For Engineers Ch4

1
CHAPTER 4
4.1 (a) For this case xi = 0 and h = x. Thus,
x2
f ( x) = f (0) + f ' (0) x + f " (0) + ⋅⋅⋅
2
f (0) = f ' (0) = f " (0) = e 0 = 1
x2
f ( x) = 1 + x + + ⋅⋅⋅
2
(b)
h2 h3
f ( xi +1 ) = e − xi − e − xi h + e − xi − e − xi + ⋅⋅⋅
2 6
for xi = 0.2, xi+1 = 1 and h = 0.8. True value = e–1 = 0.367879.
zero order:
f (1) = e −0.2 = 0.818731

0.367879 − 0.818731
εt = × 100% = 122.55%
0.367879
first order:
f (1) = 0.818731 − 0.818731(0.8) = 0.163746

0.367879 − 0.163746
εt = × 100% = 55.49%
0.367879
second order:
0.8 2
f (1) = 0.818731 − 0.818731(0.8) + 0.818731 = 0.42574
2
0.367879 − 0.42574
εt = × 100% = 15.73%
0.367879
third order:
0.8 2 0.8 3
f (1) = 0.818731 − 0.818731(0.8) + 0.818731 − 0.818731 = 0.355875
2 6
0.367879 − 0.355875
εt = × 100% = 3.26%
0.367879
PROPRIETARY MATERIAL. © The McGraw-Hill Companies, Inc. All rights reserved. No part of this Manual may
be displayed, reproduced or distributed in any form or by any means, without the prior written permission of the
publisher, or used beyond the limited distribution to teachers and educators permitted by McGraw-Hill for their individual
course preparation. If you are a student using this Manual, you are using it without permission.
2
4.2 Use the stopping criterion
ε s = 0.5 × 10 2 − 2 % = 0.5%
True value: cos(π/3) = 0.5
zero order:
π 
cos  = 1
3
0.5 − 1
εt = × 100% = 100%
0.5
first order:
π 
cos  = 1 −
( π / 3) 2 = 0.451689
3 2
0.451689 − 1
ε t = 9.66% εa = × 100% = 121.4%
0.451689
second order:
π 
cos  = 0.451689 +
( π / 3) 4 = 0.501796
3 24
0.501796 − 0.451689
ε t = 0.359% εa = × 100% = 9.986%
0.501796
third order:
π 
cos  = 0.501796 −
( π / 3) 6 = 0.499965
3 720
0.499965 − 0.501796
ε t = 0.00709% εa = × 100% = 0.366%
0.499965
Since the approximate error is below 0.5%, the computation can be terminated.
4.3 Use the stopping criterion: ε s = 0.5 × 10 2−2 % = 0.5%
True value: sin(π/3) = 0.866025…
3
zero order:
π  π
sin  = = 1.047198
3 3
0.866025 − 1.047198
εt = × 100% = 20.92%
0.866025
first order:
π 
sin  = 1.047198 −
( π / 3) 3 = 0.855801
3 6
0.855801 − 1.047198
ε t = 1.18% εa = × 100% = 22.36%
0.855801
second order:
π 
sin  = 0.855801 +
( π / 3) 5
= 0.866295
3 120
0.866295 − 0.855801
ε t = 0.031% εa = × 100% = 1.211%
0.866295
third order:
π 
sin  = 0.866295 −
( π / 3) 7 = 0.866021
3 5040
0.866021 − 0.866295
ε t = 0.000477% εa = × 100% = 0.0316%
0.866021
4.4 True value: f(3) = 554.
zero order:
f (3) = f (1) = −62
554 − (−62)
εt = × 100% = 111.191%
554
4
first order:
f (3) = −62 + f ' (1)(3 − 1) = −62 + 70( 2) = 78 ε t = 85.921%
second order:
f " (1) 138

