Design of genetic-algorithm-based multimode controller

Abstract

In this study, the genetic algorithm (GA) was used to search two modes filter frequencies to be applied to the multiple positive position feedback (PPF) controller. The PPF technique, which makes use of generalized displacement measurements to maximize damping in a specific mode effectively and accomplish vibration suppression, was employed as the control scheme. However, the PPF controllers require tuning of the filter frequency of controller to the structural resonance prior to compensator design. Another disadvantage of the multi mode PPF control is that a lower mode natural frequency is shifted to a lower value as the effect of higher mode control by the PPF controller. To minimize the effect of frequency shift and search the optimal filter frequencies for the multiple modes PPF controller, the GA was applied. In many applications, the sum of the square of sensor voltage output is used as the fitness function in the GA. However, in the present case, it hardly distinguishes the definite difference in the neighboring modes. Therefore, in this study, the GA with the fitness function defined as the function of the power spectrum information was adopted. As a result of GA with the proposed fitness function, the optimal filter frequencies of multi modes were found successfully.G