AP3114/Computational Methods for

Physicists and Materials Engineers

7. Solving differential
equations I
Nov. 2, 2015

Contents
Introduction to differential equations
Symbolic solutions to ordinary differenti
al equations
Applications of ODEs: analysis of damped
and undamped free oscillations

Introduction to differential equations

Differential equations are equations involving derivatives of a function.
1. Ordinary differential equation
When the dependent variable is a function of a single independent variable,
the differential equation is said to be an ordinary differential equation (ODE).
q = -k(dT/dx)
2. Partial differential equation
If the dependent variable is a function of more than one variable, a
differential equation involving derivatives of this dependent variable is said to
be a partial differential equation (PDE).
2/x2+2/y2 + /x= 0.
3. Order and degree of an differential equation
The order of a differential equation is the order of the highest-order
derivative involved in the
equation. Thus, the ODE
q = -k(dT/dx)
is a first-order equation,

Introduction to differential equations

The degree of a differential equation is the highest power to which the
highest-order derivative
is raised. Therefore, the equation
(d3y/dt3)2+(d2y/dx2)5-xy = ex,
is a third order, second-degree ODE, while the equation
y/t = c(y/x),
is a first-order, first-degree PDE.
4. Linear and non-linear equations
An equation in which the dependent variable and all its pertinent derivatives
are of the first
degree is referred to as a linear differential equation. Otherwise, the equation
is said to be
non-linear. Examples of linear differential equations are:
d2x/dt2+ (dx/dt) + 0x = Asin(ft),
and
C/t + u(C/x) = D(2C/x2).

Introduction to differential equations

5. Constant or variable coefficients
The following equation:
d3y/dt3+(d2y/dx2)2-5y = ex,
where all the coefficients accompanying the dependent variable and its
derivative are constant, would be classified as a third order, linear ODE with
constant coefficients. Instead, the equation
2C/t2 u(x,t)(C/x) = 0,
would be classified as a second-order, linear PDE with variable coefficients.
6. Homogeneous and non-homogeneous equations
Typically, differential equations are arranged so that all the terms involving
the dependent
variable are placed on the left-hand side of the equation leaving only
constant terms or terms
involving the independent variable(s) only in the right-hand side. When
arranged in this
fashion, a differential equation that has a zero right-hand side is referred to

Introduction to differential equations

On the other hand, if the right-hand side of the equation, after placing the
terms involving the
dependent variable and its derivatives on the left-hand side, is non-zero, the
equation is said
to be non-homogeneous. Non-homogeneous versions of the last two
equations are:
d2x/dt2+ (dx/dt) + ox = Aoe-t/,
and
(x-1)(dy/dx) + 2xy = x2-2x.
7. Solutions
A solution to a differential equation is a function of the independent
variable(s) that, when
replaced in the equation, produces an expression that can be reduced,
through algebraic
manipulation, to the form 0 = 0. For example, the function
y = sin x,

Introduction to differential equations

8. General and particular solutions
A general solution is one involving integration constants so that any choice of
those constants represents a solution to the differential equation. For
example, the function
x = Ce-t,
is a general solution to the equation
dx/dt + x = 0,
because, substituting Ce-t for x in the equation produces
- Ce-t+ Ce-t= 0.
A particular solution is a solution corresponding to a specific value of the
integration
constants. For example, the function
y = x2/2
is a particular solution to the equation,
dy/dx x = 0.
A general solution for this equation would be
y = x2/2 + C,
where C is an arbitrary integration constant.
Given the solution of the homogeneous equation, yh(x), the solution of the
corresponding nonhomogeneous equation, y(x), can be written as

Introduction to differential equations

9. Verifying solutions using SCILAB
Since SCILAB is not a symbolic environment, it is not suitable for the
verification of solutions other than polynomial solutions SCILAB provides
function derivat to calculate derivatives of polynomials. If we have a function
that can be expressed as a polynomial, we can use function derivat to check
if that function satisfies a particular differential equation as illustrated in the
example below:
Check that the function y(x) = x3-2x+4 is a solution to the differential
equation, d2y/dx2- 6x = 0
using SCILAB:

Introduction to differential equations

10. Initial conditions and boundary conditions
To determine the specific value of the constant(s) of integration, we need to
provide values of the solution, or of one or more of its derivatives, at specific
points. These values are referred to as the conditions of the solution. For
example, we could specify that the solution to the equation.
d2y/dt2+y = 0,
must satisfy the conditions
y(0) = -5,
and
dy/dt = -1 at t = 5.
Initial conditions are provided at a single value of the independent variable so
that after
evaluating those conditions at that point all the integration constants are
uniquely specified.
In general, first order differential equations include one integration constant,
requiring only
one condition to be evaluated to uniquely determine the solution. Thus, this
type of equations

