notes

uni notes
git clone git://popovic.xyz/notes.git
Log | Files | Refs

commit b87c4ff6232fd1ed6b64dd3f916739894690e7ff
parent 2c07f08b13fa5a7e325fe0fbb5b2da9c95d3a874
Author: miksa <milutin@popovic.xyz>
Date:   Thu,  2 Jun 2022 18:47:22 +0200

boundary conditions and checking

Diffstat:
Aapp_pde/appendix.tex | 60++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Mapp_pde/basics_fluids.tex | 176++++++++++++++++++++++++++++++++++++++++++++++++++++++-------------------------
Mapp_pde/build/basics_fluids.pdf | 0
3 files changed, 180 insertions(+), 56 deletions(-)

diff --git a/app_pde/appendix.tex b/app_pde/appendix.tex @@ -0,0 +1,60 @@ +\appendix +\section{Appendix: Mathematical Preliminaries} +\subsection{Leibniz Rule of Integration} +\label{appendix:leibniz} +The Leibniz integral rule for differentiation under the integral sign +initiates with an integral +\begin{align} + \mathcal{I}(t, x) = \int_{a(t)}^{b(t)} f(t, x) dx = \mathcal{I}(t, a(t, + a(t), b(t))). +\end{align} +And upon differentiation w.r.t. $t$, utilizes the chain rule on $a(t)$ and +$b(t)$ respectively, by +\begin{align} + \frac{d\mathcal{I}}{dt} = + \frac{\partial \mathcal{I}}{\partial t}+ + \frac{\partial \mathcal{I}}{\partial a}\frac{\partial a}{\partial t}+ + \frac{\partial \mathcal{I}}{\partial b}\frac{\partial b}{\partial t}. +\end{align} +Which in integral representation reads +\begin{align} + \frac{d\mathcal{I}}{dt} = \int_{a(t)}^{b(t)}\frac{\partial f(t, + x)}{\partial t} dx + f(t, b(t)) \frac{\partial b(t)}{\partial t} + - f(t, a(t)) \frac{\partial a(t)}{\partial t} +\end{align} + +\subsection{Gaussian Integration Law} +\label{appendix:gauss integration} +This should explain the Gaussian integration law + +\subsection{Identity for Vorticity} +We start off with the standard material derivative +\begin{align} + \frac{D\mathbf{u}}{Dt} = \frac{\partial \mathbf{u}}{\partial t} + (\mathbf{u}\nabla)\mathbf{u}. +\end{align} +We will use Einstein's Summation Convention, where we sum over indices that +both appear at as the bottom as the top index, to rewrite the second part of +the material derivative $(\mathbf{u}\nabla)\mathbf{u}$ into +\begin{align} + (\mathbf{u}\times (\nabla \times \mathbf{u}))_k + &= \varepsilon^{ijk}u_j(\nabla \times \mathbf{u})_k \\ + &= \varepsilon^{ijk}u_j\varepsilon_{klm}\partial^l u^m\\ + &=(\delta^i_l\delta^j_m-\delta^i_m\delta^j_l)u_j\partial^l u^m\\ + &=u_m\partial^i u^m - u_l \partial^l u^i.\label{eq:identity split} +\end{align} +Now the first part in equation \ref{eq:identity split} can be rewritten into +\begin{align} + u_m\partial^i u^m =\partial^i (\frac{1}{2}u_mu^m) . +\end{align} +Thus we get +\begin{align} + (\mathbf{u}\times (\nabla \times \mathbf{u}))_k + = \frac{1}{2}\partial^i(u_m u^m) + u_l \partial^l u^i, +\end{align} +which is +\begin{align} + (\mathbf{u}\nabla)\mathbf{u} = \nabla(\frac{1}{2}\mathbf{u}\mathbf{u}) - + \left(\mathbf{u}\times (\nabla \times \mathbf{u})\right) +\end{align} + diff --git a/app_pde/basics_fluids.tex b/app_pde/basics_fluids.tex @@ -6,23 +6,24 @@ \tableofcontents \section{Governing Equations of Fluid Dynamics} -We first of start with a fluid with density/mass +We first start of with a fluid with a density \begin{align} \rho(\mathbf{x}, t), \end{align} -with a position $\mathbf{x} = (x, y, z)$ in three dimensional space at time +in three dimensional Cartesian coordinates $\mathbf{x} = (x, y, z)$ at time $t$. For water-wave applications, we should note that we take -$\rho=\text{constant}$. The fluid moves in time and space with a velocity field +$\rho=\text{constant}$, but we will go into this fact later. The fluid moves +in time and space with a velocity field \begin{align} \mathbf{u}(\mathbf{x}, t) = (u, v, w). \end{align} -Additionally it is also described by the pressure of the fluid +Additionally it is also described by its pressure \begin{align} P(\mathbf{x}, t), \end{align} generally depending on time and position. When thinking of e.g. water the pressure increases the deeper we go, that is with decreasing or increasing $z$ -coordinate (depending how we set up our system $z$ pointing up or down +direction (depending how we set up our system $z$ pointing up or down respectively). The general assumption in fluid dynamics is the \textbf{Continuum @@ -30,16 +31,16 @@ Hypothesis}, which assumes continuity of $\textbf{u}, \rho$ and $P$ in $\mathbf{x}$ and $t$. In other words, we premise that the velocity field, density and pressure are ''nice enough`` functions of position and time, such that we can do all the differential operations we desire in the framework of -fluid dynamics. +differential analysis. \subsection{Mass Conservation} Our aim is to derive a model of the fluid and its dynamics, with respect to -time and position, in the most general way. This is generally done thinking -of the density of a given fluid, which is a unit of mass per unit volume, -intrinsically an integral representation to derive these equations is going -to be used. +time and position, in the most general way. This is usually done thinking +of the density of a given fluid, which is a unit mass per unit volume, +intrinsically an integral representation to derive these equations suggests +by itself. Let us now thing of an arbitrary fluid. Within this fluid we define a fixed -volume $V$ relative to a chosen inertial frame and bounded by a surface $S$ +volume $V$ relative to a chosen inertial frame and bound it by a surface $S$ within the fluid, such that the fluid motion $\mathbf{u}(\mathbf{x}, t)$ may cross the surface $S$. The fluid density is given by $\rho(\mathbf{x}, t)$, thereby the mass of the fluid in the defined Volume $V$ is an integral @@ -79,22 +80,23 @@ The figure bellow \ref{fig:volume}, expresses the above described picture. with a surface normal vector $\mathbf{n}$ \label{fig:volume}} \end{figure} -Since we want to figure out the fluid dynamics, we can consider now the rate -of change of mass in of this completely arbitrary $V$, which needs to be -disappear, i.e. is equal to zero since we cannot lose mass. Matter (mass) is -neither created nor destroyed anywhere in the fluid +Since we want to figure out the fluid's dynamics, we can consider the rate +of change in the completely arbitrary $V$. The rate of change of mass needs to +disappear, i.e. it is equal to zero since we cannot lose mass. Matter (mass) is +neither created nor destroyed anywhere in the fluid, leading us to \begin{align} \frac{d}{dt}\left( \int_V \rho(\mathbf{x}, t)\ dV \right) = 0. \end{align} -We may get more information with simply ''differentiating under the integral +To get more information we simply ''differentiate under the integral sign``, also known as the Leibniz Rule of Integration, see appendix -\ref{appendix:leibniz}, the above integral equation reads +\ref{appendix:leibniz}, the integral equation representing the rate of change +of mass reads \begin{align}\label{eq:mass balance} \frac{dm}{dt} = \int_V \frac{\partial \rho(\mathbf{x}, t)}{\partial t}\ dV +\int_{\partial V} \rho(\mathbf{x}, t) \mathbf{u}\cdot\mathbf{n}\ dS = 0. \end{align} -The above equation \ref{eq:mass balance} is an underlying equation, describing that the rate of +The above equation in \ref{eq:mass balance} is an underlying equation, describing that the rate of change of mass in V is brought about, only by the rate of mass flowing