Sensitivity Analysis for a Guarded-Hot-Plate Apparatus: A Methodology Based on Orthogonal Experiment Designs

Results and analyses of a sensitivity study of six controlled variables on the response of the National Institute of Standards and Technology (NIST) 500mm guarded-hot-plate apparatus are presented. The effects of four factors held constant as well as three uncontrolled environmental variables were also examined. The goal of the study is to derive a sensitivity analysis ranking of the relative importance of factors and interactions affecting the apparatus. Sixty-six thermal conductivity measurements were conducted across three experiments at a mean temperature of 310 K for a pair of fibrous-glass specimens (120 kg.m(-3)) having nominal dimensions 500mm in diameter and 26mm in thickness. The apparatus response was studied using an orthogonal fractional factorial design, a one-factor-at-a-time design, and a full factorial design for a subset of factors from the fractionated design. The results indicate that most important factor affecting the thermal conductivity measurement was the temperature difference across the air space separating the central meter plate and the surrounding guard plate, described here as the gap temperature difference (Delta T-g). The study also revealed an interaction between the gap temperature difference and the temperature difference across the specimen (Delta T-avg). An empirical model for the results of the sensitivity study is presented. Results of the gap temperature difference (Delta T-g) are similar to published results from another guarded-hot-plate apparatus. Improvements for equipment operation, as well as insights to the sources of experimental uncertainty, are presented.