Scribd
Upload
19 views

Lecture 10 12

Uploaded by

123 amaxel
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
19 views

Lecture 10 12

Uploaded by

123 amaxel
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 33

2024학년도 2학기 화공수학

Fourier Analysis: Fourier


Transforms
화학공학과
담당교수: 이용진
[email protected]
Discrete Fourier Transform (DFT)
• Fourier Transform ~ assume that a function f(x) is given on some
interval.
• Often, a function f(x) is given only in terms of values at finitely
many points.
• The main application of such a “discrete Fourier analysis”
concerns large amounts of equally spaced data.
• Dealing with sampled values rather than with functions, we can
replace the Fourier transform by the so-called discrete Fourier
transform (DFT).
• Let f(x) be periodic, (2π), we assume that N measurements of f(x)
are taken over the interval 0 ≤ 𝑥 ≤ 2𝜋 at regularly space points
Discrete Fourier Transform (DFT)
• We also say that f(x) is being sampled at these points.
• We now want to determine a complex trigonometric polynomial
Discrete Fourier Transform (DFT)
Discrete Fourier Transform (DFT)
Discrete Fourier Transform (DFT)
Discrete Fourier Transform (DFT)-Example
2024학년도 2학기 화공수학

Partial Differential
Equations: Basic Concepts
of PDEs
화학공학과
담당교수: 이용진
[email protected]
Partial Differential Equations
• Function of interest depends on two or more independent
variables → typically time and one or more spatial variables.
• Only the simplest physical systems can be modeled by ODEs
• Most problems in science and engineering - including heat and mass transfer,
fluid mechanics, quantum mechanics etc. lead to PDEs
Defining PDEs
• Just like with ODEs, characterization of your PDE will indicate your solution
method
• Order is defined by the order of the highest derivative
• A PDE is linear if it is of the first degree in the dependent variable and its
partial derivatives.
e.g.
Partial Differential Equations
• A linear PDE is homogeneous if every term contains the
dependent variable or one of its derivatives
Ex:

Notation
• Often, 𝑢 is the dependent variable; e.g. 𝑢=𝑢(𝑥,𝑦,z,𝑡),
𝜕2 𝑢
∂𝑢/∂𝑡 can be written as 𝑢𝑡 , 2 can be written 𝑢𝑥𝑥 ,
𝜕𝑥
∇𝑢=∂𝑢/∂𝑥+∂𝑢/∂𝑦+∂𝑢/∂𝑧

∴ if 𝑢𝑡 = −𝑐∇2 𝑢,
𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢
• ∂𝑢/∂𝑡=−𝑐[ + + ]
𝜕𝑥 2 𝜕𝑦 2 𝜕𝑧 2
Partial Differential Equations
Common PDE’s in Chemical Engineering
• There are many important linear PDEs of the second order in chemical
engineering, particularly if you go to grad school. Some of the key equations
are:
• Heat Transfer or diffusion

• Wave equations
- Arise in fields like acoustics, electromagnetics and fluid dynamics
- Describe any kind of waves – mechanical (e.g. water, sound, seismic) or light
waves
- A) 𝑢𝑡𝑡 = 𝑐 2 𝑢𝑥𝑥 → 1𝐷, B) 𝑢𝑡𝑡 = 𝑐 2 (𝑢𝑥𝑥 + 𝑢𝑦𝑦 ) → 2𝐷
• Laplace equation ( ∆~ the Laplace operator)
- Simplest example of an elliptic PDE (special type of linear second order PDE)
- Solutions to these equations are the harmonic function -> important in many
fields of science – e.g. astronomy + fluid dynamics -> describe the behavior of
fluid potentials; also represents the steady state heat equation (no dependence
on time)
Special Solution to PDE’s
• There are often many possible general solution to PDEs. For
example, the 2D Laplace equation ∇2 𝑢 = 0 is satisfied by very
unique functions:

• In-class exercise:
Is ln(𝑥 2 + 𝑦 2 ) a solution to this differential equation?
Boundary/Initial Condition in PDEs
• The particular solution for any problem is determined by boundary
conditions and, if time is a variable, initial condition.
- Conditions must match up with the partial derivatives
- E.g. 𝑢𝑡𝑡 = 𝑐 2 𝑢𝑥𝑥 -> requires 1 IC + 2BC’s -> If you tried 2 IC’s and 1BC,
these would not be able to define the physical reality of the problem.
• Just as with ODEs, the superposition / linearity principle applies: if we
have two linearly independent solutions of our PDE in some region R,
then the sum and multiples of those two solution is also a solution
- If 𝑢1 and 𝑢2 are both solution, then 𝑢3 = 𝑐1 𝑢1 + 𝑐2 𝑢2 is as well
Examples
• Example 1: 𝑢𝑥𝑥 − 𝑢 = 0
- Since no derivative with regard to y occur, we can solve like 𝑢′′ −
𝑢 = 0, keeping in mind that our arbitrary constants will need to
be functions of y:
- λ2 − 1 = 0 -> λ = ±1 and 𝑢 𝑥, 𝑦 = 𝐴 𝑦 𝑒 𝑥 + 𝐵(𝑦)𝑒 −𝑥
• Example 2: 𝑢𝑥𝑦 = −𝑢𝑥
- We can use reduction of order to simplify.
Let 𝑝 = 𝑢𝑥 . Then PDE becomes

and
Examples
• Solve for u= u(x,y)
- 25𝑢𝑦𝑦 − 4𝑢 = 0

- 𝑢𝑥𝑦 = 𝑢𝑥
Solution by Separation Variables
• The model of a vibrating elastic string (a violin string, for instance)
consists of the one-dimensional wave equation

