Stories from the Field,Fundamental Questions and Solutions - Measurement While Drilling Signal Analysis, Optimization and Design - page 22

1.2.5.4 Wind tunnel analysis: studying new approaches.
While computer models are useful screening tools, they alone are not
enough. Gridding effects mask the finest flow details that can be uncovered
only through actual testing. The use of wind tunnels in modeling downhole mud
flow was first proposed and used by the last author of this paper during his
tenure with Schlumberger. Technical details and justification are disclosed, for
instance, in Gavignet, Bradbury and Quetier (1987) who used the method to
study flows beneath drill bits nozzles. This counter-intuitive (but correct)
approach to modeling drilling muds provides a strategically important
alternative to traditional testing and reduces the time and cost of developing new
MWD systems.
The CNPC MWD wind tunnel test facility consists of two components, a
“short flow loop” where principal flow properties and tool characteristics are
measured, and a “long loop” (driven by the flow in the short wind tunnel)
designed for telemetry concept testing, signal processing and noise removal
algorithm evaluation. Field testing procedures and software algorithms for tool
properties and surface processing are developed and tested in wind tunnel
applications first and then moved effortlessly to the field for evaluation in real
mud flows. This provides a degree of efficiency not possible with “mud loop
only” approaches.
Our “short wind tunnel,” actually housed at a suburban site, is shown in
Figure 1.5a. This laboratory location was selected because loud, low-frequency
signals are not conducive to office work flow. The created signals are as loud as
motorcycle noise and require hearing protection for long duration tests. More
remarkable is the fact that internal pipe pressures are several orders of
magnitude louder than the waves that escape - in practice, this is further
multiplied by the (large) ratio of mud to air density, about 800 in the case of
water. Thus, careful and precise acoustic signal measurement is required to
accurately extrapolate those to mud conditions. Similarly, torques acting on
sirens in air are at least 800 times lower. In fact, air-to-mud torque scaling is
simply proportional to the dynamic head “UU 2 ” ratio, where U is the oncoming
speed. Thus, wind tunnel tests can be run at lower speeds with inexpensive
blowers provided a quadratic correction factor is applied for downhole flow
extrapolation. The MWD turbine, similarly designed and tested, is not discussed
in this paper.
In Figure 1.5a, a powerful (blue) blower with its own power supply pumps
more or less constant flow rate air regardless of siren blockage. A sensitive flow
meter is used to record average flow rate. Flow straighteners ensure uniform
flow into the siren and to remove downstream swirl for accurate differential
pressure measurement. The siren test section deserves special comment. The
motion of the rotor is governed by its own electrical controller and is able to
effect position-modulated motions as required for telemetry testing.