You are on page 1of 21

Chapter 4

Systems of Non-linear Equations

PREPARED BY:
ENGR. ROMANO A. GABRILLO
MENGG-MFG.E

Introduction
In this chapter, a solution technique for solving a

system of non-linear equations will be discussed


namely the Newton-Raphson Method.
In the real world no system behaves in a linear

manner. There is no ideal material, ideal support


condition and a perfect structure. Due to
imperfections the behavior can always be described
by a set of non-linear equations only.

Newton-Raphson Method
Suppose there are n unknowns:

{x}T = x1 x2 xn
is a solution for non-linear equations

f1 (x1, x2,
f2 (x1, x2,

fn (x1, x2,

... xn)=0
... xn)=0

... xn)=0

If x approximates x bar, then the increment from x to


x bar be denoted by:
x1 x1 x1
x x x
2
2
2

{x} {x x} x3 x3 x3
... ... ...


xn xn xn
Where xj is the increment from xj to xj for j=1, 2,

-n
It is now required to find vector x, and to find the
direction and distance to more from x (in n
space), to get the desired point: x bar = x + x

It is necessary to seek the exact increment x


that satisfies:
f i ( x x) 0 for i 1, 2, .... .... n.
Expanding this in the Taylors series and neglecting

higher order terms, we get:


f
f
f
f1 ( x x) f1 ( x)
dx1
dx2 ...
dxn 0
x1
x2
xn
f
f
f
f 2 ( x x) f 2 ( x)
dx1
dx2 ...
dxn 0
x1
x2
xn

f
f
f
f n ( x x) f n ( x)
dx1
dx2 ...
dxn 0
x1
x2
xn

Where partial derivatives are evaluated at the current values of x=x bar.
This linear system can be expressed in matrix form as:

f1 ( x)
x
1

...
...
f n ( x)

x1

f1 ( x)
x2
...

...
...

...
...
f n ( x)
...
x2

f1 ( x)
xn

...
...
f n ( x)

xn

x1k 1
k 1
x2

...
xnk 1
xk

... x1k
k
... x2

... ...
... xnk

f1 ( x )

f 2 ( x)
f ( x)
n x

or
x1

x2
x
n

k 1

x1
f1 ( x)

1
x2 [k ] xk f 2 ( x)
x
f ( x)
n
n x

f1

x1

where [k ]

f
n
x
1

f1
x2

f n
x2

f1

xn

f n
xn x k

This method based on this formula is called


Newton-Raphson Method
We have seen this method for single degree of

freedom system as:

f ( xk )
xk 1 xk
f ' ( xk )
There

is a number but a non-linear set of


equations. It is a matrix [k]-1 . The above
equation is written for non-linear system of
equations as:
1
f ' ( xk )

{x}k 1 {x}k [k ]xk1{ f ( x)}xk

Example No. 1
Consider a non-linear spring problem as shown

below. The springs used are non-linear whose force


deflection characteristics are given by:
Spring 1 F1 = 598x1 + 6060x13
Spring 2 F2 = 657x2 + 919x23
Spring 3 F3 = 69x3 + 196x33

The loads are P1 = -120, P2 = 398

It is required to find the deflection of the spring


x1 = deflection in spring 1 = U1
x2 = deflection in spring 2 = (U3 U1)
x3 = deflection in spring 3 = U3
The equilibrium equations are

F 1 F 2 = P1
F2 + F 3 = P2
F1 = 598 U1 + 6060 U13
F2 = 657 (U3 U1) + 919 (U3 U1)3
F3 = 69 U3 + 196 U33

Substituting:
1255 U1 + 6060 U13 657 U3 919 (U3 U1)3 = -120
-657 U1 + 919 (U3 U1)3 + 726U3 + 196U33 = 398
If U1 and U3 are not exact we will get residue as:
f1(x) = r1 = 120 + 1255U1 + 6060U13 - 657U3 - 919(U3 U1)3
f2(x) = r2 = -398 - 657U1 + 919(U3 U1)3 + 726U3 + 196U33

