RBF神经网络在油水两相流含率超声测量中的应用

Application of RBF Neural Network to Ultrasonic Holdup Measurement for Water-Oil Two Phase Flow

  • 摘要: 介绍了一种油水两相流含率测量新方法——超声散射法.该方法利用超声波散射效应测量流体中的分相含率,通过非集流方式,在基本不改变流体流动状态的情况下实现油水两相含水率的测量.动态实验结果表明,该方法在油水两相条件下,可以实现分相的测量.但由于受流型、油泡直径及吸收系数等因素的非线性影响,使得测量模型难于建立.文章通过构建RBF神经网络,对含水率进行预测,提高了含水率测量精度.

     

    Abstract: A new method of holdup measurement for water-oil two phase flow is introduced, i.e. ultrasonic diffusion. The method uses the effect of ultrasonic diffusion to measure holdup of discrete phase. Then the discrete phase holdup can be measured by non-packer without changing flow state. Because of the nonlinear effect caused by flow pattern, oil diameter and absorption coefficient etc, it is difficult to establish the measurement model. The paper establishes RBF neural network to predict water holdup, and the measurement precision can be improved.

     

/

返回文章
返回