基于BP人工神经网络的含蜡原油触变应力计算

含蜡原油的流变性对其管道输送有着重要的影响.当原油温度逐渐接近凝点时,原油表现为非牛顿流体,其流变性表现出异常复杂的触变性.利用人工神经网络较强的非线性逼近、良好的自适应和预测性能,采用误差反向传播算法(即BP算法)对含蜡原油的触变剪切应力进行了计算,计算结果与实验结果和采用R-G模型方程计算的结果进行了对比.结果表明,BP网络计算的精度高于R-G模型方程计算的精度,BP网络和R-G模型的计算结果与实验结果的最大相对误差分别为5.0%和12.7%.

Calculation of thixotropic stress of waxy crude oil by BP artificial neural network

The rheological behavior of waxy crude oil has important influence on the long distance pipe transportation of crude oil. When the temperature of waxy crude oil decreases and approaches its solidification point, the crude oil behaves as the non-Newtonian fluid and has strong thixotropy. The artificial neural network has strong ability of non-linear approaching, self-adjusting and predicting, so the BP artificial neural network is used to predict the thixotropic shear stress of the waxy crude oil. Results obtained are compared with the experimental ones and those calculated by the R-G model. The results show that the maximum relative errors of thixotropic stress between the experimental and calculated ones are 5.0% and 12.7% by using the BP artificial neural network and R-G model respectively. So the BP artificial neural network can predict thixotropic stress of waxy crude oil and has better prediction results than those by using the R-G model.