CHAPTER 2

NUMERICAL SOLUTIONS OF
ORDINARY DIFFERENTIAL EQUATIONS
• Solutions of 1st order ODEs
• Euler’s method
• Improved Euler Method
• Runge Kutta (RK) Methods
• Adaptive stepping
• Stability of RK methods
• Solutions of 2nd and higher order ODEs
• Solutions of coupled ODEs
• Solutions of boundary value problems

İ.H. TUNCER & Y. ÖZYÖRÜK 21 AE305 - Numerical Methods in Aerospace Engineering @ 20201

Differential Equations
An equation which is composed of an unknown function and its derivatives with respect to
independent variables is called a differential equation .

• When the function involves one independent variable, it is called an

Ordinary Differential Equation, ODE .

E.g. Airplane accelerating on ground :

W dV
= T − D(V ) − kfric [W − L(V )]
g dt

• This is in contrast to a Partial Differential Equation, PDE that involves two or more
independent variables.
E.g. Unsteady heat diffusion through a 1-D bar :
∂T ∂2T
=α 2
∂t ∂x

İ.H. TUNCER & Y. ÖZYÖRÜK 22 AE305 - Numerical Methods in Aerospace Engineering @ 20201

Differential Equations (cont’d)

• Differential equations are also classified according to their order : The highest derivative
determines the order of the differential equation. A second order eqn. would include a
second derivative, and so on so forth.
First-order examples
dI (t) P − I (t)
A Covid-19 spread model : =β I (t), P : population, I : infected # of people
dt P
W dV
A/C velocity on ground : = T − D(V ) − kfric [W − L(V )], T : thrust, D : drag, L : lift
g dt
∂u ∂u
Convection equation : +a = 0, u : quantity convected, a : convection speed
∂t ∂x

Second-order examples
d 2x dx
Mass-spring-damper : m +c + kx = 0, m : mass, c : damping cons, k : spring cons
dt 2 dt
2 2
∂ y ∂ y
Wave equation for a string : − a2 2 = 0, y : discplacement, a : wave speed
∂t 2 ∂x

İ.H. TUNCER & Y. ÖZYÖRÜK 23 AE305 - Numerical Methods in Aerospace Engineering @ 20201
Solution of ODEs - Boundary Conditions
For the solution of an nth order ordinary differential equation, n conditions are required to
obtain a unique solution.

• If all the conditions are specified at the beginning of the solution domain, then the
problem is called an initial-value problem, IVP .
2
m ddt x2 + c dx
dt + kx = 0 2nd-order ODE
Conditions:
x|t=0 = x0 −→ IC #1
dx
dt |t=0 = v0 −→ IC #2
(illustration : wiki)

• In boundary-value problems, BVP, the conditions are specified at both ends of the
solution domain.
2
Equation : ddxT2 + kq̇ = 0 2nd-order ODE
Boundary Conditions:
T |x=0 = T0 −→ BC #1
−k dT |
dx x=L
= qL00 −→ BC #2

İ.H. TUNCER & Y. ÖZYÖRÜK 24 AE305 - Numerical Methods in Aerospace Engineering @ 20201

Solution of ODEs - Variable Discretization

• Numerical solutions can only be obtained for discrete values of the independent variables
• Consider a function y (x), where x is the spatial coordinate. y is the dependent variable,
and x is the independent variable.
• Then, treating the problem numerically requires discretization of the independent variable
first, and assigning indices to the discrete points.

• If spacing between two consequtive discrete

points is fixed and equal to ∆x:
xi = xo + (i − 1)∆x, i = spatial index 1..N
• The dependent variable at the discrete point
(i) is then denoted by yi ≡ y (xi ) yi ≡ y (xi )

• If the indepenent variable is time, the

convention is to use n as the index: tn and
y n ≡ y (tn ).

İ.H. TUNCER & Y. ÖZYÖRÜK 25 AE305 - Numerical Methods in Aerospace Engineering @ 20201

Solution of 1st Order ODEs - Euler’s Method

• A first order ordinary differential equation forms an initial value problem and is of the form
dy
y0 ≡ = f (x, y )
dx

• Numerical methods to solve an initial value problem is based on the first order Taylor
Series Expansion, which is known as Euler’s method :

dy 
y (x + ∆x) = y (x) + ∆x + O(∆x 2 )
dx (x,y )
y (x + ∆x) = y (x) + ∆x y 0 (x, y )
y (x + ∆x) = y (x) + ∆x f (x, y )
yi+1 = yi + ∆x f (xi , yi )

which is of the form Next-Value = Previous-Value + Step-Size × SLOPE≡ODE

• Note that the slope in TSE is provided by the ODE itself. This formula can be applied
step by step to compute the solution along the independent variable.
İ.H. TUNCER & Y. ÖZYÖRÜK 26 AE305 - Numerical Methods in Aerospace Engineering @ 20201
Solution of 1st Order ODEs - Euler’s Method (cont’d)
Example: Free-falling body motion

