Background Image
Previous Page  13 / 21 Next Page
Information
Show Menu
Previous Page 13 / 21 Next Page
Page Background

This results in the numerical method of solving the quasi-one-dimensi-

onal equations of gas-dynamics (physically specified well), underlaid by

the predictor-corrector method and a variant of the non-linear quasi-

monotonous compact difference scheme of the higher order of accuracy.

In this case, for the “predictor” stage, the system of Euler’s quasi-one-

dimensional equations is used in a general non-divergent characteristic

form, in which the unknown values are written relative to the Riemann

quasiinvariants. At the “corrector” stage, the divergent form of Euler’s

quasi-one-dimensional equations is used.

Time increment

Δ

t

, necessary for integrating the difference scheme

given above, is selected from the conditions of satisfying the Courant-

Friedrichs-Lewy stability criterion.

To approximate the second derivatives that are included into the

“viscous” part of the equations set of the dynamic and thermal boundary

layer, we used the finite-difference presentation of

(

μf

)

0

z

the variables on

the mesh introduced earlier by the following formulae [7]:

(

μf

)

0

z

h

μ

i

1

f

0

i

1

+ 8

μ

i

+1

/

2

f

0

i

+1

/

2

μ

i

1

/

2

f

0

i

1

/

2

μ

i

+1

f

0

i

+1

i

;

μ

i

±

1

/

2

=

[

μ

i

±

2

+ 9 (

μ

i

+

μ

i

±

1

)

μ

i

1

]

16

;

f

0

i

±

1

/

2

[27 (

f

i

±

f

i

±

1

)

±

(

f

i

1

f

i

±

2

)]

12Δ

;

f

0

i

±

1

/

2

[

f

i

2

±

6

f

i

1

18

f

i

±

10

f

i

±

1

±

3

f

i

±

2

]

12Δ

.

Mixed derivatives

∂x

μ

∂f

∂y

i,j

have been approximated by the known

central differencies of the fourth order (or with the formulas given):

(

μf

)

0

z

[

f

i

2

+ 8 (

f

i

+1

f

i

1

)

f

i

+2

]

12Δ

. The approximation effectiveness

of using “viscous” members of the boundary layer of high order accuracy

operators is particularly noticeable in the tasks relating to the computation

of the aerodynamic shock tubes. Let us use the displacement operator

T

m

u

j

=

u

j

+

m

=

u

(

z

j

+

m

Δ)

and the symbol of unit operator

E

=

T

0

to

determine the derivatives

∂v

∂r

r

=

R

,

∂ζ

∂r

r

=

R

δ

at the boundary of the design

area:

u

0

|

r

=

R

=

1

60Δ

[

12

T

5

+ 75

T

4

200

T

3

+ 300

T

2

300

T

1

+ 137

E

] +

Δ

5

6

u

VI

;

ISSN 0236-3941. HERALD of the BMSTU. Series “Mechanical Engineering”. 2014. No. 1 15