Learning Speed-Gradient Synchronization Control of the Two-Rotor Vibration Setup

In this paper the implementation of the speed-gradient single or multiple synchronization control algorithm with learning for a two-rotor vibration setup is considered. The main contribution of the paper is demonstration of possibility of neural network control of vibration setups under uncertainty. We propose the structure of the learning control system which is developed to expand the range of the vibration machine efficiency and increase the performance of its operation with uncertainty system parameters. The neural network was trained on simulation results for various loads and desired total system energy values. The presented results of computer simulation demonstrate the possibility of using a learning vibration setup control system based on an artificial neural network and speed gradient algorithm to compensate the platform mass change and keep the desired vibration level due to various bulk media feeding modes.