La commande des actionneurs électropneumatiques a été un sujet largement traité au cours des dernières décennies. Le caractère fortement non-linéaire de son comportement en a fait un cas d'étude particulièrement pertinent dans le cadre d'une démarche d'application de la théorie de la commande des systèmes non-linéaires. L'utilisation de ces techniques a été comparée aux approches linéaires traditionnelles et généralement jugée largement supérieure notamment en termes de précision ou de temps de réponse. Dans ce manuscrit nous abordons très spécifiquement l'aspect multivariable du système et introduisons la transformée A-T, similaire à la transformée de Park appliquée classiquement aux systèmes électriques, afin de donner une forme strict feedback à son modèle d'état, de clarifier les phénomènes physiques mis en jeu lors de sa commande et de distinguer les deux degrés de liberté du système. Cette transformée permet en outre une comparaison directe avec les moteurs électriques décrits dans le repère de Park. Ce parallèle rend notamment possible la solution du problème délicat de l'observation de la position à vitesse nulle en transférant des méthodologies déjà validées sur des systèmes électriques. L'exploitation des deux degrés de liberté est illustrée par la synthèse de lois de commande combinant le suivi d'une trajectoire de position du piston au respect d'un second critère variable (réglage de la pressurisation moyenne, optimisation de la consommation instantanée). L'utilisation d'un actionneur électropneumatique asservi comme actionneur à compliance variable est étudiée. Une loi de commande basée sur la transformée A-T est proposée pour contrôler simultanément la position et la raideur pneumatique de l'actionneur. Une méthodologie de réglage des gains de commande est proposée pour définir l'impédance en boucle fermée du système. L'influence de la raideur pneumatique sur la raideur en boucle fermée est étudiée. L'utilisation d'une source d'énergie alternative (de l'hélium sous pression) est également pour la première fois mise en œuvre. L'influence du changement de gaz sur le dimensionnement de l'actionneur électropneumatique est étudiée et une méthodologie permettant d'utiliser les lois de commande prévues pour de l'air est proposée. L'ensemble des propositions faites dans ce manuscrit est testée et validée sur un banc d'essais à la structure inédite. Ce dernier allie deux actionneurs, l'un électropneumatique (l'actionneur étudié) et l'autre, un moteur plat électrique (l'actionneur de charge). L'utilisation de celui-ci permet la génération d'efforts perturbateurs dans une large bande passante ainsi que la modification en temps réel des paramètres mécaniques dynamiques de la charge. / Control of electropneumatic actuator is a subject which has been thoroughly dealt with in the past decades. The strongly non-linear behavior of those systems made them a particularly relevant case study for applications of the non linear control theory. The use of those techniques has been compared to the traditional linear approaches and is generally considered as more efficient in terms of precision and response time. In this work, we specifically deal with the multivariable character of this actuator and we introduce the A-T transformation, similar to the Park Transformation used for electrical systems study, in order to give a strict feedback form to its state model, clarify the physical phenomena occurring during its control and distinguish the system's two degrees of freedom. This transformation allows a comparison with electrical motors described in the Park's frame. This parallel allows to solution the complex problem of zero speed observation of the piston's position by transferring the methodologies which have been previously used on electrical systems. The utilization of the two degrees of freedom is illustrated by the synthesis of control law which combines the piston position trajectory tracking with the respect of a given criterion (which can be the pressurization level or the instantaneous gas consummation optimization). The use of an electropneumatic actuator as a variable stiffness actuator (VSA) is studied. A control law base on the A-T transform is proposed in order to simultaneously control the position and the pneumatic stiffness of the actuator. A control gains tuning strategy is proposed to allow the system closed loop impedance tuning. The influence of the pneumatic stiffness over the closed loop stiffness is studied. An alternative energy source (pressurized helium) is also used for the first time. The influence of the change of gas over the actuator sizing is studied and a methodology allowing using control laws designed for air applications is proposed. Every proposition are tested and validated using an innovative test bench with a unique structure. The latter combines two actuators. One is an electropneumatic actuator (the one studied) and the other one is a linear motor (used as the load actuator). The use of the latter allows disturbance generation in a wide bandwidth and the real-time tuning of the load dynamic characteristics. Finally, the tools developed in this PHD are used in order to validate the implementation of an electropneumatic solution to the thrust vector control problem of a nano scale space launcher. Data provided by the CNES (the French National Space Agency) are used to de ne specific requirements and the flight conditions are reproduced on the test bench to assess the solution quality.

