Page 162 - Mechanical Engineers' Handbook (Volume 2)

2 Thermocouples 151

Radiation error plays an important role at high temperatures, especially in low-velocity
situations. One approach to this problem is to use an aspirated probe to increase the gas
velocity past the thermocouple junction, increasing the convective heat transfer coefficient
and reducing radiation error. Suction pyrometer probes were investigated at the National
Bureau of Standards during the 1950s for use in combustion chamber and afterburner studies,
as reported by Lalos. 21 Norton et al. 22 adapted this method for use in a high-temperature
glass furnace. Another approach is to use two probes with different (but predictable) error
sensitivities and calculate the radiation error from the difference in the readings of the two
24
23
probes. This approach has been used by Moffat and Brohez et al. The main problem with
this approach is that the correction can be very large, and if the uncertainty in the raw data
exceeds a few percent, then the correction becomes highly uncertain. The behavior of ther-
mocouple probe designs in fire environments was investigated by Blevins and Pitts, 25 who
modeled the behavior of bare thermocouples, single-shielded aspirated probes, and double-
shielded aspirated probes under fire conditions. Pitts et al., 26 of NIST’s Building and Fire
Research Laboratory, investigated the uncertainties in the corrected readings of bare and
shielded, aspirated probes. Their results confirmed that the largest uncertainties were found
when the corrections were largest.
Measurements in high-velocity gases suffer mainly from problems with the recovery
factor of the probes. High-recovery-factor probes read close to stagnation temperature, but
their recovery factor is sensitive to manufacturing tolerances and, therefore, is somewhat
uncertain. Low-recovery-factor probes require larger corrections for the residual velocity
error, but their simpler geometry can reduce the uncertainty in the final reading. Moffat 27
recommended a probe using a spherical tip (a ball bearing) whose recovery factor is very
well known and repeatable, while Vasquez and Sanchez 28 describe the design of a system
using a Kiel-type shield for a high-recovery probe. Both references dealt with applications
requiring high accuracy in the measurement of small temperature differences—the temper-
ature rise across a single stage of a compressor or turbine.

2.14 Thermocouple Installations for Surface Temperature Measurement

Two problems dominate the accuracy of surface temperature measurement with thermocou-
ples: ensuring good contact between the sensor and the surface and avoiding disturbing the
surface temperature by the presence of the thermocouple.
Keltner and Beck 29 analyzed the steady-state and transient errors of thermocouples at-
tached to thick walls using the unsteady surface element method. Sobolik, Keltner, and
Beck dealt with thin plates, with special emphasis on the effect of the measurement errors
30
on the interpretation of data from thin-foil heat flux gages.
Analysis of the surface measurement problem shows that thin-film thermocouples should
be ideal sensors. Han and Wei 31 describe a method of making thin-film thermocouples on
nonmetallic surfaces that yields excellent steady-state and transient measurement accuracy.
They applied this technique to a Zirconia coating on an engine piston, but the thin-film
concept would also be well suited to the electronics industry, which uses nonmetallic pack-
aging and requires accurate measurement.
Not all thermocouple probes are attached to the surface—spring-loaded contact probes
are often used for measurements in situations where no permanent attachment would be
permitted. The error in such applications is dominated by the contact resistance. Osman,
Eilers, and Beck 32 used an inverse conduction solution to develop a method for correcting
transient measurements based on a steady-state calibration.