Compensación adaptativa feedback feedforward para el rechazo de perturbaciones en una estructura flexible de un grado de libertad: análisis comparativo
No Thumbnail Available
Date
Advisors
Evaluators
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad Industrial de Santander
Abstract
Description
En este artículo se estudia un Control Vibratorio Activo (AVC) para un problema de tres carritos. Se comparan los algoritmos de mínimos cuadrados filtrados (FxLMS) y mínimos cuadrados recurrentes (RLS) en términos de rechazo de perturbaciones, costo computacional y esfuerzo de control cuando se dispone de una medición correlacionada de la perturbación. El compensador RLS propuesto considera un acoplamiento de retroalimentación entre el compensador y la perturbación. La ruta de propagación secundaria de la planta se estimó utilizando el algoritmo LMS normalizado (NLMS). El acoplamiento positivo interno se modela como un filtro FIR estimado por los parámetros reales de la planta. Las simulaciones mostraron un rendimiento superior del algoritmo RLS con un costo informático razonable. El análisis comparativo se realizó comparando la compensación entre el orden del filtro y la magnitud del rechazo.
In this paper an Active Vibrational Control (AVC) for a three-cart problem is studied. The Filtered-x Least Mean Square (FxLMS) and Recursive Least Square (RLS) algorithms are compared in terms of disturbance rejection, computational cost and control effort when a correlated measurement of the disturbance is available. The proposed RLS compensator considers a feedback coupling between the compensator and the disturbance. The secondary propagation path of the plant was estimated using normalized LMS (NLMS) algorithm. The internal positive coupling is modeled as a FIR filter estimated by the real plant parameters. Simulations showed a superior performance of RLS algorithm with a reasonable computer cost. The comparative analysis was performed comparing the tradeoff between the filter order and the magnitude of the rejection.
In this paper an Active Vibrational Control (AVC) for a three-cart problem is studied. The Filtered-x Least Mean Square (FxLMS) and Recursive Least Square (RLS) algorithms are compared in terms of disturbance rejection, computational cost and control effort when a correlated measurement of the disturbance is available. The proposed RLS compensator considers a feedback coupling between the compensator and the disturbance. The secondary propagation path of the plant was estimated using normalized LMS (NLMS) algorithm. The internal positive coupling is modeled as a FIR filter estimated by the real plant parameters. Simulations showed a superior performance of RLS algorithm with a reasonable computer cost. The comparative analysis was performed comparing the tradeoff between the filter order and the magnitude of the rejection.
Keywords
Index terms –active vibration control, fir adaptive filter, the filtered-x least mean square, recursive least square, Términos del índice -active vibration control, fir filtro adaptativo, el cuadrado medio menos filtrado de x, recursive least square