Dear Saad, I am solving two phase flow with Nektar++. In my case, I should alter viscosity with the Phase function nu(phi). I implement variable viscosity by m_var in the HelmSolve function. m_fields[i]->HelmSolve(F[i], m_fields[i]->UpdateCoeffs(), NullFlagList, factors, m_var); that m_var is m_var[StdRegions::eVarCoeffD00] = VarCoeff; m_var[StdRegions::eVarCoeffD11] = VarCoeff; and VarCoeff = nu(phi) and in your case is nu(T). Noting that if temperature vary in the domain with time you should add factors[StdRegions::eFactorTime] = time; before the HelmSolve. This explanation also has been described in the LaplaceSolve (https://doc.nektar.info/doxygen/4.1.0/class_nektar_1_1_multi_regions_1_1_con... (https://doc.nektar.info/doxygen/4.1.0/class_nektar_1_1_multi_regions_1_1_con...)) , see the LaplaceSolve). By using FactorTime, if you do not use the last version of the nektar++, you will encounter memory leakage. This issue was solved in the last version of the Nektar++. Also if the coefficient is time-dependent, the Cholesky decomposition needs to be updated every step. The simulation would become very slow. Best Regards, Mohammad 14 Aabaan 1401 2:01 PM, "Debbahi Saad" <saad.debbahi@gmail.com (mailto:saad.debbahi@gmail.com?to=%22Debbahi%20Saad%22%20<saad.debbahi@gmail.com>)> wrote: This email from saad.debbahi@gmail.com (mailto:saad.debbahi@gmail.com) originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list (https://spam.ic.ac.uk/SpamConsole/Senders.aspx) to disable email stamping for this address. Dear Nektar users, I'm using Nektar++ Incompressible Navier-Stokes Solver in fully 3D mode, with added temperature field, in a plane poiseuille flow heated at the bottom wall, solver running very well, and as if it wasn't already enough, I'm asking for more here ! 'hopefully', --> My question is about the possibility to handle variable viscosity or temperature dependent viscosity, So is there any approach to account for temperature dependent viscosity nu(T) ? I'm asking for this, since the temperature gradient at the heated wall is known to alter the viscosity (Ref: Physics of transitional shear flow, A. V. Boiko et al. Sec. 6.5 page 112) which I have to account for when carrying flow transition calculations. --> In case it is not directly achievable through Nektar's functionalities, is it conceivable to direct my effort toward employing an ad-hoc approach modifying the momentum transfer to account for lower liquid viscosity at higher temperature? That is, most likely adding a forcing term which mimics the effect of varying viscosity by altering the momentum transfer accordingly, since the stress tensor for a newtonian fluid tau=mu.du/dy. Any comment on the sanity of this approach ? Thank you all in advance, any help or general guidelines on this model variation greatly appreciated. Regards, Saad