Sources of Error

Associate Prof. Ahmed Eldessouky

FEEE
BAS225 Numerical Methods

1
 Two sources of numerical error
1) Round off error
2) Truncation error

2
 Round off Error
• Caused by representing a number
approximately

1
0.333333
3
2 1.4142...

3
 Problems created by round off error

• 28 Americans were killed on February 25,

1991 by an Iraqi Scud missile in Dhahran,
Saudi Arabia.
• The patriot defense system failed to track
and intercept the Scud. Why?

4
Problem with Patriot missile
• Clock cycle of 1/10 seconds was
represented in 24-bit fixed point
register created an error of 9.5 x
10-8 seconds.
• The battery was on for 100
consecutive hours, thus causing
an inaccuracy of

8 s 3600s
9.5 10 100hr 
0.1s 1hr
0.342s
5
 Problem (cont.)
• The shift calculated in the ranging system
of the missile was 687 meters.
• The target was considered to be out of
range at a distance greater than 137
meters.

6
Effect of Carrying Significant
Digits in Calculations

7
Find the contraction in the
diameter
Tc

D D  (T )dT
Ta

Ta=80oF; Tc=-108oF;
D=12.363”
α = a0+ a1T + a2T 2

8
 Thermal Expansion
Coefficient vs Temperature
T(oF) α (μin/in/oF)
-340 2.45
-300 3.07
-220 4.08
-160 4.72
D  D  T
-80 5.43
0 6.00
40 6.24
80 6.47

9
Regressing Data in Excel
(general format)

Alpha ( in/in/ F)
8.00

o
4.00

0.00
-400 -200 0 200
o
T ( F)

α = -1E-05T2 + 0.0062T + 6.0234 10

Observed and Predicted Values
α = -1E-05T2 + 0.0062T + 6.0234
T(oF) α (μin/in/oF) α (μin/in/oF)
Given Predicted
-340 2.45 2.76
-300 3.07 3.26
-220 4.08 4.18
-160 4.72 4.78
-80 5.43 5.46
0 6.00 6.02
40 6.24 6.26
80 6.47 6.46 11
Regressing Data in Excel
(scientific format)

Alpha ( in/in/ o F)
8.00

4.00

0.00
-400 -200 0 200

T ( o F)

α = -1.2360E-05T2 + 6.2714E-03T + 6.0234

12
Observed and Predicted Values
α = -1.2360E-05T2 + 6.2714E-03T + 6.0234
T(oF) α (μin/in/oF) α (μin/in/oF)
Given Predicted
-340 2.45 2.46
-300 3.07 3.03
-220 4.08 4.05
-160 4.72 4.70
-80 5.43 5.44
0 6.00 6.02
40 6.24 6.25
80 6.47 6.45 13
Observed and Predicted Values
α = -1.2360E-05T2 + 6.2714E-03T + 6.0234
α = -1E-05T2 + 0.0062T + 6.0234
T(oF) α (μin/in/oF) α (μin/in/oF) α (μin/in/oF)
Given Predicted Predicted
-340 2.45 2.46 2.76
-300 3.07 3.03 3.26
-220 4.08 4.05 4.18
-160 4.72 4.70 4.78
-80 5.43 5.44 5.46
0 6.00 6.02 6.02
40 6.24 6.25 6.26
80 6.47 6.45 6.46
14
Truncation Error

15
 Truncation error
• Error caused by truncating or
approximating a mathematical
procedure.

16
Example of Truncation Error

Taking only a few terms of a Maclaurin

x
series to
approximat e
e x x 2
x 3
e 1  x    ....................
2! 3!
If only 3 terms are used,
 x 2

Truncation Error e   1  x  
x

 2! 