f (3) = 78 + (3 − 1) 2 = 78 + 4 = 354 ε t = 36.101%
2 2
third order:
( 3)
f (1) 150
f (3) = 354 + (3 − 1) 3 = 354 + 8 = 554 ε t = 0%
6 6
Thus, the third-order result is perfect because the original function is a third-order
polynomial.
4.5 True value: f(2.5) = ln(2.5) = 0.916291...
zero order:
f (2.5) = f (1) = 0
0.916291 − 0
εt = × 100% = 100%
0.916291
first order:
f (2.5) = f (1) + f ' (1)(2.5 − 1) = 0 + 1(1.5) = 1.5
0.916291 − 1.5
εt = × 100% = 63.704%
0.916291
second order:
f " (1) −1 2
f (2.5) = 1.5 + ( 2.5 − 1) 2 = 1.5 + 1.5 = 0.375
2 2
0.916291 − 0.375
εt = × 100% = 59.074%
0.916291
third order:
( 3)
f (1) 2
f (2.5) = 0.375 + ( 2.5 − 1) 3 = 0.375 + 1.5 3 = 1.5
6 6
5
0.916291 − 1.5
εt = × 100% = 63.704%
0.916291
fourth order:
( 4)
f (1) −6 4
f (2.5) = 1.5 + ( 2.5 − 1) 4 = 1.5 + 1.5 = 0.234375
24 24
0.916291 − 0.234375
εt = × 100% = 74.421%
0.916291
Thus, the process seems to be diverging suggesting that a smaller step would be required for
convergence.
4.6 True value:
f ' ( x) = 75 x 2 − 12 x + 7
f ' (2) = 75(2) 2 − 12(2) + 7 = 283
function values:
xi–1 = 1.8 f(xi–1) = 50.96

xi = 2 f(xi) = 102
xi+1 = 2.2 f(xi+1) = 164.56
forward:
164.56 − 102
f ' (2) = = 312.8
0.2
283 − 312.8
εt = × 100% = 10.53%
283
backward:
102 − 50.96
f ' (2) = = 255.2
0.2
283 − 255.2
εt = × 100% = 9.823%
283
centered:
6
164.56 − 50.96
f ' (2) = = 284
2(0.2)
283 − 284
εt = × 100% = 0.353%
283
Both the forward and backward have errors that can be approximated by (recall Eq. 4.15),
f " ( xi )
Et ≈ h
2
f " ( 2) = 150 x − 12 = 150( 2) − 12 = 288
288
Et ≈ 0.2 = 28.8
2
This is very close to the actual error that occurred in the approximations
forward: E t ≈ 283 − 312.8 = 29.8

backward: E t ≈ 283 − 255.2 = 27.8
The centered approximation has an error that can be approximated by,
( 3)
f ( xi ) 2 150
Et ≈ − h =− 0.2 2 = −1
6 6
which is exact: Et = 283 – 284 = –1. This result occurs because the original function is a cubic
equation which has zero fourth and higher derivatives.
4.7 True value:
f " ( x) = 150 x − 12
f " ( 2) = 150( 2) − 12 = 288
h = 0.25:
f ( 2.25) − 2 f (2) + f (1.75) 182.1406 − 2(102) + 39.85938

f " ( 2) = = = 288
0.25 2 0.25 2
h = 0.125:
f ( 2.125) − 2 f (2) + f (1.875) 139.6738 − 2(102) + 68.82617

