基于多传感器数据融合的油气水三相流相态组合的辨识

发布时间：2018-02-06 01:12

本文关键词： 多相流流态辨识数据融合特征参数　出处：《东北石油大学》2016年硕士论文　论文类型：学位论文

【摘要】：目前国内陆上各大油田处于高含水开发期,地层压力和油井流压偏低,油井普遍呈现脱气状态,大多数油气井都为油气水三相的混合流动状态。多相流动特性复杂,而且相间存在界面效应和相对速度,致使参数检测难度较大。目前尚没有成熟的高含水、低产液的三相流参数检测技术。当前的多相流参数检测对多传感器精度和流型依赖性强,带来了很大的测量误差。为克服上述问题,本文提出了基于缓冲器在线分离的测量方法,不依赖于传感器绝对精度和流型的流速检测,可实现油气水分相流量的测量。具体研究内容如下:基于在线分离和相态转换的三相流测量方法依据重力分异原理,控制流体动态分离使得油气水分时分相通过测量出口端。通过判断气相液相转换点和油相水相转换点,通过体积法就可获得各相流量。传感器只是用来做流态判断,故对传感器绝对精度要求不高;通过罐体缓冲和重力分异作用,降低流型对测量精度的影响,提高测量精度。为识别流体相组合流态,提出了基于多传感器特征提取的方法。采用小波技术对流动信号进行分析,提取对信号敏感的均值、方差、标准差和导数组成流态识别特征参数。在特征提取基础上,建立了基于最小错误率的贝叶斯算法流态在线识别模型。该模型的建立实现了油气水三相流流态的自动识别,为分相流量测量打下了基础。为验证基于流态转换的三相流测量方法的正确性,在多相流平台上,对测量系统开展了多相流实验研究。实验结果表明,本测量方法适应于油气水三相流测量,在线流态识别模型可以实现油气水组合流体的识别,实现了较高的分相流量测量精度。
[Abstract]:At present, the major onshore oil fields in China are in the period of high water cut development, the formation pressure and the oil well flow pressure are low, and the oil wells generally present the state of degassing. Most oil and gas wells are in the mixed flow state of oil-gas-water three-phase. The multi-phase flow characteristics are complex, and the interphase effect and relative velocity exist between the phases, which makes the parameter detection difficult. At present, there is no mature high water cut. In order to overcome the above problems, the current multiphase flow parameter detection has a strong dependence on the accuracy and flow pattern of multi-sensor and brings great measurement error. In this paper, an on-line separation method based on buffer is proposed, which does not depend on the sensor's absolute accuracy and flow pattern. The specific research contents are as follows: the three-phase flow measurement method based on on-line separation and phase transition is based on the principle of gravity differentiation. By controlling the dynamic separation of the fluid, the time division phase of oil, gas and water can be measured through the exit, and the gas phase liquid phase transition point and the oil phase water phase transition point can be determined. The flow rate of each phase can be obtained by volume method. The sensor is only used to judge the flow state, so the absolute precision of the sensor is not high. By the action of tank buffer and gravity differentiation, the influence of flow pattern on measurement accuracy is reduced, and the measurement accuracy is improved. A method of feature extraction based on multi-sensor is proposed. The wavelet technique is used to analyze the flow signal to extract the mean value and variance sensitive to the signal. The standard deviation and derivative are used to identify the characteristic parameters of the flow state. A Bayesian algorithm based on minimum error rate (MER) was established to identify the flow state of oil, gas and water, and the model was established to realize the automatic recognition of the flow state of oil-gas-water three-phase flow. In order to verify the correctness of the three-phase flow measurement method based on the flow state conversion, the multiphase flow experimental research was carried out on the multiphase flow platform. This measurement method is suitable for oil-gas-water three-phase flow measurement. On-line flow pattern identification model can realize the identification of oil-gas-water combination fluid and achieve high accuracy of phase separation flow measurement.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TE937

【相似文献】