17
Another Example of Truncation
Error

Using a x to f (x)
finite f ( x)  f ( x  xapproximate
)  f ( x)
x

secant line
P

tangent line

Figure 1. Approximate derivative using finite Δx

18
Another Example of Truncation
Error

Using finite rectangles to approximate

an integral.
y

90

2
y=x
60

30

0 x
0 1.5 3 4.5 6 7.5 9 10.5 12

19
 Example 1 —Maclaurin series
1.2
Calculate the value e with an absolute
of
relative approximate error of less than
2 3
1%.e 1  1.2 
1.2 1.2

1.2
 ...................
2! 3!
n 1.2 a %
e Ea
1 1 __ ___
2 2.2 1.2 54.545
3 2.92 0.72 24.658
4 3.208 0.288 8.9776
5 3.2944 0.0864 2.6226
6 3.3151 0.020736 0.62550

6 terms are required. How many are required 20to

get at least 1 significant digit correct in your
Example 2 —Differentiation

f ( x)  x 2 f ( x  x)  f ( x)
Find f (3)
for using f ( x) 
x
an x 0.2
d ' f (3  0.2)  f (3)
f (3) 
0.2
2 2
f (3.2)  f (3) x 30..2  3 10.24  9 1.24
  1   6.2
0 .2 0 .2 0.2 0.2

The actual value

is f ' ( x) 2 x, f ' (3) 2 3 6
Truncation error is 6  6.2  0.2
then,
Can you find the truncation error 21
 Example 3 — Integration

Use two rectangles of equal width to

approximate the area under the curve
f ( x)  x 2 over the
for [3,9]
y interval 9

x
2
90
dx
3

2
y=x
60

30

0 x
0 3 6 9 12

22
 Integration example (cont.)

Choosing a width of 3, we
havex 2 dx ( x 2 ) (6  3)  ( x 2 ) (9  6)
9


3
x 3 x 6

(3 2 )3  (6 2 )3
27  108 135
Actual value is given by
9 9
 x3   9 3
 33


2
x dx       234
3  3 3  3 
Truncation error is then
234  135 99
Can you find the truncation error with 4 23
Propagation of Errors
 Propagation of Errors

In numerical methods, the calculations are

not made with exact numbers. How do these
inaccuracies propagate through the
calculations?

http://numericalmethods.eng.usf.e
du
25
 Example 1:
Find the bounds for the propagation in adding two numbers. For example
if one is calculating X +Y where
X = 1.5 ± 0.05
Y = 3.4 ± 0.04
Solution
Maximum possible value of X = 1.55 and Y = 3.44

Maximum possible value of X + Y = 1.55 + 3.44 = 4.99

Minimum possible value of X = 1.45 and Y = 3.36.

Minimum possible value of X + Y = 1.45 + 3.36 = 4.81

Hence
4.81 ≤ X + Y ≤4.99.

http://numericalmethods.eng.usf.e
du
26
Propagation of Errors In Formulas

If f is a function of several variables

X 1 , X 2 , X 3 ,......., X n  1 , X n
then the maximum possible value of the error f in is

f f f f
f  X 1  X 2  .......  X n  1  X n
X 1 X 2 X n  1 X n

http://numericalmethods.eng.usf.e
du
27
 Example 2:

The strain in an axial member of a square

cross-section is given by
F
 2
h E
Given
F 72 0.9 N
h 4 0.1 mm
E 70 1.5 GPa

Find the maximum possible error in the

measured strain.
http://numericalmethods.eng.usf.e
du
28
 Example 2:
Solut
ion 72
 3 2 9
(4 10 ) (70 10 )
64.286 10  6
64.286

  
  F  h  E
F h E
http://numericalmethods.eng.usf.e
du
29
 Example 2:
 1  2F  F
 2  3  2 2
F h E h hE E h E
Thus
1 2F F
E  2 F  3 h  2 2 E
h E hE h E
1 2 72
 3 2 9
0.9  3 3 9
0.0001
(4 10 ) (70 10 ) (4 10 ) (70 10 )
72 9
 1.5 10
(4 10 3 ) 2 (70 109 ) 2
5.3955
Hence
(64.286 5.3955 )
http://numericalmethods.eng.usf.e
du
30
 Example 3:

Subtraction of numbers that are nearly equal can create unwanted

inaccuracies. Using the formula for error propagation, show that
this is true.

Solution
Let z  x  y
z z
z 
Then x  y
x y
 (1)x  ( 1)y
 x  y

Sothe x  change
z relative y is

z x y http://numericalmethods.eng.usf.e
du
31
 Example 3:

For example if
x 2 0.001
y 2.003 0.001

z 0.001  0.001

z | 2  2.003 |
=
0.6667
=
66.67%
http://numericalmethods.eng.usf.e
du
32
THE END