Assessment of a common nonlinear eddy-viscosity turbulence model in capturing laminarization in mixed convection flows

Figures & data

Figure 1. Normalized Nusselt number distribution in ascending mixed convection flows.

Table 1. DNS cases from You et al. [18].

Case	Gr/Re^2	Bo	Thermal-hydraulic regime
A	0	0	Forced convection
B	0.252	0.13	Early-onset mixed convection
C	0.348	0.18	Laminarization
D	0.964	0.50	Recovery

Table 2. Functions appearing in the LS and Suga turbulence models.

Variable	LS k-ε model	Suga k-ε model
	− νtSij + (2/3)kδij
Cμ	0.09
fμ		1 − exp[ − (Ret/90)^1/2 − (Ret/400)^2]
Eε

Table 3. Constants appearing in the LS and Suga models.

Constant	LS k-ε model	Suga k-ε model
σk	1.0	1.0
σε	1.3	1.3
Cϵ1	1.44	1.44
Cϵ2	1.92	1.92

Table 4. Constants appearing in the Reynolds shear stress equation of the Suga model.

c1	c2	c3	c4	c5	c6	c7
−0.1	0.1	0.26	−10 Cμ^2	0	−5 Cμ^2	5 Cμ^2

Table 5. Results for fully developed forced convection.

Model/Technique	Nu0	% diff.	cf0	% diff.
DNS of You et al. [18]	18.3	–	9.28 × 10^−3	–
Launder–Sharma model	17.4	−4.9	8.52 × 10^−3	−8.2
Suga model	18.3	0	8.93 × 10^−3	−3.8

Figure 2. Normalized friction coefficient distribution in ascending mixed convection flows.

Figure 3. Mean flow profiles obtained using the Suga and Launder–Sharma models.

Figure 4. Turbulence parameters of the Suga and LS models.

Figure 5. Turbulent viscosity ratio and damping function from the Suga and LS models.

Figure 6. Budgets of turbulent kinetic energy obtained using the Suga and Launder–Sharma model; (a) Case A (forced convection) and (b) Case C (laminarization).

Figure 7. Distribution of various low-Reynolds number functions in the Suga model.

Figure 8. Effects of setting C_μ = 0.09 in the Suga model.

Figure 9. Effects of using a new C_μ expression in the Suga model.

Figure 10. Effects of replacing the E-term of the original Suga model with that of the LS model.