Application Study Of Damping Performance Of Magnetic Fluid Porous Medium Damper On Basis Of BP Neutral Network

Large-scale precision instruments such as spacecraft is easy to be influenced by perturbation and vibrates.The vibration has a variety of characteristics such as low frequency and small amplitude.It is necessary to install magnetic fluid damper or magnetic fluid porous media damper on spacecraft.Compared to the magnetic fluid damper,the magnetic fluid porous media damper is applicable to a wider frequency vibration damping,and has better damping effect.But the number,the radius and the length of the holes is hard to measure.So it is hard to accurately calculate the magnetic fluid damper’s damping effect.To solve the problem,the BP neural network is applied to this damper,and the damping effect can be calculated accurately by this method.In addition the paper uses the method of quantum mechanics to analyze chain formation of magnetic fluid and calculate the volume of adsorbed magnetic fluid,for neural network training data.The thesis includes four parts as following:1.Using the quantum mechanical analysis chain formation of magnetic particles in the external magnetic field.Establish physical models of magnetic particles’ chain formation.Put forward reasonable assumption to simplify chain formation model.2.The shape of the adsorbed magnetic fluid is treated as oval.Analysis the chain formation of the magnetic particles at the surface of adsorbed magnetic fluid,and solve the thickness and width of adsorbed magnetic fluid.The volume of adsorbed magnetic fluid can be calculated accurately,and it will be the input data of BP neural network.3.The magnetic fluid porous media damper is designed,and set up damping test platform.Basing on damping data,draw the picture of relationship diagram between vibration isolation coefficient and vibration coefficient.4.Design the structure of BP neural network,and input damping data into the BP neural network for training.The effective neural network which can calculate the vibration isolation of the damper is got.Optimize the neural network,and improve the computing performance of the neural network.