f " ( 2) = = = 288
0.125 2 0.125 2
7
Both results are exact because the errors are a function of 4th and higher derivatives which are
zero for a 3rd-order polynomial.
4.8
(
∂v cgte − ( c / m )t − gm 1 − e − ( c / m )t
=
)
= −1.38666
∂c c2
∂v ~
∆v(c~ ) = ∆c = 1.38666(1.5) = 2.079989
∂c
v(12.5) =
9.8(50)
12.5
( )
1 − e −12.5( 6 ) / 50 = 30.4533
v = 30.4533 ± 2.079989
Thus, the bounds computed with the first-order analysis range from 28.3733 to 32.5333. This
result can be verified by computing the exact values as
v(c − ∆c) =
9.8(50)
11
( )
1 − e − (11 / 50) 6 = 32.6458
v(c + ∆c) =
9.8(50)
14
( )
1 − e −(14 / 50) 6 = 28.4769
Thus, the range of ± 2.0844 is close to the first-order estimate.
4.9
v(12.5) =
9.8(50)
12.5
( )
1 − e −12.5( 6 ) / 50 = 30.4533
~ ) = ∂v ∆c~ + ∂v ∆m
∆v(c~, m ~
∂c ∂m
(
∂v cgte − ( c / m )t − gm 1 − e − ( c / m )t
=
)
= −1.38666
∂c c2
∂v gt −( c / m ) t g
= e
∂m m
( )
+ 1 − e −( c / m ) t = 0.871467
c
∆v(c~, m
~ ) = − 1.38666 (1.5) + 0.871467 ( 2) = 2.079989 + 1.742934 = 3.822923
v = 30.4533 ± 3.822923
8
4.10 For ∆T~ =20,
~ ∂H ~
∆H (T ) = ∆T
∂T
∂H
= 4 AeσT 3 = 4(0.15)0.9(5.67 × 10 −8 )650 3 = 8.408
∂T
~
∆H (T ) = 8.408( 20) = 168.169
Exact error:
H (670) − H (630) 1542.468 − 1205.81

∆H true = = = 168.3286
2 2
Thus, the first-order approximation is close to the exact result.
For ∆T~ =40,
~
∆H (T ) = 8.408( 40) = 336.3387
Exact error:
H (690) − H (610) 1735.055 − 1059.83

∆H true = = = 337.6124
2 2
Again, the first-order approximation is close to the exact result. The results are good because
H(T) is nearly linear over the ranges we are examining. This is illustrated by the following
plot.
2000
1600
1200
800
550 600 650 700 750
4.11 For a sphere, A = 4πr2. Therefore,
H = 4πr 2 eσT 4
At the mean values of the parameters,
9
H (0.15,0.9,550) = 4π (0.15) 2 0.90(5.67 × 10 −8 )(550) 4 = 1320.288
∂H ~ ∂H ~ ∂H ~
∆H = ∆r + ∆e + ∆T
∂r ∂e ∂T
∂H
= 8πreσT 4 = 17,603.84
∂r
∂H
= 4πr 2σT 4 = 1466.987
∂e
∂H
= 16πr 2 eσT 3 = 9.6021
∂T
∆H = 17603.84(0.01) + 1466.987(0.05) + 9.6021( 20) = 441.4297
To check this result, we can compute
H (0.14,0.85,530) = 4π (0.14) 2 0.85(5.67 × 10 −8 )(530) 4 = 936.6372
H (0.16,0.95,570) = 4π (0.16) 2 0.95(5.67 × 10 −8 )(570) 4 = 1829.178
1829.178 − 936.6372
∆H true = = 446.2703
2
4.12 The condition number is computed as

~
x f ' (~
x)
CN = ~
f (x )
 1 
1.00001 
(a)  2 1.00001 − 1  = 1.00001(158.1139) = 157.617
CN =
1.00001 − 1 + 1 1.003162
The result is ill-conditioned because the derivative is large near x = 1.
10(−e −10 ) 10(−4.54 × 10 −5 )

(b) CN = = = −10
e −10 4.54 × 10 −5
The result is ill-conditioned because x is large.
10
 300 
300 − 1
(c) −6
 300 + 1  = 300( −5.555556 × 10 ) = −0.99999444
2
CN =
300 2 + 1 − 300 0.0016667
The result is well-conditioned.
− xe − x − e − x + 1
x
x2 0.001(0.499667)
(d) CN = =
− 0.9995
= −0.0005
 e −1
−x
 
 x 
 
The result is well-conditioned.
(1 + cos x) cos x + sin x(sin x)

