Surface temperature and emissivity measurement for materials exposed to a flame through two-color IR-thermography

TL;DR

This study evaluates the accuracy of two-color IR thermography for measuring surface temperatures of materials exposed to flames, emphasizing calibration, material effects, and error reduction techniques for high-temperature fire testing applications.

Contribution

It introduces an improved calibration workflow and detailed error analysis for two-color IR thermography in fire resistance testing of composite materials and metals.

Findings

Proper calibration reduces temperature measurement errors below 50 K.

Emissivity values vary significantly between pristine and burnt samples.

Material-specific factors influence IR measurement accuracy.

Abstract

Two-color (2C) pyrometry has long been used for flame temperature and soot concentration studies and is now becoming more widely used to measure surface temperatures of burning materials. With the obvious advantage of being a contact-free method that requires only minimal optical access, 2C pyrometry combined with high-speed acquisition is a promising diagnostic tool to obtain exceptional temporal and spatial resolution of thermally degrading samples. However, its conceptual simplicity relies on a set of basic assumptions that when violated can result in large errors. In this work, we use an experimental configuration representative for fire resistance testing for aerospace and naval applications to analyze the impact of camera parameters and test setup on the accuracy of the surface temperature results obtained. Two types of fibre reinforced polymer composites and a steel plate are used to investigate material specific aspects that effect the measurements. An improved workflow for camera calibration is presented that takes the actual experimental setup into account. The temperature and emissivity mapping obtained trough in-situ IR measurements is compared against data acquired trough thermocouples and post-fire hemispherical directional reflectance measurements at room temperature. This comparison illustrates the necessity for proper post-processing and demonstrates that emissivity values obtained from pristine or burnt samples are not well suited to obtain accurate surface temperatures through conventional (single color) IR thermography. We also present a detailed error budget and suggestions for calibration measurements to keep the overall error well below 50 K in a temperature range from 673 K - 1473 K.