[gaurav_bhutani]

Is it mathematically incorrect to have a 3rd order boundary condition for a 2nd order Partial Differential Equation?

[agd]

I assume you are referring to a condition on the third derivative of the function to be solved for - in that case the answer is yes - it is incorrect. You would specify conditions on the function or the derivative of the function (or a mixture). The differential equation itself only defines the nth derivative of the function, so n+1 order derivative information on the boundaries is really meaningless.

If you are referring to the order of accuracy of a discretization scheme, then the statement makes sense mathematically, since you could have a third-order accurate approximation to a first-derivative. But it may or may not adversely interact with the 2nd-order approximation of the interior scheme.

[gaurav_bhutani]

Thanks a lot for the explanation.