1
$\begingroup$

I want to study the kth vorticity equation. The NS equation is provided as

\begin{align}\label{eq1} &\dfrac{\partial }{\partial t} \textbf{u} + \left(\textbf{u}\cdot \nabla \right)\textbf{u} = \nu \bigtriangleup \textbf{u} -\dfrac{1}{\rho} \textbf{Grad}(p) \end{align}

which implies

\begin{align}\label{eq2} &\dfrac{\partial }{\partial t} \textbf{curl}^k(\textbf{u}) + \textbf{curl}^k\biggl(\left(\textbf{u}\cdot \nabla \right)\textbf{u}\biggr) = \nu~ \bigtriangleup \textbf{curl}^k( \textbf{u}) \tag{1} \end{align}

with $$\textbf{curl}^k(\textbf{u}(x,t)) = \textbf{curl}(\textbf{curl}\cdots \cdots \textbf{curl})(\textbf{u}(x,t))$$ where we apply the curl $k$ times in succession. I want to compute $\textbf{curl}^k\biggl(\left(\textbf{u}\cdot \nabla \right)\textbf{u}\biggr)$ for each $k$. We know for $k=1$ we have $$\dfrac{\partial }{\partial t} \textbf{v} + \biggl((\textbf{u}\cdot \nabla )\textbf{v} -(\textbf{v}\cdot \nabla )\textbf{u} \biggr) = \nu~\bigtriangleup \textbf{v} , ~~~~~ \textbf{v}= \textbf{curl}(\textbf{u}).$$ Is it true in general $$\dfrac{\partial }{\partial t} \textbf{v}^k + \biggl((\textbf{u}\cdot \nabla )\textbf{v}^k-(\textbf{v}^k\cdot \nabla )\textbf{u} \biggr) = \nu~\bigtriangleup \textbf{v}^k, ~~~ \textbf{v}^k:= \textbf{curl}^k(\textbf{u}(x,t))$$

$\endgroup$
4
  • 2
    $\begingroup$ your $k=1$ equation has the wrong sign for the two terms between $\bigl( \cdots \bigr)$; see for example math.stackexchange.com/a/4595461/87355 ; the generalization to larger $k$ will include more terms than just these two $\endgroup$ Commented Dec 29, 2023 at 18:37
  • $\begingroup$ @CarloBeenakker thank you sorry about that. I thought induction would work but maybe not if the terms don't cancel. $\endgroup$
    – MrPie
    Commented Dec 29, 2023 at 18:47
  • $\begingroup$ @CarloBeenakker if no terms cancel we should get 2 more terms from each term involved. So we have $2^k$ terms for each $k$. Or some of them cancel $\endgroup$
    – MrPie
    Commented Dec 29, 2023 at 18:52
  • $\begingroup$ what is general formula for $$\textbf{curl}\left(\left(\textbf{a}\cdot \nabla \right)\textbf{b} - \left(\textbf{b}\cdot \nabla \right)\textbf{a}\right) $$ $\endgroup$
    – MrPie
    Commented Dec 29, 2023 at 18:56

1 Answer 1

2
$\begingroup$

Use that the velocity field is incompressible, $\nabla\cdot u=0$, to rewrite $$(u\cdot\nabla)u_j=\sum_{i} \nabla_i (u_iu_j).$$ You seek the curl of the curl of this expression, use that $$[\operatorname{curl}\operatorname{curl}f]_j=\sum_{i}\bigl(\nabla_j\nabla_i f_i-\nabla_i\nabla_if_j\bigr).$$ Hence $$[\operatorname{curl}\operatorname{curl} (u\cdot\nabla u)]_j=\sum_{i}\sum_k\nabla_i\nabla_k\bigl(\nabla_j(u_ku_i)-\nabla_i (u_ku_j)\bigr).$$


To check that the conjectured formula in the OP does not hold, consider the divergence-free velocity field $u=(x+y,x-y,xy)$. On the one hand, $\operatorname{curl}\operatorname{curl} u=0$, so the formula in the OP gives 0, but $\operatorname{curl}\operatorname{curl} (u\cdot\nabla)u=(0,0,-4)$ is nonzero.

$\endgroup$
9
  • $\begingroup$ Im not so used to that notation with subscript under $\nabla$. I aim to solve the kth curl of non-linear term. The formula you propose is the second curl or the kth one? $\endgroup$
    – MrPie
    Commented Dec 30, 2023 at 2:59
  • $\begingroup$ $\nabla_i \equiv \partial/\partial x_i$; so, for example, $\sum_i \nabla_i u_i = \operatorname{div} u$; this is the formula for the curl of the curl (second curl). $\endgroup$ Commented Dec 30, 2023 at 7:49
  • $\begingroup$ So this is only the second curl? My observation here is if we push $k$ to be large then maybe the kth curl of non-linear term will slowly vanish. So that at large values we can approximate the kth curl by solution to heat equation. $\endgroup$
    – MrPie
    Commented Dec 31, 2023 at 18:42
  • $\begingroup$ I don't think so; try $u=\{e^{x+y},-e^{x+y},e^x+e^y\}$, the $2k$-th curl of the non-linear term is $\{0,0,(-5)^k e^{x+y}(e^x-e^y)\}$, so it does not vanish for large $k$. $\endgroup$ Commented Dec 31, 2023 at 21:11
  • $\begingroup$ hmmmm if it doesnt vanish what can we say about it? $\endgroup$
    – MrPie
    Commented Dec 31, 2023 at 22:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .