Quasi-periodic non-stationary solutions

of 3D Euler equations for incompressible flow

Sergey V. Ershkov,
Institute for Time Nature Explorations,
M.V. Lomonosov's Moscow State University,
Leninskie gory, 1-12, Moscow 119991, Russia
e-mail: sergej-ershkov@yandex.ru

A novel derivation of non-stationary solutions of 3D Euler equations for incompressible

inviscid flow is considered here. Such a solution is the product of 2 separated parts: -
one consisting of the spatial component and the other being related to the time
dependent part.
Spatial part of a solution could be determined if we substitute such a solution to the
equations of motion (equation of momentum) with the requirement of scale-similarity
in regard to the proper component of spatial velocity. So, the time-dependent part of
equations of momentum should depend on the time-parameter only.
The main result, which should be outlined, is that the governing (time-dependent)
ODE-system consist of 2 Riccati-type equations in regard to each other, which has no
solution in general case. But we obtain conditions when each component of time-
dependent part is proved to be determined by the proper elliptical integral in regard to
the time-parameter t, which is a generalization of the class of inverse periodic
functions.

Keywords: Euler equations, Riccati equations, non-stationary solutions.

1
1. Introduction: the Euler system of equations.

In accordance with [1-3], the Euler system of equations for incompressible flow of
inviscid fluid should be presented in the Cartesian coordinates as below (under the
proper initial conditions):


 u  0 , 1.1
 
u   p
 (u  ) u    F, 1.2
t 

- where u is the flow velocity, a vector field; ρ is the fluid density, p is the pressure, F
represents external force (per unit of mass in a volume) acting on the fluid; besides, we
assume external force F above to be the central force, which has a potential 
represented by F = - .

2. The originating system of PDE for Euler Eqs.

Using the identity (u∇)u = (1/2)∇(u2) – u×(∇×u), we could present the Euler equations

in the case of incompressible flow of inviscid fluid u = {u₁, u₂, u₃} as below [4-5]:


u  0 ,
( 2.1)

u   1  p 
 u      (u 2 )     
t 2  

- here we denote the curl field  = u, a pseudovector field (time-dependent) [6]:

  u u    u1  u 3    u 2  u1 
 ,  y ,  z     3  2  ,    ,     (2.2)
  y z   z x  x y
x
 

2
- also we denote  = {fx, fy, fz} in (2.1); besides, let us choose  = 1 for simplicity.

3. Conditions for the space-part of exact solution.

Let us search for solutions {u, p} of the system (2.1) in a form below:

u₁ = U(t)u(x,y,z), u₂ = V(t)v(x,y,z), u₃ = W(t)w(x,y,z), p = P(t)p(x,y,z) (3.1)

- then we should obtain from (2.1) and expression (2.2) the proper system of PDE:

  u1 1  
 (u 2   z  u 3   y )  (u 1  u 2  u 3 )  p fx ,
2 2 2
 t 2 x x



  u2 1  
  (u 3   x  u 1   z )  (u 1  u 2  u 3 ) 
2 2 2
p fy , 3.2
 t 2 y y


  u3 1  
 (u 1   y  u 2   x )  (u 1  u 2  u 3 )  p  fz ,
2 2 2
 t 2 z z

Besides, there exists the proper restriction from continuity equation (1.1) as below
(u/x ≠ 0):

u v w
U (t )  V (t )  W (t )  0, 
x y z
(3.3)
v u w u
 ,   , {, }  const
y x z x

3
 The system of equations (3.2) should be transformed under conditions (3.1) as below:

  v u   u w 
 V (t )  v ( x, y, z )  V (t )   U (t )    W (t )  w( x, y , z )  U (t )   W (t )  
 d U (t )   x y   z  x 

 dt u ( x, y , z )

    
 P(t )  p( x, y, z )  f x 
 1  x (U (t )  u ( x, y, z )  V (t )  v ( x, y, z )  W (t )  w ( x, y, z )) 
2 2 2 2 2 2
x
  ,
 2 u ( x, y , z ) u ( x, y , z )


  w v  v u 
 W (t )  w( x, y , z )  W (t )   V (t )    U (t )  u( x, y, z )  V (t )   U (t )  
 y z  x  y 
 d V (t )   3.4
 dt v ( x, y , z )

    
 P(t )  p( x, y, z )  f y 
 1  y (U (t )  u ( x, y, z )  V (t )  v ( x, y , z )  W (t )  w ( x, y , z )) 
2 2 2 2 2 2
y
  ,
 2 v ( x, y , z ) v ( x, y , z )


  u w   w v
 U (t )  u( x, y , z )  U (t )   W (t )    V (t )  v( x, y, z )  W (t )   V (t )  
 d W (t )   z x   y z

 dt w( x, y, z )

    
(U ( t ) 2
 u 2
( x , y , z )  V 2
( t )  v 2
( x , y , z )  W 2
( t )  w 2
( x , y , z ))  P(t )  p( x, y , z )  f z 
 1 z z
  ,
 2 w( x, y, z ) w( x, y, z )

- thus, from the 1-st of Eqs. (3.4) we should assume ({ai} = const, i =1,…,9):

v u u w
v ( x, y , z )  v ( x, y , z )  w( x, y, z )  w( x, y, z ) 
x y z x
 a1,   a2,   a 3,  a4,
u ( x, y , z ) u( x, y , z ) u ( x, y , z ) u( x, y , z )

 2  2  2
(u ( x, y, z )) (v ( x, y, z )) ( w ( x, y, z ))
1 x 1 x 1 x
  a5,   a6,   a7, (3.5)
2 u( x, y , z ) 2 u ( x, y , z ) 2 u( x, y , z )


p ( x, y , z )
x fx
  a8,   a9,
u( x, y , z ) u( x, y , z )

4
- but the 2-nd of Eqs. (3.4) yields as below ({bi} = const, i =1,…,9):

w v v u
w( x , y , z )  w( x , y , z )  u ( x, y , z )  u( x, y , z ) 
y z x y
 b1 ,   b2 ,   b3 ,  b4 ,
v ( x, y , z ) v ( x, y , z ) v ( x, y , z ) v ( x, y , z )

 2  2  2
(u ( x, y, z )) (v ( x, y, z )) ( w ( x, y, z ))
1 y 1 y 1 y
  b5 ,   b6 ,   b7 , (3.6)
2 v ( x, y , z ) 2 v ( x, y , z ) 2 v ( x, y , z )


p ( x, y , z )
y fy
  b8 ,   b9 ,
v ( x, y , z ) v ( x, y , z )

- besides, 3-d of Eqs. (3.4) yields ({ci} = const, i =1,…,9):

u w w v
u ( x, y , z )  u( x, y , z )  v ( x, y , z )  v ( x, y , z ) 
z x y z
 c1,   c2,   c3,  c4,
w( x, y, z ) w( x, y, z ) w( x, y, z ) w( x , y , z )

 2  2  2
(u ( x, y, z )) (v ( x, y, z )) ( w ( x, y, z ))
1 z 1 z 1 z
  c5,   c6,   c7, (3.7)
2 w( x, y, z ) 2 w( x , y , z ) 2 w( x , y , z )


p ( x, y , z )
z fz
  c8,   c9 .
w( x , y , z ) w( x, y, z )

4. The space-part of exact solution.

As for the structure of space part of exact solution (3.1), the system of equations (3.5)-
(3.7) could be solved by the proper analytical way as below:

5
 - Eq. (3.5) yields

v u
v ( x, y , z )  w( x, y, z ) 
x  u  a 2   v z w  a4  u
 a1,   ,   a 3,     ,
u( x, y , z )  y  a 1   x u( x, y , z ) x a  z
 3
(4.1)
 
(u ( x, y, z ))   a 5 a 6   a 1 , a 7   a 4 , p ( x, y, z )  a 8  u ( x, y, z ), f x   a 9  u( x, y, z ),
x x

- Eq. (3.6) yields

w  v   v 
w( x , y , z )   u ( x, y , z )    v ( x, y , z )  
y  w  b1   v  x   x 
 b1 ,   ,    b3  a 1 , 
v ( x, y , z ) y b  z  v ( x, y , z )   u ( x, y , z ) 
 
2
  
  
v  b3 
 (  b 3  a 1 ) ,  u ( x, y , z )      v( x, y, z ),  b  a ,
x  ( b3  a 1 )  3 5
 