into V across S, and thus the mass does not change. @@ -115,8 +117,8 @@ called the \textbf{Continuity Equation} \partial_t \rho + \nabla(\rho \mathbf{u}) = 0 \end{align} -In the light of the results of the equation of mass conservation -\ref{eq:continuity}, an expansion, of the nabla gives +In light of the results of the equation of mass conservation +in \ref{eq:continuity}, an product rule gives \begin{align} \partial_t \rho + (\nabla \rho)\mathbf{u} + \rho(\nabla \mathbf{u}), \end{align} @@ -125,7 +127,7 @@ as follows \begin{align} \frac{D}{Dt} = \partial_t + \mathbf{u}\nabla. \end{align} -Thus the equation of mass conservation becomes +With the material derivative the equation of mass conservation reads \begin{align} \frac{D\rho}{Dt} + \rho \nabla\mathbf{u} = 0 \end{align} @@ -153,7 +155,7 @@ element in the fluid. The integral formulation of the force would be \int_V \rho \mathbf{F}\ dV - \int_S P\mathbf{n}\ dV. \end{align} Now applying the Gaussian rule of integration on the second integral over the -surface, the resluting force in per unit volume is +surface, the resulting force in per unit volume is \begin{align} \int_V \left(\rho \mathbf{F} - \nabla P\right)\ dV. \end{align} @@ -161,54 +163,116 @@ The acceleration of the fluid particles is given by $\frac{D\mathbf{u}}{Dt}$, and thus the total force per unit volume on the other hand is \begin{align} \int_V \rho \frac{D\mathbf{u}}{Dt}\ dV = - \int_V \left(\rho \mathbf{F} - \nabla P\right)\ dV + \int_V \left(\rho \mathbf{F} - \nabla P\right)\ dV. \end{align} - - - - - - - - -\newpage -\appendix -\section{Appendix: Mathematical Preliminaries} -\subsection{Leibniz Rule of Integration} -\label{appendix:leibniz} -The Leibniz integral rule for differentiation under the integral sign -initiates with an integral +Newton's Second Law for a fluid in an Volume is essentially saying that the +rate of change of momentum of the fluid in the fixed volume $V$, which is the particle +acceleration is the resulting force acting on V together with the rate of +flow of momentum across the surface $S$ into the volume $V$. Hence we arrive +at the \textbf{Euler's Equation(s) of Motion} \begin{align} - \mathcal{I}(t, x) = \int_{a(t)}^{b(t)} f(t, x) dx = \mathcal{I}(t, a(t, - a(t), b(t))). + \frac{D\mathbf{u}}{Dt} = \left(\frac{\partial \mathbf{u}}{\partial t} + (\mathbf{u}\nabla)\mathbf{u}\right) = + -\frac{1}{\rho}\nabla P + \mathbf{F}. \end{align} -And upon differentiation w.r.t. $t$, utilizes the chain rule on $a(t)$ and -$b(t)$ respectively, by +As a side note we have mentioned that there is another contribution if the +fluid is viscid. Indeed there is a tangential force due to the velocity +gradient, which into introduces the additional term \begin{align} - \frac{d\mathcal{I}}{dt} = - \frac{\partial \mathcal{I}}{\partial t}+ - \frac{\partial \mathcal{I}}{\partial a}\frac{\partial a}{\partial t}+ - \frac{\partial \mathcal{I}}{\partial b}\frac{\partial b}{\partial t}. + \mu \nabla^2 \mathbf{u}, \qquad + \mu = \text{viscosity of the Fluid}. \end{align} -Which in integral representation reads +Thereby the equations become \begin{align} - \frac{d\mathcal{I}}{dt} = \int_{a(t)}^{b(t)}\frac{\partial f(t, - x)}{\partial t} dx + f(t, b(t)) \frac{\partial b(t)}{\partial t} - - f(t, a(t)) \frac{\partial a(t)}{\partial t} + \rho\frac{D\mathbf{u}}{Dt} + = -\nabla P + \rho \mathbf{F} + \mu \nabla^2 \mathbf{u}. \end{align} -\subsection{Gaussian Integration Law} -\label{appendix:gauss integration} -This should explain the Gaussian integration law - - - +For now we have separated two simplifications, that define an +\textbf{idealized/perfect fluid} +\begin{enumerate} + \item \textbf{incompressible} $\qquad \mu=0$ + \item \textbf{inviscid} $\quad \rho = \text{const.