Introduction to differential equations

Boundary conditions, on the other hand, are provided at more than one value
of the independent variable(s). The term boundary conditions is used
because the function is
evaluated at the boundaries of the solution domain in order to specify the
solution.
An example of initial conditions used in a solution will be to solve the
equation
d2u/dt2+ 2(du/dt) = 0,
given
u(0) = 1, du/dt|t=0= -1.
An example of boundary conditions used in a solution will be to solve the
equation
d2y/dx2+y = Asinx,
using
y(0) = A/2, and y(1) = -A/2.
In general, the solution of a n-th order ODE requires n conditions.

Symbolic solutions to ordinary differenti

al equations
By symbolic solutions we understand those solutions that can be expressed
as a closed-form function of the independent variable.
1. Solution techniques for first-order, linear ODEs with constant coefficients
A first order equation is an equation of the form
a(dy/dx)n+ bym= f(x),
where a, b, n and m are, in general, real numbers. Some specific techniques
for linear
equations, i.e., when n = m = 1, follow. This catalog of solutions for linear
ODEs is intended as
a review of the techniques.
Equations of the form: dy/dx = f(x)
An equation of the form dy/dx = f(x) can be re-written as
dy= f(x)dx,
and a general solution found by direct integration,

Symbolic solutions to ordinary differenti

al equations
dy = f(x)dx,
or
y = f(x)dx+C.
If an initial condition y(xo) = yo, is given, then the integration can be
calculated as
or

If the function f(x) is a polynomial, we can use the SCILAB user-defined

function intpoly to produce the indefinite integral. Consider the example in
which dy/dx = f(x) = 3x3+ x + 2.
The solution, with the help of SCILAB, is calculated as:

Symbolic solutions to ordinary differenti

al equations

ser-defined function intpoly:

Symbolic solutions to ordinary differenti

al equations

Thus, the general solution is :

y(x) = 2x+0.5*x2+0.75*x4+ C

Symbolic solutions to ordinary differenti

al equations
2. Solutions of homogeneous linear equations of any order with constant
coefficients
Consider the linear, constant-coefficient, homogeneous ODE of order n:
where the coefficients b0, b1, , bn-1, are constant. We can use the operator
D(k)= d (k) /dx(k)
to re-write the equation as
Treating the operators D(k), (k = n, n-1, , 1), as algebraic terms, the
equation is re-written as
The idea is that the linear combination of the operators, shown above in
square brackets, is
applied to the function y(x), in a similar manner as algebraic terms would be
multiplied to it.

Symbolic solutions to ordinary differenti

al equations
Suppose that the characteristic equation has n independent roots, then the
general solution of
the linear, constant-coefficient, homogeneous ODE of order n given earlier is
If out of the n roots there is one that has multiplicity m, then the m terms
corresponding to
this root in the solution, will be
Example 1 Determine the general solution to the homogeneous equation

In terms of the D operator, this ODE can be written as

Symbolic solutions to ordinary differenti

al equations
The characteristic equation corresponding to this ODE is
To obtain solutions to this equation in SCILAB use:

Thus, a general solution to the ODE under consideration is:

Applications of ODEs: analysis of damped and

undamped free oscillations
Consider the mass-spring system shown in the figure below. The mass is
removed from its equilibrium position (x = 0) and released at a position x =
x0 at t=0. At the moment of its release the body was moving with a speed v
= vo. The diagram shows the body of mass m being acted upon by the
restoring force of the spring, Fs = - kx, and by a viscous damping force, Fv = v = -(dx/dt).

Applications of ODEs: analysis of damped and

undamped free oscillations
Newtons second law, when applied in the x-direction to the mass m is written
as:
which results in the second-order, linear, ordinary differential equation:

Undamped motion
Let us first consider the case in which the motion is undamped, i.e., = 0.
The equation in
this case reduces to

e corresponding characteristic equation is

with solutions,

Applications of ODEs: analysis of damped and

undamped free oscillations

s result suggest a solution of the form

lternatively, by taking

he solution can be written as

The quantity
is known as the natural angular frequency of the harmonic motion that
results when no viscous
damping is present.

