Round robin study of total heat flux gauge calibration at fire laboratories
Introduction
Many types of experimental measurements are important for fire science and engineering. In recent years there has been an increased emphasis on improving the quality of these measurements and the need to quantify their uncertainties. In 2000 the Building and Fire Research Laboratory (BFRL) at the National Institute of Standards and Technology (NIST) hosted a workshop on this topic entitled “Measurement Needs for Fire Safety: An International Workshop” at the request of the FORUM for International Cooperation in Fire Research [1]. One of the recommendations generated by the workshop participants was that a collaborative effort among the member laboratories should focus on characterizing the uncertainties associated with heat flux measurement in fire environments.
As a result of this meeting the FORUM Heat Flux Measurement Working Group was formed in July, 2000. During initial discussions it was discovered that a variety of different approaches were being employed in the laboratories to calibrate heat flux gauges. Schmidt-Boelter [2] and Gardon gauges [3] were typically used for total heat flux measurements. To the knowledge of the participants, the various calibration approaches had not been directly compared previously, and it was decided to conduct a calibration round robin in order to assess the consistency of the different approaches. Ultimately two independent round robins using Schmidt-Boelter and Gardon gauges were completed with five fire laboratories participating. Additionally, the four gauges were calibrated by the manufacturer, and the two gauges used during the first round robin were calibrated by two additional calibration laboratories, LNE in France and the Physics Laboratory (PL) at NIST.
The need for improved calibration facilities and procedures has been generally recognized in a variety of engineering fields requiring accurate heat flux measurement. A workshop held at NIST in 1999 summarized the needs of several industries, along with those of fire science [4]. The workshop proceedings include a summary of an effort within the International Standards Organization aimed at standardizing heat flux gauge calibrations for use in fire science and engineering studies. Much of the work supporting the development of this new standard has been performed as part of a European cooperative project known as HFCAL. Even though some laboratories have participated in both the HFCAL study and the current round robin, the investigations have been performed independently.
Section snippets
Round robin design
The fire laboratories participating in the round robin were BRE/FRS, FM Global, BFRL, SINTEF, and SP. The heat flux gauge calibration facilities utilized by the participating laboratories differ substantially from one another. An approach was adopted in which each fire laboratory carried out calibrations using their existing method following standard internal procedures.
The pattern (nomenclature utilized here is taken from the Guide for Interlaboratory Comparisons [5]) was the “basic circular”
BRE/FRS
The calibration of heat flux gauges for use as working standards is carried out by comparing their response at various levels of irradiance with the response of a secondary standard gauge at the same levels of irradiance. The measurements are made at multiple heat flux levels by varying the distance between the radiant source and the gauges.
The radiant source is a 0.3 m×0.3 m porous refractory burner operating on pre-mixed natural gas with air. It is mounted vertically and is operated in the
First round robin
Each fire laboratory reported the results for the calibrations of the two heat flux gauges in terms of the coefficients determined from linear least squares curve fits of the data when plotted as kW/m2 versus the response of the gauge in mV. In some cases the y-intercept was forced to pass through the plot origin, while in others it was allowed to “float,” and a y-intercept value from the fit is reported. Normal laboratory protocols were followed.
There are many different approaches that might
Discussion
Given the significant differences between the calibration methods used in the various fire laboratories, the agreement for the calibrations, as reflected by the variations from the average values for the nominal full-scale readings of the four gauges, is viewed as good. However, there are systematic variations in calibration results. SINTEF reported the lowest sensitivity for the two gauges tested during the first round robin, while BRE/FRS reported the highest. When the results from these two
Summary
The results of the two round robins of heat flux gauge calibrations have demonstrated that the agreement between the five fire laboratories is satisfactory for most fire research purposes and standard fire testing. Absolute values and variations in results between these laboratories are comparable to those reported by measurement laboratories. Systematic variations observed in the results suggest it should be possible to further improve agreement between calibration results from the fire
Recommendations
The role of convective heat transfer in the response of gauges sensitive to both radiation and convection is not very well understood and needs to be elucidated, both for calibration purposes and for their use in fire tests and other applications. The absorptivity, which is a property of the surface coating, also requires additional study because it affects the relative amounts of radiative and convective heat transfer to the gauge surface.
When the calibration is performed as a transfer
References (12)
- Ohlemiller T, Johnsson EL, Gann RG, editors. Measurement needs for fire safety. In: Proceedings of an international...
- Kidd CT, Nelson CG. How the Schmidt-Boelter gage really works. In: Proceedings of the 41st international...
- et al.
Transient response of circular foil heat-flux gauges to radiative fluxes
Rev Sci Instrum
(1975) - Grosshandler WL. Heat flux transducer calibration. In: Summary of the second workshop. National Institute of Standards...
- Guide for interlaboratory comparisons. Recommended Practice RP-15, National Conference of Standards Laboratories;...
- Fire tests—calibration and use of heat flux meters: part 2: primary calibration methods. ISO/DIS 14934-2, under FDIS...
Cited by (43)
A heat flux sensor leveraging the transverse Seebeck effect in elemental antimony
2023, Sensors and Actuators A: PhysicalNecessary condition for measurement of heat flux sensor using a thermoelectric module
2021, Fire Safety JournalFundamental measurement techniques
2021, Flammability Testing of Materials Used in Construction, Transport, and Mining, Second EditionA Heat-Transfer Rate Inducing System (H-TRIS)Test Method
2019, Fire Safety JournalCitation Excerpt :It is known that hot gases accumulated around heat flux gauges significantly influence their incident irradiation measurements [81,82], and that this may result in an underestimation of up to 20% when the gauge is placed flush with a target surface; this is due to free convective flows that are created by heating of the surface. Given that water-cooled heat flux gauges are calibrated under natural convection conditions, or alternatively in a vacuum [83], to minimise the influence of hot gases around the gauge, it is recommended that this type of heat flux gauge should not be mounted flush to a surface [81]. When calibrating H-TRIS the heat flux gauge was mounted on a self-cooling probe (Fig. 10).
Gardon gauge heat flux sensor verification by new working facility
2019, Measurement: Journal of the International Measurement ConfederationCitation Excerpt :NIST [6,7] has used a VTBB (Variable-Temperature Black Body) as heat flux source generator up to 50 kW·m−2 with a Schmidt-Boelter [8] type reference sensor, assuming axial one-dimensional heat flow, a range 0–110 kW·m−2 of heat flux and 200 °C as operative temperature limit. Another calibration system has developed at NASA’s Dryden Flight Research Center that has considered also a numerical method to evaluate radiant heat flux calibration uncertainties [9]. The aim was to reduce numerically, if possible, errors in the calibration system.