Student Notes
My notes
Last updated
My notes
Last updated
Attached are the handwritten notes, collected during lectures by Dr. Rusak. They are mostly an exercise in wrote just as brainless is and attacked for completeness.
Fluid mechanics is a topic in classical physics of fluid is a substance gas or liquid where intermolecular forces are medium to week and there's a significant random movement of molecules and disorder in the arrangement
Continuum hypothesis a fluid is modeled as a Continuum or material that is exhibiting no structure however small is divided
$\lambda$ is the mean free path
the mean free path is a distance between molecular collisions 10-7 M for air at STP greater than a meter at the edge of the atmosphere 10 -10 m in water liquid
Knudsen number
$\text{Kn} = \frac{\lambda}{L} << 1$
but L characteristic length of Interest much greater than Lambda Kn = lambda/L <<< 1 Collision between thermal equilibrium tau is less than 10 to the -8 second and air at STP free continuing Behavior how is much less than the time of Interest which we measure
Is important to recognize that for very small distances approximately within an order of magnitude of the mean free path of the situation at hand
density as a function control volume $\rho = \lim_{\Delta V\rightarrow min} \frac{\Delta m}{\Delta Vol}$
specific volume $ v = 1/\rho$
To find the pressure of a hydrostatic fluid look at the unit tetrahedra each of the minor faces has a force orthogonal to eachother. The normal force balancing has to be the balaning them allso that the pressure is the force pre unit area of some arbitrary are
Inreality there are viscous effects but are very small for most slightly viscous fluids so the viscous component can usually be neglected
Definition
Average kinetic energy
k = 1.38\timex 10^{-26}J/K
Specific properties
internal energy :$du = c_vdT - [T(\frac{\partial p}{\partial T}-p)]\frac{d\rho}{\rho^2}$
enthalpy: $h = u + \frac{P}{\rho}$ or $dh=du+d(p/\rho)$
entropy: $Tds = cVdT - (\frac{\partial P}{\partial T}){cost. \rho}\frac{d\rho}{\rho^2}$ leads to $Tds = dh - \frac{d\rho}{\rho}$ or the gibbs eqn
simple compressible fluid: any two properties define the rest $P = f(\rho,T)$ but that is not the only choice. Just need to link the fluid to the critical point
Critical point $(P_c ,\rho_c, T_c)$ is the location where first and second partials of pressure with respect to density for constant temperature is 0
Liquid: high first derivative so we can usually hold \rho constant rate of temperature is not necessarily high tho
Gas vapor is moderate so that small changes in density create comporable changes in presssure
Carried out on the variable $(\cdot)$ by dividing by the corresponding $(\cdot)_c$ so that $(\cdot)_R = (\cdot)/(\cdot)_c$. for $P,\rho = 1/\mathcal{V},T$ and the compressibility factor $Z = \frac { P } { \rho RT}$ where $R = \mathcal{R}/M$ the universal gas constant div the molar weight Z is a function of $P_R,T_R$
For low pressure and density, we have the conditions for a thermodynamically perfect ideal gas $P = \rho RT$
Vander Waals Eqn of State
def $\left( P + a \frac { 1 } { V { m } ^ { 2 } } \right) \left( V { m } - b \right) = R T$ or as he preferes $P = \rho RT/(1-b\rho)-a\rho^2$
Redlich-Kwong Eqn. of State
def: $p = \frac { R T } { V { m } - b } - \frac { a } { \sqrt { T } V { m } \left( V _ { m } + b \right) }$
Scalars are properties without a preferred direction $(\rho,P,T,\mu,h,s)$ so that the normal operations are defined on
Vectors are a properties with a prefered direction so that ${\bf V} = u{\bf e}_x + v {\bf e}_y + w{\bf e}_z = V_x{\bf e}_x + V_y{\bf e}_y + V_z{\bf e}_z$
Here with tensors or order 1 or greater, operations are not commuitive, directions matter
component extraction ${\bf c} = k{\bf a} = ka_i{\bf e}_i\rightarrow c_i = ka_i $
addition subtraction
${\bf c} = {\bf a} \pm {\bf n} = (a_i + b_i){\bf e}_i$
multiplication is a measure of orthogonality of two vectors
dot inner product : c is now a scalar $c = {\bf a\cdot b} =a_i \space b_i$
cross outer product: c is of the same rank $\bf c = a\times b$ or $ci = \epsilon{ijk}a_jb_k$
Rusak’s notation is kinda fucked up. I’m going to copy one verbatim and describe my qualms
$\stackrel{T}{\underline{\underline{}}} = \stackrel{\rightarrow}{a}\stackrel{\rightarrow}{b}$ and does not define the operation which takes place between. Obviously it’s an outer dyatic product but notationally it does not make sense unless indexes are applied
$\textcolor{red}{fucked up }$
what happens to a single element
all the points which have passed through a point from some $t_0$
hold time constant and show tangent line of the velocity field or path element would take if time was constant so that $d{\bf l} \times {\bf V} = 0 $.
Construct field so that
Follow the element in time so that every variable has a definite derivative wrt time.
Let the quantity of interest be $\bf V$.
