基于BP神经网络的摩擦系数预测方法研究

The Research of Friction Coefficient Prediction Based on BP Neural Network

准确的获取摩擦系数是提高摩阻计算精度的有效方式之一,摩擦系数隐含了力学分析模型假设与实际情况的差异以及井眼形状、钻井液体系等因素的影响,使摩阻系数表现出非稳定的随机性和个体差异.充分利用多层前馈神经网络具有的自学习、自组织、自适应和非线性动态处理等特性,将BP网络与摩擦系数预测活动进行了有效的结合,建立了摩擦系数及其影响因素之间的隐含关系预测模型.通过仿真实验表明可利用神经网络实现从摩擦系数的角度提高摩阻扭矩预测精度的目的,解决了钻井过程中存在的摩擦系数个体差异问题.

In this paper we discuss a Predictor-Corrector Approach for the Numerical Solution of Fractional Differential Equations. We select the simulation research on Chen chaotic system. First, discuss under certain initial conditions, the fractional Chen chaotic system is chaotic and still present a rich and complex fractional order chaotic dynamics behavior. With the Predictor-Corrector Approach, we use controller of piecewise quadratic function on Chen chaotic system to obtain the equilibrium point. By programming in MATLAB, we get the stable phase diagram of fractional order Chen chaotic systems.