基于管柱声场模型的油井动液面深度检测信号处理

发布时间：2018-04-24 20:09

本文选题：动液面 + 检测　；参考：《西安石油大学》2017年硕士论文

【摘要】：油井动液面深度的准确检测对于提高采油效率具有重要意义。目前主流的回波法存在发声装置不够安全可靠和液面反射波难以辨识的缺陷,还不能较好地满足油田的开采要求。贾威等人提出了基于管柱声场模型的油井动液面检测方法,为复杂油井动液面深度检测提供了一种新的途径。本文在此新方法的基础上做了进一步研究,提高了检测精度。首先,对共振模型进行改进,建立了相邻两阶共振频率的间距(35)f与动液面深度的数学模型,有效避免了辨识nf的阶数n的问题。其次,油田生产中瞬间脉冲声波可激发油井套管中空气柱共振,但共振声波能量衰减严重,非常不利于检测。针对这一缺陷,采用连续白噪声的发声方式,可激发井下空气柱产生多阶频率的共振;且通过在井口连续发声,可不停地给井下的共振声波补充能量。再次,针对共振声波中包含大量的随机白噪声、各种障碍物的反射声波和井下电机运转产生的机械噪声,提出了一种基于Welch功率谱估计的频谱FFT方法,可极大地消除噪声的干扰,准确方便地提取出了共振谐波的频率差。此外,针对FFT变换中由于频谱泄漏和栅栏效应导致频率值存在误差的问题,研究了目前国内外经典常用的插值FFT算法。通过分析发现,现有算法的校正精度不够稳定且与频率偏差存在很大关系。提出了一种改进的迭代插值算法,校正精度高有了进一步提高。利用此方法对Welch功率谱估计的FFT结果进行校正,进一步提高了动液面的检测精度。最后,将以上信号处理的方法用于模拟井下套管的管道长度检测实验中,130米内管道长度的检测误差在0.5米以内,验证了课题信号处理方法的可行性。
[Abstract]:It is very important to measure the depth of moving level of oil well to improve the oil recovery efficiency. At present, the mainstream echo method has some defects such as the sound generating device is not safe and reliable, and the reflection wave of liquid surface is difficult to identify, so it can not meet the requirements of oilfield exploitation. Javert et al put forward a new method of oil well moving level detection based on pipe string sound field model, which provides a new way for the detection of moving liquid level depth in complex oil wells. Based on the new method, the detection accuracy is improved. Firstly, the resonance model is improved, and the mathematical model of the distance between the adjacent two order resonance frequencies and the depth of the moving liquid surface is established, which effectively avoids the problem of identifying the order n of the nf. Secondly, the instantaneous pulse acoustic wave can excite the hollow gas column resonance in the oil well casing, but the acoustic energy of the resonance sound wave attenuates seriously, which is very unfavorable to the detection. In order to solve this problem, continuous white noise can be used to excite the resonance of multiple frequencies in underground air column, and by continuously producing sound at the well head, it can continuously replenish the energy of resonance acoustic wave in the well. Thirdly, a frequency spectrum FFT method based on Welch power spectrum estimation is proposed, which includes a large number of random white noise, reflected acoustic waves from various obstacles and mechanical noise generated by underground motor operation. The noise interference can be eliminated greatly and the frequency difference of resonance harmonics can be extracted accurately and conveniently. In addition, in order to solve the problem of frequency error caused by spectrum leakage and fence effect in FFT transform, the classical interpolation FFT algorithm is studied at home and abroad. Through analysis, it is found that the correction accuracy of the existing algorithms is not stable and has a great relationship with the frequency deviation. An improved iterative interpolation algorithm is proposed, and the correction accuracy is further improved. This method is used to correct the FFT results of Welch power spectrum estimation, and the detection accuracy of moving liquid surface is further improved. Finally, the above signal processing method is used to simulate the detection error of the pipe length within 130 meters in the experiment of the downhole casing. The feasibility of the signal processing method is verified.
【学位授予单位】：西安石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TE938

【相似文献】