• For the unknown deflection u(x,t) of the string, a PDE that we have
just obtained, and some additional conditions, which we shall now
derive.
• Since the string is fastened at the ends x=0 and x=L, we have the
two boundary conditions

• Furthermore, the form of the motion of the string will depend on


its initial deflection (deflection at time t =0), call it f(x), and on its
initial velocity (velocity at t=0), call it g(x). We thus have two initial
conditions
Solution by Separation Variables
• We now have to find a solution of the PDE satisfying the
conditions. This well be the solution of the problem. We shall do
this in three steps, as follows.
Solution by Separation Variables
• Method to solve some linear PDEs with linear boundary condtitions
(heat equation, wave equation, laplace equation, Helmholtz
equation, and others).
• Rests on the assumption that a multivariable function, like u(x,y,t)
can be represented as a product of several single variable functions
i.e. u(x,y,t) = X(x)Y(y)T(t)
• This will allow us to turn a PDE into two or more ODEs that we
already know how to solve
• This does not always work, just like it didn't always work for 1st
order ODE's
Worked Example : 1D Heat Transfer in a Bar
• Ex : 𝑇𝑡 𝑥 = 𝛼𝑇𝑥𝑥 (1D heat transfer)
• Here 𝛼 is the thermal diffusivity with units of 𝑚2 /𝑠.

• First, we assume 𝑇 𝑥, 𝑡 = 𝑋(𝑥) ∙ 𝑌(𝑡)


Worked Example : 1D Heat Transfer in a Bar
𝛼 𝑑2𝑋
• Call LHS = (1/Y)(dY/dT)=F(t) and RHS = =𝐺 𝑥 .
𝑋 𝑑𝑥 2

• F(t) can equal G(x) only if both are equal to some constant, 𝜇.
• One final note : 𝜇 has units of 1/time as written.
• Now, we can split our PDE into two ODEs.
Solution for Y(t)

• At this point we haven’t know anything about 𝜇. However, we can


say that t should be negative.
• If 𝜇 was positive, we would have a temperature that explodes to ∞,
which is unphysical.
Worked Example : 1D Heat Transfer in a Bar
• To make things easier to see, we replace 𝜆 = −𝜇.
• 𝜆 will be positive, but the exponential will tend to zero at long times.
𝑌 𝑡 = 𝑐 ∙ exp 𝜇𝑡 = 𝑐 ∙ exp −𝜆𝑡
Solution for X(x)

Linear 2nd order homogeneous


constant coefficients

Characteristic equation

Using Euler’s formula yields


Finally, the general solution
Application of Boundary Conditions
• A particular solution is identified by applying one IC and two BCs that
represent the physical reality of the problem.
• Let’s assume the ends of the bar are held at a fixed reference
temperature, which will call 0.
• This could be replaced by any constant 𝑇𝑟𝑒𝑓 , since this is a linear
equation.

• We need three conditions because this is second order in x (need


two conditions), and first order in t (need one initial condition).
1) BC #1: Apply T(0,t) = 0

This whole thing has to be 0. c=0 would only lead the trivial solution.
So, A has to be 0.
Application of Boundary Conditions
2) BC #2: Apply T(L,t) = 0

Again, we don’t want a trivial solution. So we want to pick 𝜆 such that


the sin term is zero.
𝜆
This is only the case if 𝐿 = 𝑛𝜋
𝛼

Solving for 𝜆:
This implies there’s actually a series of solutions.
The solution is then
This is super interesting. The timescale for heat dissipation is related to
the solution form.
Application of Boundary Conditions
• Let’s talk about the last piece of information we have T(x,0)=f(x)
• We haven’t actually specified what the initial conditions are, and
that’s how we are going to determine 𝑐𝑛 .
Initial condition
• Let’s say that the initial condition f(x) is a sinusoid (the initial
temperature profile is sinusoid).
• This matches our solution if n = 1, so
• Here,
• Let’s go ahead and plot this solution for 𝛼 = 0.1 𝑚2 /𝑠 and L = 1m.
Exercise 1
Find the temperature u(x,t) in a bar of silver of length 10cm and constant cross section of area
cal
1cm2 (density 10.6 g/cm3, thermal conductivity 1.04 , specific heat 0.056 cal/(g ℃)
cm sec ℃
that is perfectly insulated laterally, with ends kept at temperature 0 ℃ and initial temperature
f x = 4 − 0.8 𝑥 − 5 ℃.
Exercise 2
• A bar with heat generation of constant rate H(>0) is modeled by
ut = c 2 uxx + H. Solve this problem if L = π and the ends of the
bar are kept at 0℃.

You might also like