From: m35products (m35prod@optonline.net)
Date: Sun Oct 03 2004 - 13:07:33 PDT
>From NCC1701spock@aol.com who is presently unable to post to this list:
At Atomics International in the early 1960's there was a lot of research
done concerning fluid flow through pipes. It was discovered that any sort
of perturbation [Note: TURBulent] in laminar flow in a pipe required a
length of twelve (12) diameters of the pipe without further disturbance to
restore laminar flow.
Laminar flow is parabolic in cross section along the axis of movement, with
the highest velocity at the center of the pipe. Just making the
measurements proved quite difficult because the instrumentation required
caused disturbances.
The upshot was that, for the purposes of pumping fluids through pipes to use
the fluid to transfer heat from one place to another, laminar flow is NOT
desirable; the heat tends to remain at or near the inner surface of the pipe
where the fluid velocity is zero (or, at least, not very fast). By
introducing turbulence into the flow, a much greater amount of heat is
transferred into and out of the fluid.
This all relates to the deuce question thusly: There is not any place in
the exhaust flow near the turbine housing where the flow proceeds
undisturbed for twelve pipe diameters; it is therefore not necessary to
place the thermocouple on the axis of the pipe. If it is placed against the
outside of the pipe, there will be some "lag" in reporting the temperature
of the gas inside the pipe, but the reading is not necessarily inaccurate by
virtue of being too low. Indeed, if the temperature of the gas inside the
pipe is decreasing after the thermocouple reading achieved equilibrium with
the [previous] temperature of the gas, the themocouple reading will higher
than the [present] gas temperature.
What does all of this mean? Well, as long as the thermocouple is reasonably
proximate to the place you want to measure and the thermal "coupling" to the
exhaust gas is good, you are probably getting a reading that is as accurate
as you need. There is a lot of math, particularly statistics, involved in
computing how close the reading is to what is "true" at a given moment; if
one were to take the square root of the sum of the squares of the errors of
each link in the chain, it is usually found that one can be 100% certain
that the measurement is within plus or minus 50% of the actual conditions.
Depending upon a lot of factors, it works out that one can be 90% certain
that the measurement is within 34% and so on. It is impossible to be 100%
cetain that the reading is correct, no matter what.
Put the thermocouple in some reasonable place and relax; it's probably close
enough.
Ed.
This archive was generated by hypermail 2.1.4 : Sat May 07 2005 - 20:36:49 PDT