油包水型乳状液蜡沉积速率预测研究

发布时间：2018-01-15 11:14

本文关键词：油包水型乳状液蜡沉积速率预测研究　出处：《西南石油大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着深水油田的开发和海底管线的建设,油水乳状液的蜡沉积问题已成为多相混输技术面临的一个新的问题。油水乳状液物性复杂,导致蜡沉积模型研究困难加大,目前国内外学者建立的蜡沉积模型多为单相模型。本文利用已报道的蜡沉积实验数据开展油包水型乳状液蜡沉积预测速率研究。本文在跟踪国内外该领域内相关研究成果的基础上,以215组不同油水乳状液蜡沉积数据为研究对象,提出以组合预测法建立油包水型乳状液粘度预测模型,根据经验公式计算蜡沉积相关参数,完善蜡沉积实验数据样本。本文引入PSO-RBF神经网络建立油包水型乳状液蜡沉积速率预测模型,与RBF神经网络模型和改进黄启玉模型进行比较和验证,同时预测实际管道中蜡沉积速率沿线分布规律。本文主要的研究内容与结论如下:(1)通过文献调研收集油包水型乳状液蜡沉积速率的实验数据,整理、分析后获得蜡沉积模型需要的数据。(2)通过粘度组合预测模型、Vand模型、Brinkman模型、Pal模型及大庆油田模型分别对5种油水乳状液的204组粘度数据进行比较。结果表明,粘度组合预测模型适用性好、预测精度高,其平均相对误差仅为5.4%。(3)本文在215组油水乳状液蜡沉积数据样本的基础上,引入PSO-RBF神经网络模型预测蜡沉积速率,并与RBF神经网络模型、改进黄启玉模型对应的预测结果进行对比。结果表明,PSO-RBF神经网络模型预测的蜡沉积速率与实验值的相对误差较小,具有精度高,预测能力稳定等特点。(4)蜡沉积影响因素多而复杂,确定各因素的主次对蜡沉积规律研究尤为重要。采用RBF神经网络中输出变量对输入变量的偏导的敏感性分析,确定了蜡沉积影响因素之间的关系,量化了各影响因素的敏感性系数,并对影响因素的影响程度进行排序。(5)利用本文的蜡沉积模型预测了不同工况下实际管道蜡沉积速率沿线分布规律,并根据蜡沉积模型预测的沿程摩阻增加值和现场沿程摩阻增加值进行对比分析,间接验证蜡沉积模型预测结果的准确性。结果表明,PSO-RBF神经网络模型可用于预测油水混输管道的蜡沉积分布规律。
[Abstract]:With the development of deep water oil field and the construction of submarine pipeline, wax deposition of oil-water emulsion has become a new problem in multiphase mixed transportation technology. The physical properties of oil-water emulsion are complex. Therefore, it is difficult to study wax deposition model. Most of the wax deposition models established by scholars at home and abroad are single-phase models. In this paper, the prediction rate of oil-in-water emulsion wax deposition is studied based on the reported experimental data of wax deposition. This paper tracks the correlation in this field at home and abroad. Based on the research results. Taking 215 groups of different oil-water emulsion wax deposition data as the research object, a prediction model of viscosity of oil-in-water emulsion was established by combination prediction method, and the related parameters of wax deposition were calculated according to empirical formula. In this paper, PSO-RBF neural network was introduced to establish the oil / water emulsion wax deposition rate prediction model. Compared with RBF neural network model and modified Huang Qi-yu model. The main research contents and conclusions of this paper are as follows: 1) collect the experimental data of oil-in-water emulsion wax deposition rate through literature investigation and collate. After analysis, the data needed for wax deposition model are obtained. (2) through viscosity combination prediction model, Vand model and Brinkman model are obtained. The Pal model and Daqing oil field model compare 204 sets of viscosity data of five kinds of oil-water emulsion respectively. The results show that the viscosity combination prediction model has good applicability and high prediction accuracy. The average relative error is only 5.40.The paper introduces PSO-RBF neural network model to predict wax deposition rate on the basis of 215 samples of oil-water emulsion wax deposition data. Compared with the RBF neural network model and the modified Huang Qi-yu model, the results show that the relative error between the wax deposition rate predicted by the PSO-RBF neural network model and the experimental value is small. It has the characteristics of high precision, stable prediction ability and so on. (4) the influence factors of wax deposition are many and complex. It is very important to determine the primary and secondary factors for wax deposition. The sensitivity analysis of output variables to the deviation of input variables in RBF neural network is used to determine the relationship between the influencing factors of wax deposition. The sensitivity coefficient of each influencing factor is quantified, and the influence degree of the influencing factor is sorted. 5) the distribution law of wax deposition rate along the actual pipeline under different working conditions is predicted by using the wax deposition model in this paper. The accuracy of the prediction results of wax deposition model is indirectly verified by comparing and analyzing the increase value of friction along the line and the increase value of friction along the field, which is predicted by the wax deposition model, and the results show that the prediction results of the model are accurate. PSO-RBF neural network model can be used to predict the wax deposition distribution of oil and water mixed pipeline.
【学位授予单位】：西南石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TE832

【相似文献】