(4.2)
u  b4   v  a 2   b4  
     ,       , b5   b 4 , (v ( x, y, z ))   b 6 , b 7   b1 ,
y b  x a  b   y
 3    
1 3

   b3 
p ( x, y, z )   b8  v( x, y, z ), p ( x, y , z )   a 8     v ( x, y, z ), f   b  v( x, y, z ),
y x  ( b3  a 1 )  y 9
 

- and Eq. (3.7) yields

 u   u 
 u( x, y, z )    w( x, y, z )   ( a 3  c1  b3  a 1 )
 z   z  u v
   a 3 c1 ,   ( a 3  c1 ) ,   ,
 w( x, y, z )   u( x, y, z )  z z b3
  
  
     
w( x, y, z )   
a3   u ( x, y , z )   a3   b3   v( x, y, z ) ,  w    c 2    u , 
 ( a 3  c1 )   ( a 3  c1 )   ( b3  a 1 )   x  c 1   z
     
(4.3)
w
v ( x, y , z ) 
 a4   c2  y w  w  c 3   v  c 3   b1 
    ,   c3,   (  b1  c 3 ) ,   ,      , c 5   c 1 ,
a  c  w( x, y, z ) y  y  c 4   z c  b 
 3  1  4  2

   a 3 b3 
c6  c4, ( w ( x, y, z ))   c 7 , p ( x, y, z )   c 8  w( x, y, z )   c 8     v( x, y, z ), f z   c 9  w( x, y, z ) .
z z  a 1 c1 
 
6
So, the space part of the solution should be presented as below:

 b3 
u     v ( x, y , z )   b  x  b  b  y  (  a  c )  z
 ( b3  a 1 )  3 3 6 3 1
 

(  a 3  c1 )
v  x  b6  y   z, (4.4)
b3

 a 3  b3  a  a b 
w     v ( x, y , z )  (  3 )  b 3  x  b 6   3 3   y  a 3  z ,
 a 1  c1  c1  a 1  c1 
   

c8
(a 1  b 3 )   1, a 2  b 6 , a 3  b 3  a 3  b 3 , a 3  c 7 , a 5  b 3 , a 8  b 3  b 6  b8 , a 8  ,
a 1  c1

 b3  b 3  (b 6 ) 2  2 c 1  a 3 2 
p ( x, y , z )  a 8    b 3  b 6  x  y  (  a 3  c 1 )  x  z  b 6 (  a 3  c 1 )  y  z  x 2    y  z
 2  2  2 b 
   3


Thus, if we choose for simplicity the proper constants as

c1 = -1, a8 = 1 ( b8 = b3b6, c8 = 1/b3 ) (4.5)

- the space part of the solution should be presented as below ( a3b3 = a3b3 ):

u   b3  x  b3  b6  y  a 3  z

a3
v  x  b6  y   z, (4.6)
b3

w  a 3  b3  x  b6  b3  a 3  y  a 3  z ,

b3  b 3  (b 6 ) 2  2 a 3 2
p ( x, y , z )   b 3  b 6  x  y  a 3  x  z  b 6 a3  y  z  x  2  y  z
2  2  2b 3
 

7
5. Time-dependent part of exact solution.

As for the structure of time-dependent part of exact solution (3.1) with space part (4.6),
it could be obtained from system of Eqs. (3.4) which should be transformed as below:

 d U (t )
 d t  V (t )  U (t )  a 2  W (t )  U (t )  a 3  U (t )  a 5  P (t )  a 8  a 9 ,
2




 d V (t )
  W (t ) V (t )  b 2  U (t ) V (t )  b 3  V 2 (t )  b 6  P (t )  b 8  b 9 , (5.1)
 dt


 d W (t )  U (t )  W (t )  c  V (t )  W (t )  c  W 2 (t )  c  P (t )  c  c ,
 d t 2 3 7 8 9

- where a2 = b6, a3b3 = a3b3, a5 = b3, a8 = 1, b8 = b3b6, c7 = a3, c8 = 1/b3; besides, it

should be accomplished along with the continuity equation (3.3):

u v w
U (t )  V (t )  W (t )  0 ,  U (t )  b 3  V (t )  b 6  W (t )  a 3  0 ,
x y z

- so, we have 4 equations for the obtaining of 4 functions U(t), V(t), W(t), P(t).

