Four methods to solving the three-moments equation in a

particular case
Aziz ATTA1*
1
Mathematics and Structural Analysis, Atta Engineering Design Office, El Jadida, Morocco
*Corresponding author: [email protected]

Abstract We will present in this article some simplified methods for the resolution of three-moments equation in
particular cases and frequently encountered in practice. We will thus state these methods and establish their
demonstrations, present the process for their implementation as well as an application on an example of a continuous
beam with five spans.

Keywords: Continuous beam, Tree-moments equation, Bending moments, Trinomial coefficients.

6 EI
1. Introduction i  ir  il  and  0  M1  2
l
Material strength is an important and indispensable
discipline for the design, calculation, sizing and 2. First method: Determinant method
verification of civil engineering structures. It is a science
which has been developed for a very long time by several The determinant method that we will present in this part
brilliant scientists in human history [1]. The three-moment is the simplest and practical method of the four methods
method (Clapeyron theorem) [2] is a well-suited method proposed. This method can be used in other fields: it is a
for solving continuous beams, established from the force general mathematical method to reverse the recurrent
method. It consists in cutting out a continuous beam into sequences of order 3 with second member. The three-
independent spans, and introducing moments on supports moments equation is written for the application of this
as hyperstatic unknowns. The rotation of each method of the following form:
intermediate support of the continuous beam is zero, this
is the condition of compatibility of the deformations. This q  1, n  1 M q  2  4M q 1  M q   q 1  3
equation gives relationship between three bending
moments at three consecutive supports. In this method, the Such that: M 1  M 0  M n1  0 . So, the only unknown
unknowns are redundant moments at the supports. We to reverse the three-moments equation is M1  0 . We
consider a continuous beam of n  1 spans subjected to
define the determinant sequence of the equation of three-
external loadings:
moments equation  3 by:
0  1 ; 1  4

 4 
 q  2  4 q 1   q  0


Figure 1. Continuous beam of n  1 spans subjected to external loadings 2.1. Determinant method statement
A span  i  is delimited by the two supports i  1 and The hyperstatic moments at the supports of a continuous
i .  ir is the rotation to the right of support i for the span beam of n  1  n  1 spans of the same length l are
 i  1 considered independent and il is the rotation to given by the following expressions:
the left of support i for span  i  considered independent.
 n 1
The equation of the three moments or the Clapeyron
theorem for the continuous beam described above is given   m nm
 M 1   0   m  0
by (1):  n  5
 q

li  l l  l q  1, n  1 M q 1     m  q  m
M i 1  2  i  i 1  M i  i 1 M i 1  6 ir  il  
 EI i 
  EI   EI  
i 1 
 EI i 1 m0

Where  m and  n is given by equations  2 and  4 .

i

We will treat a more practical particular case. In fact,

we often try to use continuous beams of the same material 2.2. Demonstration of Determinant method
 Ei =E  with spans of the same length  li =l  and constant First, we can demonstrate using the characteristic
inertia  Ii  I  . We note: &
equation of the sequence of   n n that:
  1
 
q 1 d) Calculate the hyperstatic moments
  
q 1 q 1
q  2 3  2 3 6 M q 1  q  1, n  1 
2 3
Indeed, the solutions of r 2  4r 1  0 are: r1  2  3 2.4. Expressions of forces in continuous beam
and r2  2  3 . We’ll get:
2.4.1. Bending moment in continuous beam
1 1      1
 n   r1n   r2n and       The bending moment [3] in the different spans of the
 r1 r2      4  continuous beam is given by:
And finally:
q  0, n  1
 1
 
