The Active Vibration Control of a Centrifugal Pendulum Vibration Absorber Using a Back-Propagation Neural Network

This study investigates a back-propagation (BP) neural network learning rule for control and system identification of an active pendulum vibration absorber (APV A) and develops an approach to find the bounds of learning rates based on the Lyapunov function. The use of adaptive learning rates guarantees convergence so the optimal learning rates were found. The objective of the BP algorithm was trained for tuning the system parameters in an APV A by suppressing vibration of the carrier. The simulation results for the BP neural network algorithm APVA are compared with the fuzzy BP neural network with non-neuroidentifier algorithm. The simulation results demonstrate the absorbing effectiveness of the proposed adaptive learning rates of BP neural network APVA to reduce carrier vibrations.