O trabalho de pesquisa realizado trata do comportamento dinâmico e aeroelástico em compósitos laminados de rigidez variável (LCRV). Em virtude dos avanços das técnicas de manufatura de laminados, este tema tem ganhado notoriedade internacional e sua importância se justifica pelo crescente número de trabalhos na área. Neste contexto, foram analisados dois tipos distintos de LCRV: o primeiro diz respeito a laminados fabricados com espaçamento variável entre as fibras paralelas; e o segundo trata de laminados fabricados com deposição das fibras por caminhos curvilíneos, termo em inglês tow steering. Com objetivo de explorar as características dos LCRV, foi desenvolvido um modelo aeroelástico baseado no método dos modos admitidos (Rayleigh-Ritz), utilizando as hipóteses da teoria clássica de laminação (TCL), e foi utilizado o modelo aerodinâmico baseado na teoria das faixas quase estacionária para as análises em escoamentos subsônicos, e na teoria dos pistões, para escoamentos supersônicos nas análises de flutter de painel. Assim, foi investigada a influência do efeito de diferentes funções de distribuição do volume de fibras no comportamento aeroelástico e nas três primeiras frequências naturais, constatando-se uma significativa influência nos resultados, justificando um tratamento adequado para modelagem microestrutural dos laminados com espaçamento variável. Por outro lado, com objetivo de analisar o efeito de incertezas no processo de fabricação nos LCRV fabricados com a tecnologia de tow steering, foi desenvolvida uma estratégia de identificação de incertezas e sua propagação no modelo numérico, além da otimização para obtenção de um projeto robusto. Constatou-se que a melhor configuração obtida pela otimização determinística apresentou grande dispersão quando perturbado o ângulo de definição da trajetória das fibras, diferentemente da configuração selecionada de maneira robusta, que apresentou resultados menos sensíveis a perturbações nos ângulos de deposição das fibras. Foi investigada também a viabilidade da utilização de LCRV do ponto de vista dinâmico, visando aumentar a frequência natural fundamental, e com aplicação em flutter de painel. Em ambas as aplicações foi otimizada a trajetória da deposição das fibras, com base nos polinômios interpoladores de Lagrange, sendo encontrados ganhos razoáveis quando comparados com os laminados de material composto tradicionais de rigidez constante (LCRC). Adicionalmente, foi verificado que os resultados obtidos para o LCRC e o LCRV analisados experimentalmente corroboraram os resultados obtidos numericamente no que diz respeito às frequências naturais e aos modos de vibrar. / The developed research work is related to the dynamic and aeroelastic behaviors of variable stiffness composite laminate plates (VSCL). Due to the advances in the manufacturing techniques, this research theme has been gaining international relevance and its importance is justified by the increasing number of research works in this area. In this context, two different types of VSCL are analyzed: the first regards variable fiber spacing laminates, and the second is manufactured using curvilinear paths (tow steering). In order to explore the VSCL characteristics, it was developed an aeroelastic model based on the assumed modes approach (Rayleigh-Ritz), using the hypotheses of “classical lamination theory” (CLT). Moreover, it was used the aerodynamic model based on the quasi-steady strip theory in the subsonic analyses, and the piston theory, for supersonic flows used in the evaluation of panel flutter. It was investigated the influence of different fiber volume distribution on the aeroelastic behavior and on the first three natural frequencies; it has been found that those influences are significant, which justifies the adequate treatment for the micro -structural model of VSCL. Also, to cope with uncertainties during manufacturing of steered composite laminates, it was developed a strategy for identification of those uncertainties and their propagation through the numerical model; also, an optimization procedure was proposed to achieve robust designs. It was noticed that the response of the optimal configuration obtained from deterministic optimization presented a large dispersion when the tow steering angles were perturbed, in contrast with the selected configuration obtained from robust optimization, in which the results were much less sensible to perturbations in the tow steering angles. Also, it was investigated the viability to use LCRV from the dynamic standpoint, aiming at increasing the fundamental frequency, and with application in flutter panel. For both applications, the fiber placement trajectory was optimized based on Lagrange polynomials. Reasonable gains were found with respect to constant stiffness composite laminates (CSCL). Additionally, it was verified that the obtained experimental results for VSCL and CSCL corroborate the counterparts obtained from numerical simulations regarding natural frequencies and mode shapes. / Tese (Doutorado)

This contribution describes the properties and limitations of multi-layered mechanical devices with variable flexural stiffness. Such structures are supposed to be components of new smart, self-sensing and self-controlling composite materials for lightweight constructions. To enable a proper stiffness control, reliable actuators with high actuation capabilities based on smart materials are used. Those actuators are either driven by electroactive polymers (EAPs) or shape memory alloy (SMA) wires. They control the area moment of inertia of the multi-layered bending structures. To change the area moment of inertia and, hence, the flexural stiffness of an multi-layered beam within a wide range, it is necessary to stack as many layers as possible over each other. The fundamental function of this approach is demonstrated with a three-layer stack consisting of three independent layers and four form closure actuators driven by SMAs. This experimental set-up was able to change its bending stiffness k by a factor of 14.6, with a minimum and maximum stiffness of kmin = 0.11 N mm¯¹ and kmax = 1.73 N mm¯¹, respectively. The usage of four independently controllable actuators yields nine independent flexural-stiffness states of the beam. Both analytical and numerical calculations have shown good agreement with the measured stiffness values.

No description available.