• The velocity of a free falling body may be modeled by

Newton’s second law - Conservation of Momentum • This is a 1st order ODE in the form of
equation: v 0 = f (t, v )
~
~ = m~a or d~v = F
F
dt m • Fortunately, the analytical solution may
where F~ is the net force on the body, which is be obtained by change and separation of
composed of gravitational force F ~ G and the upward variables (y = mgm−cv ).
force due to air resistance F~R . For a zero initial velocity of the object,
the solution is given by
~
F = ~G − F
F ~R
gm
= (mg − cv )êz v (t) = (1 − e −(c/m)t )
c
• Air resistance, drag, can be formulated in a variety of
ways. Above is a simple one which is linearly
proportional to velocity, where c is the drag coefficient. • However, many mathematical models
may not be solved analytically, such as
• The governing equation for downward velocity then when FR = cv 2 is assumed!.
becomes
dv mg − cv c
= =g− v
dt m m

İ.H. TUNCER & Y. ÖZYÖRÜK 27 AE305 - Numerical Methods in Aerospace Engineering @ 20201

Solution of 1st Order ODEs - Euler’s Method (cont’d)

• If the initial condition for the velocity is given at time t1 , the velocity of the object at a
later time t2 = t1 + ∆t. The new value is in turn employed to extend the computation to
the next time step t3 = t2 + ∆t and so on.
NewValue = OldValue + StepSize × Slope
= OldValue + StepSize × ODE
time step # time velocity update
n =1 t = t1 v = v 1 Initial condition

n =2 t2 = t1 + ∆t v 2 = v 1 + ∆t g − mc v 1

n =3 t3 = t2 + ∆t v 3 = v 2 + ∆t g − mc v 2

n =4 t4 = t3 + ∆t v 4 = v 3 + ∆t g − mc v 3
... ... ...

n tn = tn−1 + ∆t v n = v n−1 + ∆t g − mc v n−1

n+1 n c n
n+1 tn+1 = tn + ∆t v = v + ∆t g − m v

• Note that the slope in Euler method is evaluated at the beginning of the interval, which does not
represent the average slope over the interval!

İ.H. TUNCER & Y. ÖZYÖRÜK 28 AE305 - Numerical Methods in Aerospace Engineering @ 20201

IVPs - Euler Method (cont’d)

!−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
! . . A SIMPLE EULER SOLVER f o r ODEs ! . . Solution loop
! . . AE305 − N u m e r i c a l Methods do w h i l e ( time . l t . finaltime )
!−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−− time = time + stepsize
Module f r e e _ f a l l i n g _ b o d y _ d a t a velocity = velocity + ODE ( velocity )∗stepsize
grav = 9 . 8 1 w r i t e ( 1 , ” ( 2 f 1 2 . 3 ) ” ) time , velocity
fric = 1 2 . 5 enddo
xmass = 7 0 .
End module ! . . Close the output file
close (1)
Program EULER
c h a r a c t e r ∗40 fname ! . . P l o t t h e s o l u t i o n w i t h ” x g r a p h ” , ” xmgr ” o r ” g n u p l o t ” on←-
L i n u x X−windows
! . . Read t h e s t e p s i z e ! c a l l SYSTEM( ” xg v e l o c i t y . d a t &”)
p r i n t ∗,” ” ! c a l l SYSTEM( ” xmgr v e l o c i t y . d a t &”)
p r i n t ∗, ” E n t e r S t e p S i z e and F i n a l T i m e :> ” ! c a l l SYSTEM( ” g n u p l o t p l o t . g p l ” )
r e a d (∗ ,∗) stepsize , finaltime
stop
! . . open t h e o u t p u t f i l e End
p r i n t ∗ , ” E n t e r t h e o u t p u t f i l e name [ v e l o c i t y . d a t ] : ”
r e a d (∗ , ” ( a ) ” ) fname !−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−
i f ( fname . eq . ” ” ) fname = ” v e l o c i t y . d a t ” ! . . D e f i n e t h e ODE a s a f u n c t i o n
open ( 1 , f i l e =fname , form=” f o r m a t t e d ” ) F u n c t i o n ODE ( vel )
use free_falling_body_data
! . . S e t t h e I n i t i a l C o n d i t i o n s and o u t p u t them ODE = grav − fric / xmass ∗ vel
time = 0. return
velocity = 0 . End
w r i t e ( 1 , ” ( 2 f 1 2 . 3 ) ” ) time , velocity

İ.H. TUNCER & Y. ÖZYÖRÜK 29 AE305 - Numerical Methods in Aerospace Engineering @ 20201