x
(1 + cos x) 2 3.141907( 20,264,237)
(e) CN = = = −10,001
sin x − 6366.2
1 + cos x
The result is ill-conditioned because, as in the following plot, the function has a singularity at
x = π.
2000
1000
0
3.08 3.1 3.12 3.14 3.16 3.18 3.2
-1000
-2000
4.13 Addition and subtraction:
f (u , v) = u + v
∂f ~ ∂f ~
∆f = ∆u + ∆v
∂u ∂v
∂f ∂f
=1 =1
∂u ∂v
f (u~, v~ ) = ∆u~ + ∆v~
Multiplication:
11
f (u , v) = u ⋅ v
∂f ∂f
=v =u
∂u ∂v
f (u~, v~ ) = v~ ∆u~ + u~ ∆v~
Division:
f (u , v) = u / v
∂f 1 ∂f u
= = 2
∂u v ∂v v
1 u
f (u~, v~ ) = ∆u~ + 2 ∆v~
v v
v ∆u~ + u ∆v~
f (u~, v~ ) =
v2
4.14
f ( x) = ax 2 + bx + c
f ' ( x) = 2ax + b
f " ( x) = 2a
Substitute these relationships into Eq. (4.4),
2a 2
ax i2+1 + bx i +1 + c = axi2 + bx i + c + (2ax i + b)( xi +1 − xi ) + ( xi +1 − 2 xi +1 xi + xi2 )
2!
Collect terms
ax i2+1 + bx i +1 + c = axi2 + 2ax i ( xi +1 − xi ) + a( xi2+1 − 2 xi +1 xi + xi2 ) + bx i + b( xi +1 − xi ) + c
ax i2+1 + bx i +1 + c = axi2 + 2ax i xi +1 − 2ax i2 + ax i2+1 − 2ax i +1 xi + ax i2 + bx i + bx i +1 − bx i + c
ax i2+1 + bx i +1 + c = (ax i2 − 2axi2 + ax i2 ) + ax i2+1 + ( 2ax i xi +1 − 2ax i +1 xi ) + (bx i − bxi ) + bxi +1 + c
ax i2+1 + bx i +1 + c = axi2+1 + bxi +1 + c
12
4.15 The first-order error analysis can be written as
∂Q ∂Q
∆Q = ∆n + ∆S
∂n ∂S
∂Q 1 ( BH ) 5 / 3 ∂Q 1 ( BH ) 5 / 3 1
=− 2 S 0.5 = −50.74 = = 2536.9
∂n n ( B + 2H ) 2 / 3 ∂S n ( B + 2 H ) 2/3
2S 0.5
∆Q = − 50.74 0.003 + 2536.9 0.00003 = 0.152 + 0.076 = 0.228
The error from the roughness is about 2 times the error caused by the uncertainty in the slope.
Thus, improving the precision of the roughness measurement would be the best strategy.
4.16 Use the stopping criterion
ε s = 0.5 × 10 2 − 2 % = 0.5%
True value: 1/(1 – 0.1) = 1.111111….
zero order:
1
=1
1− x
1.11111 − 1
εt = × 100% = 10%
1.11111
first order:
1
= 1 + 0.1 = 1.1
1− x
1.1 − 1
ε t = 1% εa = × 100% = 9.0909%
1.1
second order:
1
= 1 + 0.1 + 0.01 = 1.11
1− x
1.11 − 1.1
ε t = 0.1% εa = × 100% = 0.9009009%
1.11
third order:
13
1
= 1 + 0.1 + 0.01 + 0.001 = 1.111
1− x
1.111 − 1.11
ε t = 0.01% εa = × 100% = 0.090009%
1.111
4.17
d sin φ 0
∆ ( sin φ 0 ) = ∆α
dα
d sin φ 0 −β αβ
= + 1−
dα 2 (1 + α )(1 + α − αβ ) 1+α
where β = (ve/v0)2 = 4 and α = 0.25 to give,
d sin φ 0 −4 0.25(4)
