【作者】 唐友福；

【导师】 刘树林；

【作者基本信息】 大庆石油学院 ， 工程力学， 2006， 硕士

【摘要】 潜油电泵是一种大排量、高扬程的无杆抽油设备,是油田实现高产稳产的重要手段,目前,潜油电泵在我国各油田应用极为广泛。然而,潜油电泵工作环境恶劣、造价极其昂贵,一旦发生故障,将带来巨大的经济损失,因此,开展潜油电泵机组质量可靠性研究具有重要的现实意义。本文通过对大量潜油电泵机组失效数据的统计分析,选择了机组易发生故障的主要部位为测试对象,在结合现有实验设备及现场条件的基础上,运用先进的振动测试方法,制定了试验井中潜油电泵机组振动检测方案,并采集了大量机组实验数据。本文引入了以概率统计为基础的现代信号分析与处理方法——统计量理论,通过仿真,对其开展了从低阶统计量到高阶统计量、一般统计量到循环统计量的全面系统研究,分析了各种统计量的重要特性,并将其应用于潜油电泵机组振动状态评估中。本文利用峰-峰值、有效值等统计量具有在时域上描述振动信号幅值大小的特性,依据概率统计理论中的“3σ”原则,制订了适合于潜油电泵机组振动相对标准。还利用功率谱统计量具有在频域上描述振动信号能量大小的特性,对潜油电泵机组进行了振动状态的评估,提取了不同状态下的机组频谱特征。最后还利用双谱、双相干谱、循环均值、循环谱密度以及循环双谱等高阶统计量和循环统计量来提取机组状态特征。结果表明:循环统计量相比于一般统计量、高阶统计量相比于低阶统计量而言,具有分辨率更高、包含信息更丰富等优点,更能准确有效的表征潜油电泵机组的状态特征,但是,它们也具有计算更复杂、物理意义更模糊等缺点,因此很难制定统一可行的潜油电泵振动评估标准。最后,根据现场实际需要,结合本文的主要研究内容,运用Matlab丰富的函数工具箱以及VB简便的可视化编程语言,开发了潜油电泵机组振动检测与状态评估系统,通过现场实际测试,取得了良好的应用效果,证明了该系统的可靠性。更多还原

【Abstract】 ESP (electric submersible pump) is a kind of large displacement, high-head and no pole oilpumping equipment. It is one of the most important methods for making high and steadyproduction. At recent years, the ESP is very popular in each oilfield of our country. However, itsworking condition is atrocious and its cost is expensive. Once some key part exist failure, it willoccur considerable economy losses. Therefore, it has a very important practical significance totake research on the ESP unit quality reliability.By analyzing a large amount of the ESP unit failure data, the paper selects the mainpositions of the easily damaged as testing object. Combining with existing lab equipments,using advanced modern testing technique, research of vibration detection for the ESP unit intest-well are taken. Moreover, plenty of experiment data for the ESP unit are collected.The statistic theory, which is based on the probability and stat., is introduced in this paper.Systemic research that from low statistic to high statistic, from general statistic to cyclic statisticis taken based on the simulation results. The important characters of each statistic are analyzedand it is applied into the vibration state assessment for the ESP unit.Peak-peak value statistic and virtual value statistic can describe the amplitude in timedomain, so they are used for founding the relative vibration standard for the ESP unit accordingto the “3σ” principle of the probability and stat. theory. Power spectrum statistic can describe theamplitude in frequency domain, so they are used for assessing the vibration state for the ESPunit and the characters under different state are extracted. At last, bispectrum statistic,bicoherence spectrum statistic, cyclic mean statistic, cyclic spectrum density and cyclicbispectrum are also used for extracting the states for the ESP unit. The result demonstrate thatthe cyclic statistic and the high-order statistic have better resolution and include more usefulinformation than the general statistic and the low-order statistic correspondingly, so it can detectfault feature of the ESP unit more accurately and effectively. But its evaluation is still morecomplicated and the physical meaning is fuzzier, so it is difficult to formulate a uniformvibration assessment standard for the ESP unit.Finally, acorrding to the practical need of field and combining with the main researchcontents of this paper, the software system of vibration detection and state assessment for theESP unit is developed by using the abundant function tool box of MATLAB and the simplelanguage of VB. By the local test, well application effect has been gained and reliability ofsystem is verified.更多还原