à horizon glissant (MHE) est reconnu dans la littérature pour être robuste vis-à-vis d’erreurs de modèle et de mauvaise initialisation, et les travaux de thèse ont montré qu’il était adapté en termes de performances à la problématique de l’estimation des débris en phase de rentrée. En revanche, il se fonde sur une stratégie d’optimisation qui requiert de fait un temps de calcul important. Pour pallier ce problème, une nouvelle structure d’estimation à horizon glissant a été développée, impliquant un temps de calcul faible nécessaire à l’application envisagée. Cette stratégie, appelée « estimateur à horizon glissant avec pré-estimation (MHE-PE)», prend en compte les erreurs de modèle via un estimateur auxiliaire, plutôt que de chercher à obtenir les estimées du bruit d’état sur l’horizon d’estimation, comme le fait la structure de l’estimateur MHE standard. Un théorème garantissant la stabilité de la dynamique de l’erreur d’estimation du MHE-PE a par ailleurs été proposé. Enfin, les performances de cette structure dans le cadre de l’estimation en trois dimensions des trajectoires de débris pendant la phase de rentrée se sont avérées meilleures que celles observées avec des estimateurs classiques. En particulier, sans dégrader la précision et la convergence de l’estimation, l’estimateur MHE-PE requiert moins de temps de calcul du fait du nombre réduit de paramètres à optimiser. / Space debris tracking during atmospheric re-entries will be a crucial challenge in the coming years, emphasized through many projects on space debris mitigation established by space agencies worldwide. However, this problem appears to be complex, due to model errors and difficulties to properly initialize the estimation algorithms, as a result of unknown dynamics of the debris and their disintegrations during the re-entries. A-to-be used estimator for this problem must be robust against these factors. The Moving Horizon Estimator (MHE) is known in the literature to be robust to model errors and bad initialization, and the PhD work has proved its ability to satisfy performances required by the debris tracking during the re-entries. However, its optimization-based framework induces a large computation time. To overcome this, a new MHE structure which requires smaller computation time than the classical MHE has been developed. This strategy, so-called “Moving Horizon Estimator with Pre-Estimation (MHE-PE)” takes into account model errors by using an auxiliary estimator rather than by searching for estimates of the process noise sequence over the horizon as in the classical strategy. A theorem which guarantees the stability of the dynamics of the estimation errors of the MHE-PE has also been proposed. Finally, performances of this structure in the context of 3D space debris tracking during the re-entries have been shown to be better than those obtained with classical estimators including the MHE. In particular, without degrading accuracy of the estimates and convergence of the estimator, the MHE-PE estimator requires smaller computation time than the MHE thanks to its small number of optimization variables.
2

Reconciliação dinâmica de dados baseada em estimadores em uma malha de controle MPC / Dynamic data reconciliation based on estimators in a MPC control loop

Silva, Guilherme Moura Afonso da 27 April 2017 (has links)
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The data reconciliation in process control is extremely important regarding the industries because from this it is possible to obtain a greater efficiency in the performance in industrial process control meshes aiming at a lower cost and a higher quality of the product. In this work we approach data estimation techniques for the implementation of an online dynamic data reconciliation system in order to reduce the noise and the measurement uncertainties that are submitted in the process variables. The techniques used here are: the Kalman Filter, the Preditor-Corrector DDR Algorithm, the Moving Horizon Estimator (MHE) and the Constrained Extended Kalman Filter (CEKF). The analysis is performed by applying the dynamic data reconciliation system in a simulated process, characteristic of the chemical industry, operating under MPC (Model Predictive Control). The performance of the MPC controller is also enhanced by the use of the reconciled data in the feedback control loop. / A reconciliação de dados em controle de processos é extremamente importante no que diz respeito às indústrias, pois a partir dessa é possível obter uma maior eficiência no desempenho em malhas de controle de processos industriais visando à minimização dos custos e maximizando a qualidade do produto. Neste trabalho abordam-se técnicas de estimação de dados para a implementação de um sistema de reconciliação dinâmica de dados on-line a fim de reduzir os ruídos e as incertezas de medições a que estão submetidas às variáveis do processo. As técnicas aqui empregadas são: o Filtro de Kalman, o Algoritmo DDR Preditor-Corretor, o Estimador de Horizonte Móvel (MHE) e o Filtro de Kalman Estendido com Restrições (CEKF). As análises são efetuadas aplicando o sistema de reconciliação dinâmica de dados em um processo simulado, característico da indústria química, operando sob controle preditivo (MPC). Também é efetuado o aprimoramento no desempenho do controlador MPC utilizando os dados reconciliados na malha de realimentação do controlador.

Page generated in 0.094 seconds