= + 1− = −3.1305
dα 2 (1 + 0.25)(1 + 0.25 − 0.25( 4)) 1 + 0.25
∆ ( sin φ 0 ) = 3.1305∆α
For ∆α = 0.25(0.02) = 0.005,
∆ ( sin φ 0 ) = 3.1305(0.005) = 0.015652
0.25
sin φ 0 = (1 + 0.25) 1 − 4 = 0.559017
1 + 0.25
Therefore,
max sin φ 0 = 0.559017 + 0.015652 = 0.574669
min sin φ 0 = 0.559017 − 0.015652 = 0.543365
180
max φ 0 = arcsin(0.574669) × = 35.076°
π
180
min φ 0 = arcsin(0.543365) × = 32.913°
π
4.18 The derivatives can be computed as
14
f ( x) = x − 1 − 0.5 sin x
f ' ( x) = 1 − 0.5 cos x
f " ( x) = 0.5 sin x
( 3)
f ( x) = 0.5 cos x
( 4)
f ( x) = −0.5 sin x
The first through fourth-order Taylor series expansions can be computed based on Eq. 4.5 as
First-order:
f 1 ( x) = f ( a) + f ' ( a)( x − a)
π π  π  π
f1 ( x) = − 1 − 0.5 sin + 1 − 0.5 cos  x −  = x − 1.5
2 2  2  2
Second-order:
f " (a)
f 2 ( x) = f 1 ( x) + ( x − a) 2
2
f 2 ( x) = x − 1.5 + 0.25 sin(π / 2)( x − π / 2) 2
Third-order:
( 3)
f (a)
f 3 ( x ) = f 2 ( x) + ( x − a) 3
6
0.5 cos(π / 2)
f 3 ( x) = x − 1.5 + 0.25 sin(π / 2)( x − π / 2) 2 + ( x − a) 3
6
f 3 ( x) = x − 1.5 + 0.25 sin(π / 2)( x − π / 2) 2
Fourth-order:
( 4)
f (a)
f 4 ( x ) = f 3 ( x) + ( x − a) 4
24
0.5 sin(π / 2)
f 4 ( x) = x − 1.5 + 0.25 sin(π / 2)( x − π / 2) 2 − ( x − π / 2) 4
24
15
1
f 4 ( x) = x − 1.5 + 0.25 sin(π / 2)( x − π / 2) 2 − ( x − π / 2) 4
48
Note the 2nd and 3rd Order Taylor Series functions are the same. The following MATLAB
script file which implements and plots each of the functions indicates that the 4th-order
expansion satisfies the problem requirements.
x=0:0.001:3.2;
f=x-1-0.5*sin(x);
subplot(2,2,1);
plot(x,f);grid;title('f(x)=x-1-0.5*sin(x)');hold on
f1=x-1.5;
e1=abs(f-f1); %Calculates the absolute value of the
difference/error
subplot(2,2,2);
plot(x,e1);grid;title('1st Order Taylor Series Error');
f2=x-1.5+0.25.*((x-0.5*pi).^2);
e2=abs(f-f2);
subplot(2,2,3);
plot(x,e2);grid;title('2nd/3rd Order Taylor Series Error');
f4=x-1.5+0.25.*((x-0.5*pi).^2)-(1/48)*((x-0.5*pi).^4);
e4=abs(f4-f);
subplot(2,2,4);
plot(x,e4);grid;title('4th Order Taylor Series Error');hold off
4.19 Here are Excel worksheets and charts that have been set up to solve this problem:
16
12
8
exact
backward
4
centered
forward
0
-2 0 2
-4
17
12
8
exact
backward
4
centered
forward
0
-2 0 2
-4

