Modeling and control of Shape Memory Alloy Actuator.

Abstract : This thesis propose a black box approach for the modeling and the control of SMA actuators, as an alternative to the usual phenomenological and physic models. Models using Laguerre functions are used due to their simplicity of implementation, their good convergence properties and the small amount of a prior knowledge needed. Moreover, using Volterra series, nonlinear models can be developed. All these models being linear in the parameters they can be identified on-line using classical exponential forgetting recursive least squares algorithms. An experimental study demonstrates the ability of these models to predict the output of two different setups using SMA actuators. The second part of the thesis addresses the control problem. A first adaptive predictive controller using Laguerre predictor is selected and the stability conditions are carefully studied showing conditionnal stability toward the prediction horizon. Experimental results demonstrate the limitation due to the lack of exciation and a directionnal forgetting RLS algorithm is successfully tested. To adress the problem of closed loop identification, a second controller structure is proposed that uses a global Laguerre model of the closed loop system. By doing so, the reference is used as an instrumental variable and solves the problem of stability of the identification, thus a simple RLS algorithm is used. The stability of the controller is proved and does not depend on the prediction horizon. Experimental assessment on the two actuators gives good results, moreover it reveals some parallel with model reference adaptive control.
Complete list of metadatas

Cited literature [109 references]  Display  Hide  Download

https://pastel.archives-ouvertes.fr/pastel-00724901
Contributor : Somasundar Kannan <>
Submitted on : Thursday, August 23, 2012 - 10:15:32 AM
Last modification on : Tuesday, May 28, 2019 - 3:40:10 PM
Long-term archiving on : Friday, December 16, 2016 - 7:29:50 AM

Identifiers

  • HAL Id : pastel-00724901, version 1

Citation

Somasundar Kannan. Modeling and control of Shape Memory Alloy Actuator.. Automatic Control Engineering. Ecole nationale supérieure d'arts et métiers - ENSAM, 2011. English. ⟨NNT : 06O9KF013P3⟩. ⟨pastel-00724901⟩

Share

Metrics

Record views

373

Files downloads

1254