Page 146 - Thermal Hydraulics Aspects of Liquid Metal Cooled Nuclear Reactors

Construction of experimental liquid-metal facilities 119

end, auxiliary heaters in the test section must be turned off so that the temperature profile in
the wall depends only on the conditions inside the test section.
Related to the previous point, it should be noted that the transient evolution of some variables
l
has larger characteristic times than others. As a general rule, it is important to confirm sta-
tionary conditions for all relevant variables.
l The presence of gas bubbles or oxide particles in the test section can lead to problems of
nonrepeatability of the measurements. Thus, practical solutions for purging and filtering
must be implemented and even repeated during a campaign if necessary.

3.3.3.2 In-situ calibration of instruments and data
acquisition chain

These additional considerations are particularly relevant for local temperature mea-
surements using thermocouples (TCs) for two main reasons: their tolerances and small
electric signals.
Firstly, as relatively low-temperature differences are involved in liquid-metal
experiments, very accurate measurements are required, and an additional calibration
might be necessary. TCs are designed with a letter code, according to the temperature-
voltage tables matched within defined tolerances. For liquid-metal thermohydraulic
experiments, types K and N are preferred, due to their high sensitivity and extensive
temperature range. In the industrial standard DIN-EN-60584, two tolerance classes
are defined that, for example, for TCs, types K and N at 400°C are 3.0 and
1.6K. This means that, even with the best available tolerances, the measured tem-
perature profiles might not make physical sense, particularly at low heat-flux densi-
ties. It should be noted that these tolerances refer to the absolute temperature, while the
relative differences between TCs and their repeatability are usually better. However,
careful calibration is required in any case, as will be discussed later in this section.
Secondly, the electric signals generated by the TCs based on the Seebeck effect are
in the approximate range 0–20mV for 0–500°C (type K). While this range is techni-
cally acceptable for measurements, high accuracy is needed. The mean Seebeck coef-
ficient of ca 40μVK 1 translates directly into a sensitivity of 0.025KμV 1 for the
voltage measurement. As a consequence, relatively small effects such as the resolution
of analog-to-digital converter and the temperature drift of data acquisition cards must
be taken into account. Moreover, spurious signals must be avoided, with particular
focus on two sources:
l The influence of electromagnetic noise can be reduced by proper shielding, especially from
known sources of interference, as might be an EM pump. A complex scenario is, for exam-
ple, TCs located directly at the wall of high-current electric heaters. Eventually, if the appli-
cation allows for it, the signal quality can be improved with a frequency filter.
Regarding the cold-junction compensation, for example, using an electronic ice point, spu-
l
rious signals might be superimposed if the reference temperature is not the same for all
channels and/or it does not remain constant during the data acquisition time period.
Although these issues are best-practice guidelines for temperature measurement in
general, their importance is more significant for liquid-metal tests due to the small
differences to be observed.