Stories from the Field,Fundamental Questions and Solutions - Measurement While Drilling Signal Analysis, Optimization and Design - page 27

Figure 1.5c. A pair of ganged or tandem mud sirens.
Additional signal enhancement is possible using constructive interference
of a different nature: multiple sirens arranged in series or in tandem. If the
distance between sirens is small and siren apertures are properly phased, signals
will be additive. This idea was first proposed in U.S. Patent 5,583,827 or Chin
(1996) and a possible design from that publication is reproduced in Figure 1.5c.
This design, incidentally, is not CNPC's preferred embodiment.
Two sirens, for instance, mean twice the signal. If the amplification
afforded in the previous paragraphs provides a modest signal gain of 1.5, that is,
50%, the net would be a three-fold signal increase more than enough to
overcome attenuation at the higher frequencies used. In principle, any number of
sirens can be connected to provide signal increases as needed. Performance is
determined by the single transducer in Figure 1.5b.1 (second row, far left) which
measures the signal leaving the MWD collar. The extent to which constructive
wave interference works is found by comparison with the measured differential
pressure 'p taken across the siren (e.g., see Figure 1.5h, middle and right). This
'p depends on siren geometry, flow rate and rotation speed only: it is
independent of reflections since waves pass through without interaction (that is,
reflected waves do not affect differential pressure measurements providing both
sensing ports are close).
Note that Figures 1.2b,c,d suggest that frequencies in the 50-60 Hz range
are not unrealistic, a conclusion independently reached at the website
www.prescoinc.com/science/drilling.htm (see Figure 10.5). This use of higher
frequencies is also supported by test results from actual flow loop tests with real
muds. We stress that attenuation measurements are subtle since the effects of
acoustic nodes and antinodes (which depend on frequency and flow loop
boundary conditions) must be properly accounted for. Almost all existing
papers on attenuation fail to even recognize this problem, let alone correct for it.
Our systems approach to high-data-rate design requires an equal focus on
surface systems. As implied earlier, signal strength enhancement must be
accompanied by using the most sensitive piezoelectric transducers and robust
multiple-transducer echo cancellation methods. Figure 1.5b.2 shows a
transducer array located far from the test shed. Noise can be introduced by
playing back actual field recordings. We have found, to our amusement, that the
large firecrackers used at Chinese weddings, e.g., see Figure 1.5o, provide a
useful source of low-frequency, plane-wave “pump” noise when all else is
unavailable.