},\ \nabla \mathbf{u}= + 0$ +\end{enumerate} +\subsection{Vorticity and irrotational Flow} +The curl of the velocity field $\mathbf{\omega} = \nabla \times \mathbf{u}$ +of a fluid (i.e. the vorticity), describes a spinning motion of the fluid +near a position $\mathbf{x}$ at time $t$. The vorticity is an important +property of a fluid, flows or regions of flows where $\mathbf{\omega}=0$ are +\textbf{irrotational}, and thus can be modeled and analyzed following well +known routine methods. Even though real flows are rarely irrotational +anywhere (!), in water wave theory wave problems, from the classical aspect +of vorticity have a minor contribution. Hence we can assume irrotational flow +modeling water waves. To arrive at the vorticity in the equations of motions +derived in the last section we resort to a differential identity derived in appendix +\ref{appendix:diff identity}, which gives for the material derivative +\begin{align} + \frac{D\mathbf{u}}{Dt} = \frac{\partial \mathbf{u}}{\partial t} + \nabla(\frac{1}{2}\mathbf{u}\mathbf{u)} + - \left( \mathbf{u}\times (\nabla \times \mathbf{u} \right). +\end{align} +Thus the equations of motion become +\begin{align} + \frac{\partial \mathbf{u}}{\partial t} + \nabla\left( + \frac{1}{2}\mathbf{u}\mathbf{u} + \frac{P}{\rho} + \Omega \right) + = \mathbf{u} \times \mathbf{\omega}, +\end{align} +where $\Omega$ is the force potential per +unite mass given by $\mathbf{F} = -\nabla \Omega$. +At this point we may differentiate between \textbf{stead and unsteady flow}. +For \textbf{Steady Flow} we assume that $\mathbf{u}, P$ and $\Omega$ are time +independent, thus we get +\begin{align} + \nabla\left( \frac{1}{2}\mathbf{u}\mathbf{u} + \frac{P}{\rho} + \Omega + \right) = \mathbf{u} \times \mathbf{\omega}. +\end{align} +It is general knowledge that the gradient of a function $\nabla f$ is +perpendicular the level sets of $f(\mathbf{x})$, where $f(\mathbf{x}) = +\text{const.}$. Thus $\mathbf{u} \times \mathbf{\omega}$ is orthogonal to +the surfaces where +\begin{align} \label{eq:bernoulli} + \frac{1}{2}\mathbf{u}\mathbf{u} + \frac{P}{\rho} + \Omega = + \text{const.}, +\end{align} +The above equation is called \textbf{Bernoulli's Equation}. +Secondly \textbf{Unsteady Flow} but irrotational (+ incompressible), first of +all gives us the condition for the existence of a velocity potential $\phi$ +in the sense +\begin{align} + \mathbf{\omega} = \nabla \times \mathbf{u} = 0 \quad \Rightarrow \quad + \mathbf{u} = \nabla \phi, +\end{align} +where $\phi$ needs to satisfy the Laplace equation +\begin{align} + \Delta \phi = 0. +\end{align} +According to the Theorem of Schwartz we may exchange $\frac{\partial +}{\partial t}$ and $\nabla$, giving us an expression for the material +derivative +\begin{align} + \nabla\left( \frac{\partial \phi}{\partial t} +\frac{1}{2} + \mathbf{u}\mathbf{u} + \frac{P}{\rho} + \Omega \right) = 0 +\end{align} +Thus the expression differentiated by the $\nabla$ operator is an arbitrary +function $f(\mathbf{x}, t)$, writing +\begin{align} + \frac{\partial \phi}{\partial t} +\frac{1}{2} + \mathbf{u}\mathbf{u} + \frac{P}{\rho} + \Omega = f(\mathbf{x}, t). +\end{align} +The function $f(\mathbf{x}, t)$ can be removed by gauge transformation of +$\phi \rightarrow \phi + \int f(\mathbf{x}, t)\ dt$, never the less this is +not further discussed and left to the reader in the reference. +\subsection{Boundary Conditions for water waves} +Surface, Bottom, Pressure +\newpage +\include{appendix.tex} diff --git a/app_pde/build/basics_fluids.pdf b/app_pde/build/basics_fluids.pdf Binary files differ.