基于BP神经网络的蜡沉积速率及清蜡周期预测方法

Methods for Predicting Wax Deposition Rate and Wax Removal Cycle Based on BP Neural Network

为应对XX油矿井筒结蜡对油井生产造成的严重危害,首先通过试验方法对原油基本物性以及析蜡特性进行分析,再运用SPSS大数据分析软件进行敏感性分析,发现油井产液量、含水量、生产气油比和生产时间是影响蜡沉积速率的地面生产特征,而剪切应力、原油黏度、径向温度梯度、蜡分布密度为其地下内在联系。将这4个内在影响因素作为输入参数,蜡沉积速率作为目标参数,使用SPSS Modeler软件建立神经网络模型。最终应用模型计算可以得到各油井的清蜡周期,针对蜡堵严重油井进行提前预判,采取相应清防蜡技术,降低结蜡对油田开发的不利影响,提高油井生产时率,这对油田的稳产具有重要的意义。

To cope with serious damage to the oil well production induced by wax deposition in XX Oilfield,the basic physical property of crude oil and wax precipitation characteristics were analyzed first by using experimental method,and then SPSS data analysis software was applied for sensitivity analysis,it was found out that the oil well fluid production,water,oil and gas production and production time are the surface characteristics of influencing wax deposition rate,while the shear stress,oil viscosity,radial temperature gradient,the density of wax distribution are the internal and underground connection with it.The four intrinsic factors were used as input parameters,and the wax deposition rate was as the target parameter,the neural network model is established by using SPSS Modeler software.The model is finally used to obtain the pigging period of each well,the wells with serious wax plugging are predicted in advance,the corresponding paraffin wax technology is taken to reduce the adverse effects on oilfield development,increase the oil production rate,which is of great significance for oil field production.