Uncertainty analysis of thermocouple measurements used in normal and abnormal thermal environment experiments at Sandia's Radiant Heat Facility and Lurance Canyon Burn Site.

Abstract

It would not be possible to confidently qualify weapon systems performance or validate computer codes without knowing the uncertainty of the experimental data used. This report provides uncertainty estimates associated with thermocouple data for temperature measurements from two of Sandia's large-scale thermal facilities. These two facilities (the Radiant Heat Facility (RHF) and the Lurance Canyon Burn Site (LCBS)) routinely gather data from normal and abnormal thermal environment experiments. They are managed by Fire Science & Technology Department 09132.Uncertainty analyses were performed for several thermocouple (TC) data acquisition systems (DASs) used at the RHF and LCBS. These analyses apply to Type K, chromel-alumel thermocouples of various types: fiberglass sheathed TC wire, mineral-insulated, metal-sheathed (MIMS) TC assemblies, and are easily extended to other TC materials (e.g., copper-constantan). Several DASs were analyzed: 1) A HewlettPackard (HP) 3852A system, and 2) several National Instrument (NI) systems. The uncertainty analyses were performed on the entire system from the TC to the DAS output file. Uncertainty sources include TC mounting errors, ANSI standard calibration uncertainty for Type K TC wire, potential errors due to temperature gradients inside connectors, extension wire uncertainty, DAS hardware uncertainties including noise, common mode rejection ratio, digital voltmeter accuracy, mV to temperature conversion, analog to digital conversion, and other possible sources. Typical results for "normal" environments (e.g., maximum of 300-400 K) showed the total uncertainty to be about ±1% of the reading in absolute temperature. In high temperature or high heat flux ("abnormal") thermal environments, total uncertainties range up to ±2-3% of the reading (maximum of 1300 K). The higher uncertainties in abnormal thermal environments are caused by increased errors due to the effects of imperfect TC attachment to the test item. "Best practices" are provided in Section 9 to help the user to obtain the best measurements possible.

A joint computational and experimental study to evaluate Inconel-sheathed thermocouple performance in flames.

“…6, the 1.0 mm diameter thermocouple reached a 50 K-higher steady state temperature than the 1.6 mm diameter thermocouple, with an uncertainty of 0.75% or ±5 K (95% confidence), based upon the manufacturer's estimate. A more detailed uncertainty analysis of thermocouple measurements is reported by Nakos [7]. As shown in Fig.…”

Validation experiments to determine radiation partitioning of heat flux to an object in a fully turbulent fire.

“…The following summarizes the description and uncertainty estimates for TC DASs for SNL's experimental facilities that routinely gather data from normal and abnormal thermal environment experiments, specifically the Penlight apparatus and the FRH facility. In most respects, the systems and uncertainties are the same as for two of SNL's experimental facilities that have previously been used for the same purposes; the Radiant Heat Facility (RHF) and Lurance Canyon Burn Site (LCBS) [Nakos, 2004]. The uncertainty of the entire TC DAS was estimated by looking at the individual uncertainties of each component, then combining them using the root-sum-square (RSS) method [Coleman and Steele, 1999].…”

“…Computer & Displays Nakos [2004] performed uncertainty analyses for several TC data acquisition systems used at the RHF and LCBS. These analyses apply to Type K, chromel-alumel thermocouples in MIMS TC assemblies and other applications.…”

“…Nakos et al [2004] showed that intrinsic TCs, are more accurate than MIMS TCs and that error in MIMS TCs increases roughly linearly with TC diameter. Based on recent work Nakos, 2004], a total uncertainty of ±2% for 40 mil TCs in the calorimeter will be assumed throughout the measurement range of these experiments. This uncertainty estimate is valid if the additional error due to mounting is less than 7 K at a temperature of 373 K and less than 25 K at a temperature of 1273 K.…”

Theory and experimental validation of SPLASH (Single Panel Lamp and Shroud Helper).

“…The temperature errors associated with the thermocouple readings have been studied [4] and an uncertainty of about 0.6% of the absolute temperature reading is used in this study. Considering TC error, mounting, data acquisition system, and installation errors typical total measurement errors for similar systems have been estimated [5] as 2-3%. Because of uncertainty in the thermocouple readings due to the finite size of the thermocouple required for durability at high temperatures, a two color pyrometer (Micron Model M668) was also used to measure temperature on the bottom surface of the shroud.…”