f1
The matrix [k] =
U1
f 2
U
1

f1

U 3 A B


f 2 C D
U 3

where A = 1255 + 18180U12 + 2757(U3 U1)2


B = -657 2757(U3 U1)2 = C
D = 2757(U3 U1) + 726 + 588U32

U1

U 3

k 1

f ( x)
U1
1 1
[k ]

f 2 ( x)k
U 3

First assume U1 = U3 = 0.3, k=0

U1 0.3 2.891.2 657


U 3 0.3 657 778.92 0
0

463

3720

Take the Inverse of matrix [k]


1

2891.2 657 1 0
2.891.2 657


657 778.92 0 1
657 778.92 0

E10 2891.2 657 1 0


E20 657 778.92 0 1
E11 E10 / 2891.2 1 0.227241283 0.000345877 0
629.62
0.0227241189 1
E21 E20 657E11 0

At k=0, and substituting the [k]-1


E12 E11 0.227E22 1 0 0.000345877 0.00036119
E22 E21 / 629.62 0 1 0.000360917 0.001588259
1

[k ]
1

0.000345877

0.00036119

0.000360917 0.001588259
1

U1 0.3 0.000345877 0.00036119


U 3 0.3 0.000360917 0.001588259 0

463

3720

Multiplying, and checking the =0.01


1

U1 0.3 0.0637

U 3 0.3 0.423720
1

U1 0.273

U 3 0.723
0.237 0.3
0.265 0.01
0.237

0.723 0.3
0.585 0.01
0.723

f1 ( x)
Substitute this again to find [k] &
we get
f 2 ( x)
2

1307.92
U1 0.237 2924



U 3 0.723 1307.93 1638.92 1
1

82.46

150
.
67

1307.92 1 0
2924

[k ]

1307
.
93
1638
.
92
0
1

1
1

1307.92 1 0
E10 2924

0
E2 1307.93 1638.92 0 11
E11 E10 / 2924 1 0.4473050610.000341997 0

1
0
1
E2 E2 1307.92E1 0 1098.82119 0.44730716 11

At k = 1
E12 E11 0.44E11 1 0 0.000524085 0.000407077

1
0
1
E2 E2 / 1098.82E1 0 1 0.000407079 0.0009100661
2

U1 0.237 0.000524085 0.000407077



U 3 0.723 0.000407079 0.0009100661
2

U1 0.237 0.018118242

U
0
.
723
0
.
103551909

1
3

82.46

150
.
67

U1 0.2192

U
0
.
6196

0.2192 0.237
0.0812 0.6196 0.723 0.16688
0.2192
0.6196

At k = 2
3

U1 0.2192 2570.52 1099



U 3 0.6196 1099 1393.74 2
2

2570.52 1099

[k ]
1099 1393.74 2
1

E 2570.52 1099

E 1099 1393.74 2
0
1
0
2

0
1
0
2

1 0

0
1

1 0

0
1

2
1

E E / 2570.52 1 0.427539953

1
E E 1099E1 0 923.8735917 2
1
1
1
2

7.14

13
.
43

0.000389026 0

0
.
427539574
1

At k=3
1

E12 E11 0.427E22 1 0 0.000586877 0.000462768

2
1
0
1
0
.
000462768
0
.
001082399
E2 E2 / 923.87
2
2
3

U1 0.2192 0.002024672

U
0
.
6196
0
.
01123455

2
3
3

U1 0.2174

U
0
.
60857

0.2174 0.2192
0.008
0.2174

Final Answer
Finally we get,

U1 = 0.2174
U3 = 0.6087
U2 = U3 U1 = 0.6087 0.2174=0.3913
Deflection in spring 1 = 0.2174
Deflection in spring 2 = 0.3913

Deflection in spring 3 = 0.6087

Assignment No. 7
Solve the following equations using the Newton-

Raphson Method:
equation 1:
equation 2:

x2 + 2y2 = 22
2x2 xy + 3y = 12

You might also like