${\bf a}_e = \frac{D{\bf V}_e}{Dt} = \frac{d{\bf V}_e}{dt} = \left(\frac{\partial{\bf V}_e}{\partial t} \right)_e + {\bf V}_e(t) \cdot \left(\nabla {\bf V}\right)_e $
Decompose the velocity gradient into even and odd components
$\bf D$ deformation tensor
$\bf \Omega$ spin tensor
Then the velocity vector is described in a Taylor series approximation for fixed time $\textcolor{red}{\texttt{I think he fucked up in the notes here}}$ we have comments of the velocity field in all directions not just ${\bf e}_x$
Note that $d{\bf l}= dx{\bf e}_x + dy{\bf e}_y + dz{\bf e}_z$. In the limit as the differential segment length approaches zero
When Fi is an extensive property (volume dependent), the we normalize not by volume but by mass to get a specific quantity : total to specific
As the fluid element evolves in time, we get reynolds transport theorem
Reynolds Transport theory
let $F{II} = \int\text{vol} \rho f d\text{vol}$ where $f$ is the normalized $F$ per unit mass
$f = 1 $, then $F_{II} \rightarrow \text{mass in volume at }t $
$f = {\bf V} $, then $F_{II} \rightarrow \text{momentum in volume at }t $
$f = e = u + \frac{1}{2}{\bf V}\cdot {\bf V}$, then $F_{II} \rightarrow \text{total energy in volume at }t $
Inlet
$\dot{\mathcal{F}}_\text{inlet} = -\int\rho f({\bf V \cdot n})dS$
with outlet
$\dot{\mathcal{F}}_\text{outlet} = \int\rho f({\bf V \cdot n})dS$
then diff
$\dot{\mathcal{F}}\text{inlet} - \dot{\mathcal{F}}\text{outlet} = -\int\rho f({\bf V \cdot n})dS$
end with
is the integral equation of balance on $F$ where $\dot{\mathcal{Q}}F = \int\text{Vol} \dot{q}_F d\text{vol}$
Let $f = e_t$ specific total energy , $e_t = u + \frac{1}{2} {\bf V \cdot V}$
Integral form: $\frac{d}{dt}\int_V \rho e_t d V+ \int_S \rho e_t({\bf V \cdot n}) = \dot{Q}_F$ power or energy per unit time
Conservative $\frac{d}{dt}(\rho e_t)+ \rho e_t({\bf V \cdot n}) = \dot{q}_F$
Regular $\rho\left[\frac{d}{dt}( e_t)+ e_t({\bf V \cdot n})\right] = \dot{q}_F$
a
Extensive property $F = \int_V \rho f dV$ so it’s normalized by mass but volume dependent
Integral equation
Conservation form
Note that $\dot{q}$ is $\dot{\mathcal{Q}}$ per unit volume
Regular Form through the domain
Lagrangian following a fluid element
$\rho_e(t)\left(\frac{df}{dt}\right)_e = \dot{q}_F$
The linear momentum equation where $f \rightarrow {\bf V}$ and $\dot{q}_F \neq0$
$\frac{\partial (\rho {\bf V})}{\partial t} + \nabla \cdot (\rho({\bf V\otimes V})) ={\bf \dot{q}}_F$ or $\rho(\frac{\partial ( {\bf V})}{\partial t} + \nabla \cdot ({\bf V\otimes V})) ={\bf \dot{q}}_F$
in Newton’s 2nd law form $\rho_e(t)(\frac{d{\bf V}}{dt})_e = ({\bf \dot{q}}_F)_e$
Body Force : $\int_V\rho {\bf B}dV$
Surface Force (traction) :
Then ${\bf \Theta }= -p{\bf n}+{\bf \tau}$
More Regularly
Use Cauchy to show assuming symmetric stress tensor $\bf \tau = T \cdot n$.
Use divergence theorem $\int_S {\bf A\cdot n}dS=\int_V \nabla\cdot {\bf A}dV$
Use Green’s theorem $\int_S a{\bf n}dS = \int_V\nabla a\space dV$
we produce
What is the stress tensor? Well Euler is inviscid so $\bf T = 0$.
Regular Form
2 parameter viscosity and bulk viscosity for a newtonian fluid
Nonlinear is ${\bf T} = f\left(\nabla \cdot {\bf V}\right)$
$\rho\left(\frac{\partial {\bf V}}{\partial t} + {\bf V}\cdot(\nabla {\bf V})\right)=\rho{\bf B} - \nabla p + \nabla \cdot {\bf T}$
$\rho \left( \frac{\partial e_t}{\partial t} + {\bf V} \cdot \nabla e_t\right) = \rho {\bf B\cdot V} - \nabla \cdot (p {\bf V}) + \nabla \cdot ({\bf T\cdot V})-\nabla \cdot {\bf q}$
$p = f(\rho , T) $ or hold $\rho$ as constant
$\frac{1}{\rho_e}\left(\frac{d\rho}{dt}\right)_e = - \nabla \cdot {\bf V}$
$\rho_e(t) \left(\frac{d{\bf V}}{dt}\right)_e = \rho_e{\bf B}_e - (\nabla p)_e + (\nabla \cdot {\bf T})_e$
$\rho_e(t) \left(\frac{d{e_t}}{dt}\right)_e = \rho_e{(\bf B\cdot V)}_e - (\nabla \cdot (p{\bf V}))_e + (\nabla \cdot {\bf (T\cdot V)})_e -(\nabla \cdot {q})_e$
Internal boundary condition where $\bf V\cdot n$ is some value.
Solid
Porous
Using a characteristic set of reference properties, we can create a solution which describes a family of solutions, of geometrically similar problems: care only about the shape of the airfoil
Choose a selection of characteristic properties
Consider
Consider