Stability/instability of electrorheological nano pipes used in braking nanosystems by the machine learning method

Electro-rheological fluids change their viscosity properties due to their application in an electrical field. This property could be acquired and maintained by accurately controlling the external electrical field. Besides brake systems, such electro-rheological (ER) fluids could be utilized in small-scale damping systems. In the present study, for the first time, the effects of using an electrical field in improving the dynamic tability of nanostructure containing ER fluids are presented. In this regard, a cylindrical sandwich structure, with an inner and outer layer composed of two-directional functionally graded (2D-FG) material and ER fluid core, is considered. Employing energy methods and modified power law the equations of motion of the structure are derived. Between FG and ER layers’ compatibility conditions are imposed in terms of displacement and strains. Moreover, to bypass computational complications, the inputs and outputs are utilized for training a deep neural network (DNN). In this way, the inputs and outputs are related to each other through the regression method without numerically solving the equations of motion all over again. Finally, the effects of different parameters on the frequency and loss factor characteristics of the current nanostructure will be presented.