Performance study of neural prediction and adaptive fuzzy algorithm-based active/passive hybrid isolation system

Abstract

To address the issue of unstable vibration isolation performance in traditional isolation systems under time-varying and nonlinear vibration loads, a novel intelligent hybrid isolation system, combining both passive and active elements, is proposed. Theoretical analysis of the load transmission characteristics of the hybrid isolation system is conducted, and an adaptive fuzzy control strategy based on neural network prediction is introduced. This hybrid isolation system effectively utilizes the high-frequency performance of passive isolation and the low-frequency capabilities of active control. Additionally, a neural network is employed to eliminate the time delay effects inherent in the active control process. The adoption of a fuzzy control system, coupled with the integration of a genetic algorithm for adaptability, further enhances the system’s stability when dealing with diverse vibration scenarios. Simulation and experimental results demonstrate the excellent vibration isolation performance of the hybrid system under variable frequency and amplitude excitations, the peak isolation efficiency of the experimental model of this isolation system under complex excitations is around 72%.