Applications of ODEs: analysis of damped and

undamped free oscillations

Damped motion
If damping occurs (0), the characteristic equation becomes

whose solutions are

where
The nature of the solution will depend on the relative size of the coefficients
and o, as
follows:

is real, and the solutions of the characteristic equation are

Applications of ODEs: analysis of damped and

undamped free oscillations
The solution to the ODE, therefore, is written as

The parameter

represents the damped angular frequency of the oscillation, and 1

represents the corresponding angular phase. A0 is the amplitude of the
oscillation at t = 0. If we define a variable amplitude,
then the solution to the ODE, also known as the signal, can be written as

Please notice that this solution is very similar to the case of an undamped
oscillation, except for the fact that in a damped oscillation the amplitude
decreases with time. The amplitude decreases, or decays, with time because
the parameter = /(2m)is positive. Therefore, the function exp(-t) decreases
with time.

Applications of ODEs: analysis of damped and

undamped free oscillations
If = o, then the characteristic equation produces the solution = -, with
multiplicity 2, in which case the solution becomes
This solution represents a linear function of t subjected to a decay factor,
exp(-t).
If > o, then (2-o2) = K is real, and K < ,the solutions of the
characteristic equation become
both negative. Therefore, the resulting signal can be written as:

Notice that the last two cases, namely, = o and > o, produce signals
that decay with time. These cases correspond to harmonic motions that are
said to be over-damped, i.e., the viscous damping is large enough to quickly
damp out any oscillation after the body of mass m is released.

Applications of ODEs: analysis of damped and

undamped free oscillations
The expression for the position of a damped oscillatory motion is given by

le the velocity, v(t) = dx/dt, is given by

Given the initial conditions x(t0) = x0 and v(t0) = v0, we can form a system of
two non-linear
equations in the unknowns A0 and 1, namely,

With appropriate values of the parameters , I, t0, x0, and v0, we can use
SCILAB function
fsolve to obtain the values of A0 and 1.
he expression for the acceleration

Applications of ODEs: analysis of damped and

undamped free oscillations

xample Damped oscillatory motion:

Plot position, velocity, and acceleration corresponding to the following
parameters: m = 1 kg, = 0.1Ns/m, k = 0.5 N/m. To determine the
constants Ao and 1, use initial conditions, x0= 1.5 m, and v0= -5.0 m/s.
With these values,

Since, < o, the resulting signal is that of a damped oscillation with

To solve for the constants Ao and 1 with SCILAB, we first define the set of
non-linear equations to be solved. In the function thus defined A 0 is
represented by s(1) and 1 is represented by s(2):

Applications of ODEs: analysis of damped and

undamped free oscillations
Next, we enter the known values, and select a first guess for the solution s0:

The solution for s(1) = A0 and s(2) = 1 is obtained by using fsolve:

Applications of ODEs: analysis of damped and

undamped free oscillations
Thus, A0= 7.1221364and 1= 1.3591997, and the position x(t) is given by

Expressions for the position x(t), velocity v(t), and acceleration a(t) for this
motion can be
entered into SCILAB by defining the following functions:

Applications of ODEs: analysis of damped and

undamped free oscillations
To plot the signals x(t), v(t), and a(t) in the t-interval (0,30) use the following
SCILAB commands:

Notice the oscillatory nature of the three functions, as well as their

amplitudes decay with time as expected.

Applications of ODEs: analysis of damped and

undamped free oscillations

Applications of ODEs: analysis of damped and

undamped free oscillations

eating phase portraits of oscillatory motion

A phase portrait for oscillatory, or any kind of, motion is a plot involving the
dependent variable and one of its derivatives, or two derivatives of the
dependent variable. For example, a plot of velocity, v(t), versus position, x(t),
represents a phase portrait. Other phase portraits would be a(t) vs. x(t), and
a(t) vs. v(t).
Example.
Plot the time-dependent plots and phase portraits for the signal obtained in
Example in the previous section.
To plot these phase portraits we generate data on position, velocity, and
acceleration as function of time t in the interval [0,90], as follows:

Applications of ODEs: analysis of damped and

undamped free oscillations

e phase portraits are generated as follows:

Applications of ODEs: analysis of damped and

undamped free oscillations

Applications of ODEs: analysis of damped and

undamped free oscillations

HW-8
Due: 9am, Nov 9, 2015
Late policy: 10% off per day
To be submitted online to Canvas:
Save all your codes (*.sci files) into one folder to submit;
Do use // to add comments clearly which question you
are referring to, when we execute your codes, answers
shall be shown on the screen.

34

Exercise
1. Determine the general solution to the following linear ordinary differential
equations using
the corresponding characteristic equation:

2. Plot the time variation of position, velocity, and acceleration of a damped

mechanical
oscillator for the following parameters:

Plot v-vs-x, a-vs-x, and a-vs-v phase portraits of the motions described above
problem.

Note:
Record your intermediate results using Print Screen
Save your results and submit them through your usual way.