If we are trying to simulate a flow which,despite having boundary conditions that are constant in time, still is expected to exhibit unsteady behaviour. In choosing an initial condition for the unsteady calculation, would it make sense to use the steady solution(presuming one can be found)?

Thanks

The flow would eventually become unsteady due to the numerical errors but depending on the problem being studied the time taken for reaching unsteadiness might be long if you use a steady solution. Try using a perturbed solution of the form

U_i=U_s+C*U_s*sin(w*x)......

where C you can use between 0.05-0.2 and w is the frequency which you can compute based on the domain size or you can impose arbitrary frequencies.

Hope this helps.

Thank you very much. I grasped the general meaning, however I GOT LOST in the Equation (the perturbed solution). Is there any book or website you may recommend? Thank you.

Kind Regards

There was a discussion on this topic in the context of channel flow in the current forum. search in the forum and you would find it.

Thanks Harish. ^_^