Prediction of the normal spectral emissivity of metals with different surface roughness using a modified ray-tracing method
Shixiong Liu,Qitai Eri,Bo Kong,Yue Zhou
DOI: https://doi.org/10.1016/j.ijthermalsci.2024.109147
IF: 4.779
2024-05-17
International Journal of Thermal Sciences
Abstract:Spectral emissivity is one of the most important parameters of metal radiation, with considerable influence on heat-transfer calculations, infrared signatures, pyrometry, and other high-temperature fields. Therefore, a computational model that can accurately calculate the emissivity of metal surfaces is crucial in these fields. In this study, a modified ray-tracing method was proposed to predict the normal spectral emissivity of a rough metal surface. In contrast to other methods involving a complex refractive index, this approach uses the known spectral emissivity of smooth surfaces as an input to avoid using complex refractive index data, which are more difficult to obtain. In addition to the emissivity data of pure Co and advanced high-strength steels samples (DP980) in the literature, emissivity measurements were conducted on GH5188 superalloy with different roughness values to further verify the modified ray-tracing method. A comparison of the measurement and prediction results revealed the high accuracy of the modified ray-tracing method, with the average and maximum errors of 1.92 % and 9.49 %, respectively, which are within the applicable range. As the new method is based on a geometric optics approximation ray-tracing approach, the model is primarily applicable to short-wavelength bands. Additionally, it has been observed in practical use for cases with low roughness. Moreover, some cases with long wavelength bands that exceed the geometric approximation can still yield highly accurate predictive results. When examining the distinction among various processing methods, higher accuracy can be achieved through two-dimensional measurements of the actual surface for prediction. In contrast, simplified methods, such as using one-dimensional cross-sections or surfaces generated based on the surface roughness for the prediction, yield less precise results.
engineering, mechanical,thermodynamics