For Engineers Ch4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

For Engineers Ch4

Uploaded by

Copyright:

Available Formats

1

f (0) = f ' (0) = f " (0) = e 0 = 1

for xi = 0.2, xi+1 = 1 and h = 0.8. True value = e–1 = 0.367879.

f (1) = e −0.2 = 0.818731

f (1) = 0.818731 − 0.818731(0.8) = 0.163746

4.2 Use the stopping criterion

True value: cos(π/3) = 0.5

4.3 Use the stopping criterion: ε s = 0.5 × 10 2−2 % = 0.5%

True value: sin(π/3) = 0.866025…

4.4 True value: f(3) = 554.

f (3) = f (1) = −62

f (3) = −62 + f ' (1)(3 − 1) = −62 + 70( 2) = 78 ε t = 85.921%

f " (1) 138

4.5 True value: f(2.5) = ln(2.5) = 0.916291...

f (2.5) = f (1) + f ' (1)(2.5 − 1) = 0 + 1(1.5) = 1.5

4.6 True value:

f ' (2) = 75(2) 2 − 12(2) + 7 = 283

xi–1 = 1.8 f(xi–1) = 50.96

f " ( 2) = 150 x − 12 = 150( 2) − 12 = 288

forward: E t ≈ 283 − 312.8 = 29.8

The centered approximation has an error that can be approximated by,

4.7 True value:

f " ( 2) = 150( 2) − 12 = 288

f ( 2.25) − 2 f (2) + f (1.75) 182.1406 − 2(102) + 39.85938

f ( 2.125) − 2 f (2) + f (1.875) 139.6738 − 2(102) + 68.82617

Thus, the range of ± 2.0844 is close to the first-order estimate.

4.10 For ∆T~ =20,

H (670) − H (630) 1542.468 − 1205.81

Thus, the first-order approximation is close to the exact result.

For ∆T~ =40,

H (690) − H (610) 1735.055 − 1059.83

4.11 For a sphere, A = 4πr2. Therefore,

At the mean values of the parameters,

H (0.15,0.9,550) = 4π (0.15) 2 0.90(5.67 × 10 −8 )(550) 4 = 1320.288

∆H = 17603.84(0.01) + 1466.987(0.05) + 9.6021( 20) = 441.4297

To check this result, we can compute

H (0.14,0.85,530) = 4π (0.14) 2 0.85(5.67 × 10 −8 )(530) 4 = 936.6372

H (0.16,0.95,570) = 4π (0.16) 2 0.95(5.67 × 10 −8 )(570) 4 = 1829.178

4.12 The condition number is computed as

The result is ill-conditioned because the derivative is large near x = 1.

10(−e −10 ) 10(−4.54 × 10 −5 )

The result is ill-conditioned because x is large.

The result is well-conditioned.

The result is well-conditioned.

(1 + cos x) cos x + sin x(sin x)

4.13 Addition and subtraction:

f (u~, v~ ) = ∆u~ + ∆v~

f (u~, v~ ) = v~ ∆u~ + u~ ∆v~

Substitute these relationships into Eq. (4.4),

ax i2+1 + bx i +1 + c = axi2 + 2ax i ( xi +1 − xi ) + a( xi2+1 − 2 xi +1 xi + xi2 ) + bx i + b( xi +1 − xi ) + c

ax i2+1 + bx i +1 + c = axi2 + 2ax i xi +1 − 2ax i2 + ax i2+1 − 2ax i +1 xi + ax i2 + bx i + bx i +1 − bx i + c

ax i2+1 + bx i +1 + c = (ax i2 − 2axi2 + ax i2 ) + ax i2+1 + ( 2ax i xi +1 − 2ax i +1 xi ) + (bx i − bxi ) + bxi +1 + c

ax i2+1 + bx i +1 + c = axi2+1 + bxi +1 + c

4.15 The first-order error analysis can be written as

∆Q = − 50.74 0.003 + 2536.9 0.00003 = 0.152 + 0.076 = 0.228

4.16 Use the stopping criterion

True value: 1/(1 – 0.1) = 1.111111….

where β = (ve/v0)2 = 4 and α = 0.25 to give,

For ∆α = 0.25(0.02) = 0.005,

∆ ( sin φ 0 ) = 3.1305(0.005) = 0.015652

max sin φ 0 = 0.559017 + 0.015652 = 0.574669

min sin φ 0 = 0.559017 − 0.015652 = 0.543365

4.18 The derivatives can be computed as