Home

The noise log or survey, also sometimes called the sound survey (Sonan Log, Borehole Audio Tracer Survey (BATS), AcoustiSonde Log, and others) is essentially a very sensitive detector of the sound produced by fluid flow.  The noise survey is a relative newcomer to production logging.  It was described by ARCO in 1955 as a qualitative tool, but the industry attitude was that it was no better than, if as good as, the temperature log for leak or channel detection.  But in 1973, Dr. R.M. "Mac" McKinley pointed out the utility of the noise log, both qualitatively and quantitatively, and thus was born modern noise logging, with Dr. Mac as its daddy.

The sounds of moving fluids or the hiss of escaping gas are caused by disturbances in a liquid / gas interface or by turbulence in the fluid stream.  It is not surprising that the sounds of this turbulence can be used to detect flow.  In a wellbore environment, the noise log is very effective for gas detection as it flows up through liquid, but it is also effective for the detection of various kinds of gas, water, or oil single phase flow, including channeling behind pipe (assuming there is adequate turbulence in any given situation to produce enough noise).  The noise tool itself is nothing more than a microphone (hydrophone), and associated amplification / line driving circuitry, in a pressure housing constructed to withstand the normal trials and tribulations in the life of any logging tool.  Surface equipment again amplifies the signal, and further processes it.  There are two types of noise surveys, the stationary survey with measurements made at various stations downhole, and the much less common continuous survey, usually focusing on higher audio frequencies, and used for gas entry and leak detection in casing.

Noise logs use very sensitive microphones to be able to measure sound produced by fluid flow in a well.  Measurements of sound within the audible range of frequencies (20-20,000 Hz, except for us older males with reduced high frequency response) are usually most indicative of turbulent flow behind pipe.  Stationary measurements are made in the more common noise logging methodology because tool and cable noise from scratching the casing wall would otherwise dominate the record.  In shallow wells, even machinery noise at surface can be a problem.  Surface equipment conditions the tool audio signal with typically four high-pass filters, 200, 600, 1,000, and 2,000 Hz (sometimes 4,000, 6,000 Hz, or others).  In addition to these filter "cuts", the full spectrum signal should be available, and headphones and/or a speaker should be provided.  At each depth stop, at least four noise frequency readings are taken.  Usually a coarse grid of station spacing of 50 or 100 feet is used to locate high noise areas, then much closer spacing can be employed for detailed investigation, finally down to every two or three feet.

The pertinent record features are (1) sound levels on all four cuts or "loudness", (2) variation in level of a particular cut from station to station or "character", and (3) the frequency content of the sound at a peak in level or "pitch".  Loudness relates to the severity of a problem, character to how the flow is taking place, and pitch to the type of flow (single phase, or gas through liquid).  Single phase flow through a constriction displays an amplitude peak at about 1,000 Hz, and an increase in pressure drop across said constriction increases frequency.  Two phase gas / liquid flow generates noise at lower frequencies. 

In addition to the applications mentioned above, the noise survey has been used to measure flow rate in the wellbore, measure flow rates from individual perforations, detect liquid production from gas perforations, detect sand production, and detect problems during drilling (most of these applications were also pointed out by Dr. McKinley).  In the US EPA underground injection control (UIC) program, at 40 CFR §146.8(c)(1), the noise log is authorized as a mechanical integrity test (MIT) for the demonstration of external integrity of injection wells.

The noise log is only useful if the noise generated by the phenomenon under examination is greater than background noise; take appropriate steps to limit surface noise to the extent practicable (shut down nearby machinery, etc.).  Noise tools should be run without centralizers to maximize acoustic coupling.  Prior to recording stationary measurements, wait 60 seconds or so for tool and line motion noise to dissipate.  A temperature log should usually be run with noise surveys.  Hire Dr. R.M. "Mac" McKinley to tell you what the heck your noise log means when you get done...

The foregoing is an oversimplified discussion of the noise log; most theoretical background and interpretative information has been omitted. 

10-02-00
Last 10-22-10