Improved Linear Quadratic Control Method Based on Convolutional Neural Network Vibration Recognition

Abstract

The invention discloses a linear quadratic control improvement method based on convolutional neural network vibration recognition. The method includes: collecting vibration data and controlled device data; establishing a mechanical model according to the controlled device data, inputting vibration data and using linear The quadratic control algorithm controls the vibration, and the optimization algorithm is used to solve the optimal control parameters corresponding to the linear quadratic control algorithm under each type of vibration; the wavelet transformation is performed on the vibration data to obtain the wavelet coefficient matrix and the corresponding wavelet image. The vibration type is constructed into a data-label data group, which is input into the convolutional neural network for classification training; the best convolutional neural network classifier is selected to classify the vibration input, and the optimal linear quadratic control corresponding to the vibration type is selected according to the vibration type identification result. Control parameters. This method can accurately identify different types of vibration inputs, and select the corresponding optimal control parameters according to the identification results, so as to realize the optimal control under different vibration inputs.

Description

technical field [0001] The invention relates to the technical fields of civil structural engineering and mechanical manufacturing engineering, in particular to an improved linear quadratic control method based on convolutional neural network vibration identification. Background technique [0002] Effective vibration control can reduce the adverse effects of micro-vibration in the environment, improve the use of precision instruments and the quality of finished products in industrial manufacturing. However, the vibration sources that cause vibration are often complex and changeable. For example, subways, heavy vehicles, and construction processes will generate corresponding vibration interference. However, the widely used classic LQR (Linear Quadratic Regulator, linear quadratic control) control method is difficult to achieve optimal control of different vibrations in complex environments. The main reasons are: 1) The existing LQR control parameter optimization method is when...

Images