Along with the invariant from the continuity equation (besides, a9b3 + b9b6 + c9a3 = 0)

 
d  U (t )  b 3  V (t )  b 6  W (t )  a 3 
   0,
dt

- a system of Eqs. (5.1) immediately yields the invariant for function P(t) as below

P(t )  a 8  b 3  b8  b 6  c 8  a 3    [V (t ) U (t )  a 2  W (t ) U (t )  a 3  U (t ) 2  a 5 ]  b 3 
(5.2)
 [W (t ) V (t )  b 2  U (t ) V (t )  b 3 V 2 (t )  b 6 ]  b 6  [U (t ) W (t )  c 2  V (t ) W (t )  c 3  W 2 (t )  c 7 ]  a 3 ,

8
- thus, we should exclude expression (5.2) for function P(t) from the analysis of
equations of system (5.1) and also we should exclude the continuity equation, which
means that one of 3 functions U(t), V(t), W(t) is the linear combination of two others:

b3 b6
W (t )   U (t )   V (t )  (5.3)
a3 a3

So, analyzing the system (5.1), we finally should obtain the system of 2 ordinary
differential equations of the 1-st order for any 2 of 3 functions U(t), V(t), W(t) (the last
3-rd function could be obtained by expressing it from the continuity equation above).

These governing ODE-equations form together a system of 2 Riccati-type equations in

regard to each other, which is the system of 2 ordinary differential equations of the 1-st
kind with the right parts, consisting of polynomials of the 2-nd extent in regard to the
functions U(t), V(t), W(t).
Riccati type of equations has no analytical solution in general case [6]. We should note
also that a modern methods exist for obtaining of the solution of Riccati equations with
a good approximation [7-9]. But if we choose a proper constants for the system (5.1):

a9 = b9 = c9 = 0, (a8b3 + b8b6 + c8a3) = 1, 

(b3 + b3(b6)² + a3/b3) = 1,  a3 = b3 - b3 b3 - b3b3(b6)², (5.4)

- the 1-st and 2-nd equations of system (5.1) could be transformed as presented below:

   b b  
 d U (t )  V (t )  U (t )   b 3   2 6  c 3  2 b 6   b 6  ( c 2  b 3 )  
 dt   a3  
   

  b2  b6 
  U (t ) 2  b 3  ( c 2  b 3 )  V 2 (t )  b 6    c 3  2b 6  ,
 a 
  3 

 (5.5)

 d V (t )  U (t ) V (t )   b  1  b 2   b  b  (c  b )  b  b   b 2  b 6  c  2 b   
 dt  3  a3  6 8 2 3 3 8  3 6

     a3 
   b2   b2  b6 
  V 2 (t )   b 6  1    b 6  b 8    c 3  2 b 6    U 2 (t )  b 3  b 8  ( c 2  b 3 ) ,
   a3   a 
     3 

9
- where a2 = b6, see (4.4). Besides, if we additionally choose b2 = a3, system (5.5) above
could be reduced to the simplified regular form below:

 d U (t )  
   b8 U (t )  C  U (t )  b 3  (c 3  b 6 )  b 6  (c 2  b 3 )  
 dt  
 2
  
 U (t )  b 3  (c 2  b 3 )   b8 U (t )  C   b 6  (c 3  b 6 ) ,
2
   (5.6)


 d V (t ) d U (t )
  b8   V (t )  b8 U (t )  C , C  const,
 dt dt

- where the 1-st equation of system (5.6) has a proper solution below (b8 = b3b6):

d U (t )
 dt, (5.7)
( A  U (t )  B  U (t )  D )
2

 
A   b8  b 3  (c 3  b 6 )  b8  b 6  (c 2  b 3 )  b 3  ( c 2  b 3 )  (b8 ) 2  b 6  (c 3  b 6 ) 
 
 
B  C   b 3  (c 3  b 6 )  b 6  (c 2  b 3 )  2b8  b 6  (c 3  b 6 )  , D  C 2  b 6  (c 3  b 6 )
 

The left side of expression (5.7) could be transformed to the proper elliptical integral
[10] in regard to function U(t):

 2  2 A  U (t )  B 
 arctan  , 0
d U (t )     
 ( A  U 2 (t )  B  U (t )  D ) 
   (4 A  D  B 2 ) (5.8)
 2  2 AU (t )  B 
  Arth  , 0
    

6. Discussion.

In fluid mechanics, a lot of authors have been executing their researches to obtain the
analytical solutions of Euler and Navier-Stokes equations [11], even for 3D case of

