Abstract
The present investigations provide a pathway for implementation of soft computing based Adaptive Neuro-Fuzzy Inference System (ANFIS) technique for prediction of surface heat flux from short duration temperature measurement in shock tubes or shock tunnels. Computational modeling of a co-axial thermal probe is carried out to get the necessary temperature-time histories for different temporal variations of applied heat loads. Different possible inputs are assessed while defining the most suitable ANFIS structure for the recovery of step or ramp heat loads. This proposition is then tested for recovery of heat flux in a given range or of given time history. In each case, the uncertainty band is found to be in the acceptable range. The final assessment of this novel methodology is performed for recovery of heat flux signal from temperature measurement in a shock tube-based experiment. An in-house fabricated fast response coaxial thermal probe (CTP), prepared from chromel (3.25 mm diameter and 10 mm length) and constantan (0.91mm diameter and 15 mm length) is employed for these experiments. The surface heat flux recovered from the experimental signal using ANFIS is seen to have excellent agreement with the conventional analytical method in terms of both trend and magnitude, within an uncertainty band of ± 2%. Therefore present investigations advocate the use of soft computing technique for heat flux recovery in a short duration temperature measurement due to its accuracy of prediction, lesser complexities in mathematical modeling, and being less computationally intensive.
Surface Heat Flux Measurement in Transpiration-Cooled Porous Materials using Plenum Pressure Data