Prediction Study On The Deposition Rate Of Water-in-oil Emulsion Wax

With the development of the deep water oil field and the construction of the submerged pipelines,the wax deposition problems of water-in-oil emulsion have become a new problem confronted by water-in-oil mixed transmission technique.The physical property of water-in-oil emulsion is complex causing that it is more difficult to research the wax deposition models.However,most of the current wax deposition models established by the domestic and overseas scholars are single models.This paper utilizes the reported wax deposition experiment data to conduct the research on wax deposition predicted velocity of water-in-oil emulsion.Based on tracking the domestic and overseas relevant research results in this field and taking the collected 215 groups of different water-in-oil emulsion indoor wax deposition experiment data as the research object,this paper proposes to establish the water-in-oil emulsion viscosity prediction model with combination prediction method.According to the empirical formula,the relevant parameters of wax deposition are calculated and the experiment data samples of wax deposition are completed.On the basis of the experiment data,this paper introduces PSO-RBF neutral network to establish the wax deposition velocity prediction model of water-in-oil emulsion,which is also compared and verified with RBF neutral network model and Q.Huang model.Meanwhile,the distribution law of wax deposition velocity along with the actual pipelines is predicted.The main research contents and conclusion of this paper are as follows:(1).Through document surveys,the experiment data of water-in-oil emulsion wax deposition velocity is collected,organized and analyzed;the relevant parameter data of the wax deposition demanded for modeling is obtained.(2).The viscosity combination forecasting model,Vand model,Brinkman model,Pal model and Daqing oil field model were used to compare by the viscosity data of 204 kinds of water-in-oil emulsions and it is discovered that the prediction accuracy of the combination prediction model is preferable and the applicability is relatively extensive.(3).Based on 215 groups of indoor wax deposition experiment data,this paper introduces PSO-RBF neutral network to establish water-in-oil emulsion wax deposition velocity prediction model.The error analysis on PSO-RBF neutral network model,RBF neutral network model and Q.Huang model is conducted.Through comparison,it is obtained that the error of PSO-RBF neutral network model is the minimum and the prediction result is the best.(4).There are many complex factors affecting wax deposition rate of waxy crude oil.Their priority is particularly important for building a wax deposition prediction model.Using the sensitivity analysis of the output variables in the RBF neural network to the partial deviation of the input variables,the relationship between the factors affecting the wax deposition was determined,the sensitivity coefficients of the influencing factors were quantified,and the influence degree of the influencing factors was sorted.(5).Based on the wax deposition model of this paper,the distribution law of the actual pipeline wax deposition rate under different working conditions is predicted.Based on the comparative analysis of the increase of the value of the friction loss in pipeline and the increase of the friction loss in pipeline along the field.Indirect verification of the accuracy of the prediction results of the sedimentary model.The results show that the PSO-RBF neural network model can be used to predict the wax distribution of oil-water mixed pipeline.