10
compressible gas flow [12]. But there is an essential deficiency of non-stationary
solutions indeed.
Our presentation (3.1) of the non-stationary solutions of 3D Euler equations (1.1)-(1.2)
for incompressible flow is considered here. The spatial part of such a solution is
determined by equalities (4.4), under the given initial conditions; but the time-
dependent part is determined by Eqs. (5.1)-(5.3).
Besides, the real example of exact solution is obtained. The spatial part of such a
solution is presented by the equalities (4.5)-(4.6), but the time-dependent part is
presented by equalities (5.2)-(5.3), (5.4)-(5.8).
Also, we should especially note that the components of flow velocity (4.6) of the

solution (3.1) will be uniformly increasing when (x, y, z) → ∞. So, such a solution

should be defined within the limited domain of the meanings of variables (x, y, z), it
should be given by the initial conditions.
The explicit solutions for U(t) can easily be obtained from Eqs. (5.7) and (5.6),
therefore, the explicit form of the special solutions (3.1) should be provided:

u₁ = U(t)u(x,y,z), u₂ = V(t)v(x,y,z), u₃ = W(t)w(x,y,z), p = P(t)p(x,y,z),

- where

u   b3  x  b3  b6  y  a 3  z

a3
v  x  b6  y   z,
b3

w  a 3  b3  x  b6  b3  a 3  y  a 3  z ,

b3  b 3  (b 6 ) 2  2 a 3 2
p ( x, y , z )   b 3  b 6  x  y  a 3  x  z  b 6 a 3  y  z  x2   y  z ,
2  2  2 b
  3

P(t )   [V (t ) U (t )  b 6  W (t ) U (t )  a 3  U (t ) 2  b 3 ]  b 3 

 [W (t ) V (t )  a 3  U (t ) V (t )  b 3 V 2 (t )  b 6 ]  b 6  [U (t ) W (t )  c 2  V (t ) W (t )  c 3  W 2 (t )  a 3 ]  a 3 ,

11
 b3 b6
W (t )   U (t )   V (t )  , V (t )  (b3  b6 ) U (t )  C , C  const ,
a3 a3

- here we should choose a3 = {b3 - b3 b3 - b3b3(b6)²}  0, → b3 (1+ (b6)²)  1, but

the key function U(t) should be given as below (b8 = b3b6):

    
   tan     t   B
  2 
U ( t )     , 0
 2A

   (4 A  D  B 2 )
     
    tanh    t  B
 
  2
    
, 0
 U ( t )
 2 A

 
A   b 8  b 3  (c 3  b 6 )  b 8  b 6  (c 2  b 3 )  b 3  (c 2  b 3 )  (b 8 ) 2  b 6  (c 3  b 6 ) 
 
 
B  C   b 3  (c 3  b 6 )  b 6  (c 2  b 3 )  2b 8  b 6  (c 3  b 6 )  , D  C 2  b 6  (c 3  b 6 )
 

For example, if we choose c3 =b6 (C  0) it should simplify the expression for U(t):

    
   tanh    t  B
 
 2
U (t )     
,    B2
2A

A  b 3  (c 2  b 3 )  1  (b 6 ) 2  , B  C  b 6  (c 2  b 3 ) , D  0

- and, if we additionally choose c2 = 2b3, b3 = 2/C (C  0), b6 = 1, we should obtain (see

Fig.1-5)

C2  
U (t )    tanh ( t )  1  ,    4, A  (b3 ) 2  2 , B  2 , D  0
8  

12
Fig.1. A schematic plot of the function ~ (x-y+1)*{tanh(-t) - 1},
here we designate: x  (-50, 50), t = y  (0, 25).

Fig.2. A schematic plot of the function ~ (x-y+1)*{tanh(-t) - 1},

here we designate: x  (-1, 1), t = y  (0, 1).

Fig.3. A schematic plot of the function ~ (x-y-10)*{tanh(-t) - 1},

here we designate: x  (-1, 1), t = y  (0, 1).

13
 Fig.4. A schematic plot of the function ~ (x-y+10)*{tanh(-t) - 1},
here we designate: x  (-1, 1), t = y  (0, 1).

Fig.5. A schematic plot of the function ~ (x-y+30)*{tanh(-t) - 1},

here we designate: x  (-1, 1), t = y  (0, 1).

