CVNG 1010 |

University of the
West Indies
St. Augustine

EXAMPLE: Use the Taylor Series to approximate the

true value of the root equation for a base point x0 = 1,
and h = 0.001.

1. Determine the derivatives up to a 3rd order

2. Solve the derivatives for the base point

3. Use Taylor series to solve up to assumed order

27
CVNG 1010 |
University of the
West Indies
St. Augustine

1. Determine the derivatives up to an assumed order

2. Solve the derivatives for the base point and h values

𝑓 ! 1 = 0.5 𝑓 !! 1 = −0.25 𝑓 !!! 1 = 0.375
3. Use Taylor series to solve up to assumed order

28
CVNG 1010 |
University of the
West Indies
St. Augustine

Other functions:

29
 University of the
West Indies
St. Augustine

Review of Partial Differential Equations (PDE)

30
 University of the
West Indies
St. Augustine

Review of PDE
Order – order of its highest derivative.

Linear – there is no product of U or the partial

derivatives of U. U and all the partial derivatives occur to
the first power only.

Dimension – the number of spatial variables (eg, x, y, z)

it contains

The following PDE example is second order, linear, and 2D

31
 University of the
West Indies
St. Augustine

Review of PDE
Solving the PDE involves finding the unknown function
U.

Approximation is made at discrete values of independent

variables, and the approximate scheme is implemented
in the computer programme.

32
 University of the
West Indies
St. Augustine

Review of PDE
PDE can be classified into 3 major categories:

1. Elliptic – no time derivative

2. Parabolic

3. Hyperbolic

33
 University of the
West Indies
St. Augustine

Review of PDE

PDE must have proper initial conditions and boundary

conditions to be considered a well posed problem.

Domain of dependence – PDE can only be influenced by

points within that region

Zone of influence – PDE can influence points within that

zone

34
 University of the
West Indies
St. Augustine

Finite Difference Method

The Finite Difference Method replaces all the Partial
derivatives and other terms in the PDE with
approximations.

The region over which the independent variables in the

PDE are defined is replaced by a finite grid (mesh) on
which the dependent variables are approximated.

The partial derivatives are approximated at each point

using Taylor’s theorem.

35
 University of the
West Indies
St. Augustine

Taylor’s Theorem Applied to FDM

Let us write the Taylor Series as a function of U:

36
 University of the
West Indies
St. Augustine

Taylor’s Theorem Applied to FDM

In FDM we know the values of U at the grid points. We
will replace the PDE we are solving with approximations
at the grid points.

x0 and x0+h are grid points

U(x0) and U(x0+h) are known

Truncating after the first derivative we can get

37
 University of the
West Indies
St. Augustine

Taylor’s Theorem Applied to FDM

We can rearrange the equation to get

Neglecting the error gives the first order FD

approximation to Ux(x0). This is forward difference with a
step size of h.

38
 University of the
West Indies
St. Augustine

What does this signify?

U(x)
We wish to find the
x2,U(x2) instantaneous change
x0,U(x0) at x = x0

x1,U(x1)
x
∆x ∆x

39
 University of the
West Indies
St. Augustine

What does this signify?

U(x) FD = (U(x2)- U(x0))/ (x2-x0)

x2,U(x2)
x0,U(x0)

x
∆x=h

40
 University of the
West Indies
St. Augustine

What does this signify?

U(x) BD = (U(x0)- U(x1))/ (x0-x1)

x0,U(x0)

x1,U(x1)
x
∆x ∆x

41
 University of the
West Indies
St. Augustine

What does this signify?

U(x) CD = (U(x2)- U(x1))/ (x2-x1)

CD = ((U(x2) – (U(x1))/ 2∆x

x2,U(x2)
x0,U(x0)
Most accurate
representation of the
actual slope at the
x1,U(x1) point of interest. Can
x be improved by taking
smaller values of ∆x
∆x ∆x

42
University of the
West Indies
St. Augustine

43
University of the
West Indies
St. Augustine

44
 University of the
West Indies
St. Augustine

Taylor’s Theorem Applied to FDM

Example: For the function U(x) = x2 use FD to
approximate Ux(3) using step size h = 0.1.

The exact answer is 6

so the error is 0.1

Repeating with h = 0.05 gives an

error of 0.05
45
Tutorial 1 University of the
West Indies
St. Augustine

All answers must include student name, student ID, and question
number. Save answer to html files.

Use the Taylor Series to solve the following functions for 20 steps:
1) f(x) = x^3 - 3x^2 + 5x + 10 , (x0 = 2; h= 1)
2) f(x) = sin(x), (x0 =0, h=0.1) rad

Plot the cumulative terms for the Taylor Series. All plots must
have axes labelled and accompanied legends.

Tabulate the cumulative results for the Taylor Series. Your table
must contain appropriate headings. 46
CVNG 1010 |
University of the
West Indies
St. Augustine

Reference Texts:
Numerical Methods for Engineers 7th Edition…Chapra,
Canale

Numerical Analysis Using MATLAB and

Spreadsheets…Karris

Numerical Methods for Engineers: Methods and

Applications…Ayyub, McCuen

Introduction to Numerical Methods and MATLAB

programming for engineers…Young, Mohlenkamp 47