The multispectral gas radiation modeling: A new theoretical framework based on a multidimensional approach to k-distribution methods

The Scaled SLW model of gas radiation in non-uniform media based on Planck-weighted moments of gas absorption cross-section

“…The high resolution spectra based on the HITEMP2010 [13] spectral database for H2O, and the CDSD-4000 [26] database for CO2 were used with spectral discretization ∆η = 0.01 cm -1 . More details about these spectral databases can be found in [27]. Detailed comparisons, not reported here, have shown that the use of the CDSD-4000 database instead of HITEMP2010 for CO2 in the construction of the SLW model database does not provide significant differences in full spectrum uniform scenarios.…”

“…However, because real spectra are not perfectly scaled Eqs. (27) and (28) yield, in general, different results. This will be demonstrated in a section to follow.…”

Functional data analysis applied to the multi-spectral correlated–k distribution model

“…This is mainly due to the appearance of so-called "hot lines" [13] that breaks the one-to-one correspondence between gas spectra assumed in Eq. (1).…”

“…This leads to the breakdown of the assumption of rank correlation of gas spectra in distinct states when large temperature gradients exist in the medium studied [9]. In order to circumvent this problem, we introduced a method based on functional data analysis (FDA) [10] and called the MSCk model [1,2]. The main ideas behind the MSCk approach are not new, as they share similar concepts as used in mapping methods such as described in West and Crisp [11] or in the multiscale method of Zhang [12].…”

“…The main ideas behind the MSCk approach are not new, as they share similar concepts as used in mapping methods such as described in West and Crisp [11] or in the multiscale method of Zhang [12]. The objective of this method is in fact to group together wavenumbers according to the spectral scaling functions [1] defined as the ratio between spectra in different thermophysical states. Over these intervals, the correlation assumption can be considered as exact.…”