We assume at Fig.1-5 that in the expressions (x-y-10), (x-y+1), (x-y+10), (x-y+30) set of
meanings {-10, 1, 10, 30} is varying according to the varying of the range of variable z;
besides, the factor {tanh(-t) - 1} could be schematically presented (for imagination of
the plots of solutions) by the changing of parameter t to variable y, for example.

At Fig. 6-8 we schematically imagined solutions for the case >0:

14
Fig.6. A schematic plot of the function ~ (x-y-10)*{tan(t) - 1},
here we designate: x  (-1, 1), t = y  (0, 1).

Fig.7. A schematic plot of the function ~ (x-y+1)*{tan(t) - 1},

here we designate: x  (-1, 1), t = y  (0, 1).

Fig.8. A schematic plot of the function ~ (x-y+10)*{tan(t) - 1},

here we designate: x  (-1, 1), t = y  (0, 1).

15
Also, we should note that since some solutions are unbounded (see for instance Eq.
(5.8) for >0), such a solution should be defined within the limited range of the
meanings of time-parameter t (it should be given by the initial conditions).

Besides, we should additionally note that the only periodic (and unbounded) solutions
are the ones given by U(t) for  > 0, since the hyperbolic tangent is a non periodic but
bounded function.

7. Conclusion.

A new presentation of non-stationary solutions of 3D Euler equations for

incompressible inviscid flow is considered here. Such a solution is the product of 2
separated parts: - spatial and the time-dependent parts.
Spatial part of a solution could be determined if we substitute such a solution to the
equations of motion (equation of momentum) under the demand of scale-similarity in
regard to the proper component of spatial velocity. So, the time-dependent part of
equations of momentum should depend on the time-parameter only.
The main result, which should be outlined, is that the governing (time-dependent)
ODE-system consist of 2 Riccati-type equations in regard to each other, which has no
solution in general case. But we obtain conditions when each component of time-
dependent part is proved to be determined by the proper elliptical integral in regard to
the time-parameter t, which is a generalization of the class of inverse periodic
functions. Thus, by the proper obtaining of re-inverse dependence of a solution from
time-parameter we could present the expression for each component of motion as a set
of periodic cycles.

16
References:

[1]. Landau, L.D.; Lifshitz, E.M. (1987), Fluid mechanics, Course of Theoretical
Physics 6 (2nd revised ed.), Pergamon Press, ISBN 0-08-033932-8.
[2]. Ladyzhenskaya, O.A. (1969), The Mathematical Theory of viscous
Incompressible Flow (2nd ed.), Gordon and Breach, New York.
[3]. Lighthill, M. J. (1986), An Informal Introduction to Theoretical Fluid Mechanics,
Oxford University Press, ISBN 0-19-853630-5.
[4]. Saffman, P. G. (1995), Vortex Dynamics, Cambridge University Press.
[5]. Milne-Thomson, L.M. (1950), Theoretical hydrodynamics, Macmillan.
[6]. Kamke E. (1971), Hand-book for Ordinary Differential Eq. Moscow: Science.
[7]. Carl M. Bender, Steven A. Orszag (1999), Advanced Mathematical Methods for
Scientists and Engineers. Origin.published by McGraw Hill, 1978, XIV. pp.20-22.
[8]. Rosu H.C., Cornejo-Pérez O., and Ojeda-May P. (2012), Traveling kinks in cubic
nonlinear Ginzburg-Landau equations, Phys. Rev. E 85, 037102.
[9]. Christianto V., Smarandache F. (2008), An Exact Mapping from Navier-Stokes
Equation to Schroedinger Equation, Progress in Physics, 1, p.38-39.
[10]. Lawden D. (1989) Elliptic Functions and Applications. Springer-Verlag.
See also: http://mathworld.wolfram.com/EllipticIntegral.html
[11]. Drazin, P.G. and Riley N. (2006), The Navier-Stokes Equations: A Classification
of Flows and Exact Solutions, Cambridge, Cambridge University Press.
[12]. Ershkov S.V., Schennikov V.V. (2001), Self-Similar Solutions to the Complete
System of Navier-Stokes Equations for Axially Symmetric Swirling Viscous
Compressible Gas Flow, Comput. Math. and Math. Phys. J., 41(7), p.1117-1124.

17