Technique for acoustic data analysis

• US 7,643,945 B2
• Filed: 05/23/2007
• Issued: 01/05/2010
• Est. Priority Date: 12/28/2006
• Status: Expired due to Fees

• Current Assignee: Schlumberger Technology Corporation (SLB)
• Original Assignee: Schlumberger Technology Corporation (SLB)
• Inventors: Wago, Toshimichi, Baklanov, Nikolay, Adnan, Sarmad
• Primary Examiner(s): Bui; Bryan
• Assistant Examiner(s): TEIXEIRA MOFFAT, JONATHAN CHARLES
• Application Number:
  US11/752,657
• Publication Number:
  US 20080162058A1
• Time in Patent Office: 958 Days
• Field of Search: 702/39, 702/48
• US Class Current: 702/39
• CPC Class Codes: F04B 2201/0802 Vibration
• CPC Class Codes: F04B 51/00 Testing machines, pumps, or pumping installations
• CPC Class Codes: G01M 3/24 using infrasonic, sonic, or ultrasonic vibrations

Abstract

A method of enhancing acoustic pump data for monitoring of pump condition in operation. The enhancement may be in the form of a substantial reduction in the amount of data to be processed by firmware to a much more manageable level. This may be achieved through pre-filtering of acquired acoustic data, computing frequency momentum, and by application of a median filter to the acoustic data. Regardless, a thresholding technique that uses features vector including but not limited to double-threshold derived from statistics of enhanced amplitudes, may be employed whereby enhanced acoustic data is boiled down to a series of potential acoustic events of particular widths. These widths may represent a duration of each acoustic event and be checked for correlation with the speed of the operating speed of the pump. Acoustic events that correlate with an operating speed of the pump may represent an unhealthy condition of the pump such as where the width is about the duration of one half of a cycle of the pump and a leaky pump valve presents itself for this same period of time.

Claims

1. A method of analyzing acoustic data of a pump, the method comprising:

operating the pump at a given speed;

recording the acoustic data from the pump over a series of time periods during said operating;

averaging acoustic data of the multiple cycles to minimize an effect of any inconsistent acoustics by stacking waveforms from the recorded acoustic data into a single averaged waveform;

enhancing the acoustic data for analysis;

determining upper and lower thresholds of magnitude for the recorded acoustic data based on the enhanced data; and

thresholding the acoustic data for identifying any acoustic event correlated to the speed.

2. The method of claim 1 wherein said thresholding comprises determining a width of any acoustic event, the width representing a duration of the acoustic event, the method further comprising employing a processor to compare the speed and the duration for the identifying.

3. The method of claim 2 wherein the duration is about half a stroke of the operating pump to indicate a valve leak in the pump.

4. The method of claim 1 further comprising:

noting a beginning of an acoustic event once a magnitude of the enhanced acoustic data crosses the upper threshold from below the lower threshold; and

noting an end of the acoustic event once the magnitude of the enhanced acoustic data crosses the lower threshold from above the upper threshold, the width defined by the distance between the beginning and the end.

5. The method of claim 1 wherein said establishing comprises applying a histogram technique to the enhanced acoustic data.

6. The method of claim 1 wherein said recording is of a predetermined frequency range.

7. The method of claim 1 wherein said recording occurs via an acoustic sensor coupled to a white noise filter.

8. The method of claim 1 wherein said recording is established over a plurality of frequency zones for independent analyses thereof.

9. The method of claim 1 wherein said enhancing comprises amplifying a magnitude of the acoustic data over a frequency range.

10. The method of claim 9 wherein said amplifying is applied over the frequency range in one of a linear manner and an exponential manner.

11. The method of claim 9 wherein said amplifying is applied to a greater degree at a higher end of the frequency range to enhance a presentation of the acoustic data resulting from an early stage valve leak in the pump.

12. The method of claim 1 wherein said enhancing comprises smoothing a presentation of the acoustic data with one of weighted exponential averaging and an FIR implementation of a Hilbert transform.

13. The method of claim 1 wherein enhancing comprises at least computing frequency moments of the acoustic data.

14. The method of claim 1, wherein recording comprises obtaining data in a particular frequency range with the use of a short-time fast Fourier transform (STFFT).

15. A method of enhancing acoustic data of a pump, the method comprising:

operating the pump;

recording the acoustic data from the pump in a predetermined frequency range during said operating;

amplifying a magnitude of the acoustic data over the frequency range;

integrating the amplified magnitude of the acoustic data and subsequently removing a frequency dimension of the acoustic data by stacking waveforms from the acoustic data into a single averaged waveform;

enhancing the amplified magnitude of the acoustic data by application of a median filter, said recording in the predetermined frequency range and said enhancing reducing storage of the acoustic data that is noise.

16. The method of claim 15 wherein the predetermined frequency range is between about 1 KHz and about 25 KHz.

17. The method of claim 15 wherein said recording occurs over periods of at least about 4/3 of a cycle of the operating pump.

18. The method of claim 15 wherein said operating is at a known speed, the method further comprising thresholding the acoustic data for identifying any acoustic event correlating with the speed.

19. The method of claim 15, wherein stacking comprises averaging acoustic data to minimize an effect of any inconsistent acoustics.

20. The method of claim 15, wherein recording in a predetermined frequency range comprises obtaining data with the use of a short-time fast Fourier transform (STFFT).

21. An oilfield pump assembly comprising:

a pump for operating at a given speed;

an acoustic sensor coupled to said pump for acquiring acoustic data therefrom during said operating; and

a processor coupled to said acoustic sensor for acquiring information regarding the acoustic data and the speed, said processor identifying any acoustic event correlated to the speed by determining upper and lower thresholds of magnitude for the acoustic data based on the acquired data and subsequent thresholding analysis of the acquired acoustic data by stacking waveforms from the acoustic data into a single average waveform.

22. The oilfield pump assembly of claim 21 wherein the acoustic event is an unhealthy condition of the pump.

23. The oilfield pump assembly of claim 22 wherein said pump further comprises a valve seal disposed therein, the unhealthy condition being a leak in said valve seal.

24. The oilfield pump assembly of claim 23 wherein the unhealthy condition of the leak is identified by an acoustic event persisting for about the duration of one half of a stroke of the pump.

25. The oilfield pump assembly of claim 21 wherein said pump is of a triplex configuration for a fracturing operation.