Questions about viscous terms

[tiandixuanhuang1]

Dear authors：

Hello and thank you for providing such an excellent open source toolkit! I learned a lot from this kit.

I have a question about the code in the toolkit (using OpenFoam version V8), and there is a piece of code in the UEqn.H file

fvVectorMatrix UEqn
(
(1./eps)(fvm::ddt(rho, U) + fvm::div(rhoPhiByEps, U) + MRF.DDt(rho, U))
+ (1./eps)(turbulence->divDevTau(rho, U)) // Why is it multiplied by 1/eps?
+ fvm::Sp(Drag,U)
==
fvOptions(rho, U)
);

This code is a finite volume discrete assembly of the macroscopic momentum equation multiplied by a system of linear algebraic equations, which I think is the corresponding equation (refer to the formula (21) in the paper “Multiphase flow modeling in multiscale porous media:An open-source micro-continuum approach”).

$\frac{1}{\phi}(\frac{\partial \rho\bar{\pmb{v}}}{\partial t} + \nabla\cdot(\frac{\rho}{\phi}\bar{\pmb{v}}\bar{\pmb{v}})) = -\nabla\bar{p} + \rho\pmb{g} + \nabla\cdot\bar{S} - \mu k^{-1}\bar{\pmb{v}} + \pmb{F}_c$

One of the things I am confused about is why the viscous term $\nabla\cdot\bar{S}$ is multiplied by 1/eps in the code (shown below), because the equation (21) does not multiply this coefficient, is there any key information I am missing?
(1./eps)*(turbulence->divDevTau(rho, U))

I am looking forward to your reply and best wishes : )