I'm working on a question here where I have solved these boundary conditions;

$$\vec{n} \cdot \vec{j}= -\lambda\vec{n} \cdot \nabla u + k(\rho - \rho_0)u\vec{n}\cdot\vec{g}=0$$

So my question is, how do I know which of these 2 terms is a Dirichlet/Neumann boundary condition?


$\rho$ - density of substance, $\rho_0$ - density of the medium, $\vec{g}$ - gravitational field, $u$ - concentration, $k$ - constant.

Question: Find the boundary conditions which must be fulfilled in order for no substance to leave or enter the volume V .

What I did to solve this was I calculated $\vec{n} \cdot \vec{j}=0$ (since no substance can leave the volume, this condition has to be true), where $\vec{j}=-\lambda\nabla u+ku\vec{g}(\rho-\rho_0)$.

This finally gave me the expression that I first posted here (at the top). Now my question is, with the information that is given, how do I know if I have a Neumann Boundary condition or a Dirichlet Boundary condition?

  • $\begingroup$ As it stands now, it's just a bunch of symbols with some addition, subtraction and equality operators. It doesn't even mean anything. So please do provide some context. $\endgroup$
    – Ruslan
    May 31, 2017 at 16:33
  • $\begingroup$ Thank you, I have edited my post. If there is anything else that is unclear please feel free to ask, I really want to understand this kind of problem where I'm supposed to know what kind of boundary condition I'm working with, and why. $\endgroup$
    – armara
    May 31, 2017 at 16:59

1 Answer 1


If we rewrite your equation

$$-\lambda\vec{n} \cdot \nabla u + k(\rho - \rho_0)u\vec{n}\cdot\vec{g}=0$$


$$-\lambda\frac{\partial u}{\partial\vec n} + \left(k(\rho - \rho_0)\vec{n}\cdot\vec{g}\right)u=0,$$

where $\frac{\partial u}{\partial\vec n}$ is the normal derivative, then it's immediately apparent that the boundary condition expressed here is the third type boundary condition, also known as Robin boundary condition, which is a combination of Dirichlet boundary condition (first type) and Neumann one (second type).

  • $\begingroup$ Hmm, can you really write $n \nabla u = \frac{\partial u}{\partial \vec{n}}$? I thought the gradient was a derivation with respect to the spatial coordinates (x,y,z). $\endgroup$
    – armara
    May 31, 2017 at 19:49
  • $\begingroup$ Also, how do you recognize it as a third type boundary condition? I thought that the part that has a derivative was the Neumann-condition, and the part without a derivative was a Dirichlet-condition. $\endgroup$
    – armara
    May 31, 2017 at 19:52
  • $\begingroup$ @armara for the first question please follow the link about normal derivative, for the second one note that there's both a derivative and a value of $u$ in the equation, so it's neither first-, nor second-type boundary condition — but a weighted combination thereof. $\endgroup$
    – Ruslan
    Jun 1, 2017 at 4:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.