NUMERICAL

SOLUTION USING
NEWTON-RAPHSON
METHODS
Homework 5:
• Write iterative Solver for “Rayleigh Pitot Equation”

    
   1 2
  1 
  M  
P02   2 
  1  
p  

 2
M 
2   1    1

      
  1 1  
• Use solver for Pitot problem presented on slides 53-55

2
 Homework 5:
• Key Piece of Knowledge

      1 
    1 2    1     1     1

  
 2
M 
   M 
2M 2
 1  
M 2

2 
   
M   1 
   

 2 M 2    1
  1   1
    2  1 
   1 M    1
2
    1   1   
 

3
 Numerical Solution for Mach
• Graphical Solutions are good for “sanity check” but really
Need automated solver to allow for iterative design, trade
studies, sensitivity analyses, etc.

• Use “Newton’s Method” to extract numerical solution

 1

F(M )
• Define: 
1  2  
1
  1 2  
M 
2  1
A
 *
  
M    1  2  A
• At correct Mach number (for given A/A*) …

F(M )  0
4
 Numerical Solution for Mach
(cont’d)

• Expand F(M) is Taylor’s series about some arbitrary

Mach number M(j)

 2 F 
 
2
 M 2  M  M ( j )
 F 
   
( j) 3
F(M )  F(M ( j ) )    M  M( j)   ...O M  M ( j )
 M  ( j ) 2

• Solve for M
  2 F  
 
2
 2
M  M( j) 
 M  ( j )

  
3
F(M )  F(M ( j ) )   ...O M  M ( j )
 2 
 
 
M  M( j) 
 F 
 
M  ( j )
5
 Numerical Solution for Mach
(cont’d)

• From Earlier Definition F(M ) ,thus0

  2 F  
 
2
 2
M  M( j) 
 M  ( j )

  
3
F(M ( j ) )   ...O M  M ( j )
 2 
 
 
M  M( j) 
 F 
 
M  ( j )
Still exact expression

   M  M 
2
• if M(j) is chosen to be “close” to M M  M( j) ( j)

And we can truncate after the first order terms with “little”
Loss of accuracy

6
 Numerical Solution for Mach
(cont’d)

• First Order approximation of solution for M

^ F(M ( j ) )
M  M( j) 
 F 
 
M ( j )
“Hat” indicates that solution is no longer exact

^
• However; one would anticipate that M  M  M  M( j)

“estimate is closer than original guess”

7
 Numerical Solution for Mach
(cont’d)
^
M
• If we substitute back into the approximate expression

^ ^
^ F(M )
^
M  M
 F 
  ^
M |M
^
^ ^
• And we would anticipate that MM  MM

“refined estimate” …. Iteration 1

8
 Numerical Solution for Mach
• Abstracting to a “jth” iteration
^
^ ^ F(M ( j ) )
M ( j 1)  M ( j)  Iterate until convergence
 F  j={0,1,….}
 
M |( j )
• Drop from loop when

 1
1  2     1 ^ 2
 2  1
A
   1 M ( j 1)    *
M ( j 1)    1  
^
2  A

A
A*
9
 Numerical Solution for Mach
(cont’d)
 1
^ 1  2     1 ^ 2   2  1
A
F(M ( j ) )  ^    1 M ( j)    *
M ( j )    1  2  A
 1
 
 F 
 ^
  1  2  
1 
  1 ^
M
2
 2  1

      1  ( j)
 
M  |( j )  ^
 M ( j)  M ( j)  2  

  1 
   
2  2  1 
^
M  1    1 ^ 2
   
  1 3      1  2 M ( j) 
( j )
 2  2 
2   
  ^ 2
 ^ 2
  1
M ( j )  2  M ( j ) (  1)   
  

10