管道输送原油蜡沉积速率模型研究

Investigation on the Model of Oil Pipeline Wax Deposition Velocity

分别采用人工神经网络法和逐步回归分析法对原油管道蜡沉积实验数据进行分析处理,建立蜡沉积速率模型。利用建立的蜡沉积速率模型预测管道蜡沉积速率并与实际蜡沉积速率相比,结果表明,基于人工神经网络方法建立的蜡沉积速率比基于逐步回归分析法建立的蜡沉积速率精度高,但逐步回归分析法计算速度快且具有表达蜡沉积速率与其影响因素之间亲疏关系的优点,而人工神经网络方法没有此功能。综合考虑两种方法的优缺点,在建立蜡沉积速率模型时,先用逐步回归分析法对实验数据进行预处理,得出蜡沉积速率的主要影响因素,然后再把蜡沉积速率的主要影响因素作为神经元的输入,利用人工神经网络的方法建立蜡沉积速率模型。

As the artificial neural network and stepwise regression analysis are respectively applied to the data of crude wax deposition experiment, the model of wax deposition velocity was established. Using this model to predict wax deposition velocity and comparing with actual value, the results show that the former is superior to the latter in precision. But the latter is in a good position for calculation at a high speed and can express the affinity order between the wax deposition velocity and its affection factors. Considering advantages and disadvantages of two ways, the stepwise regression analysis first is used to get the main affection factors of wax deposition velocity, then the artificial neural network is used to establish the model of wax deposition velocity with the main affection factors of wax deposition velocity as nerve cell input.