q 1

r  r     
1 q 1 q 1
q 1 q 1
q  2 3  2 3  x x
2 3
2
2 3
1
M q 1  x   mq 1  x   M q 1    M q 1 7 
 l l
Let prove the expression of M q by induction. For
 x
q
0  mq 1  x      m 1     m  .  q  m
q  0,    m 0   0 0  0  M1 : Ok! . We suppose m0  l
m0
q Where mq 1 is the static moment in the span q  1 taken
that, for q : M q 1     m vq  m and independent and subjected to the external loads which it
m0
q 1
supports.
M q     m vq 1 m . For q  2 :
m0 2.4.2. Shear force in continuous beam
The shearing force [3] in the different spans of the
M q  2   q 1  4M q 1  M q continuous beam is given by:
q q 1
  n 1  4   m  q  m    m  q 1 m q  0, n  1
m0 m0
M q  M q 1
q q
Vq 1  x   vq 1  x    8
     q 1    4 m  q  m    m 1  q  m l
m0 m 1 q
q
  0  q 1     m 1  4 m   q  m  4 0  q
  m 1  m  q m
m 1
 vq 1  x   m  0
q
l
  0  q 1    m 1  q  m  1  q
q

m 1      . m qm
q 1 q 1  vq 1  x   m0

  0  q 1    m  q 1 m  1  q     m  q 1 m l
m2 m0
Where vq 1 is the static shear force in the span q  1 taken
Which completes the demonstration by recurrence.
independent and subjected to the external loads which it
supports.
2.3. Practice of Determinant method
According to practice of the determinant method, we 2.4.3. Support reactions of continuous beam
recommend following the steps: The support reactions of the continuous beam [3] are
a) Calculate the  m for m  1; n given by the following equation:
b) Calculate the  m for m  0; n  1 using equation q  0, n  1
 4  . To further simplify the method, we propose M q  2M q 1  M q  2
the following table grouping together the first Rq 1  x   rqr1  rql1  9 
l
values of the determinants: q 1

Table 1. First values of determinants and their difference   m  2 m 1   m  2   q 1 m   q 1

n n   n   n 1  n r r
q 1 rl
q 1  m 1

-1 0 -
l
q
6   m  q  m   q 1
0 -1 -1
1 4 5
m0
2 -15 -19  rqr1  rql1 
3 56 71
l
4 -209 -265 r l
Where r q 1 and r q 1 are respectively the static right and
5 780 989
6 -2911 -2131
left support reactions in the span q  1 taken independent
7 10864 13775 and subjected to the external loads which it supports.
8 -40545 -51409
9 151316 191861 2.5. Example of application of determinant
10 -564719 -716035
method
c) Calculate the only unknown moment M1  0 We will apply this method on an example of a
continuous beam to show its simplicity and how to apply
it in practice. The continuous beam is a beam of six 1 0 0 0   M1   0 
identical spans; that is to say n  5 . The beam in question  M   
is subjected to the loading described by the following 4 1 0 0  2   1 
diagram. The concentrated loads P and 2P are applied 1    
    
to the middle of the spans. 0 1 4 1 0   M k    k 
0 0 0 0   
    
 1 4 1 0   M n 1    n  2 
0
 0 0 1 4 1   M n    n 1 
We denote this matrix Tn and we call it matrix of the
Figure 2. Example for application to continuous beam three-moments of dimension n. The characteristic
polynomial of this matrix is given by:
First, let's calculate the  i using the rotation formulas [3,4]: Pn     det Tn   I n   1   
n

 6 EI  Pl 2 Pl 2  5Pl Using the Cayley-Hamilton theorem, the inverse of the

 1     matrix of three moments is given by:
 l  24 EI 16 EI  8
n 1
  2 2
 Tn 1    1 Cnp 1Tn p
p
10 
  2  6 EI   Pl  2 Pl    Pl p 0
 l  24 EI 16 EI 
 We define the following polynomial which gives the
 6 EI  7 Pl 2 2 Pl 2  13Pl trinomial coefficient [5] for x  1 :
 3     min  k , n 
 l  360 EI 16 EI  15
  nk 2  x    Cnp C pk  p x 2 p  k k  0; 2n
  4  6 EI   Pl    2 Pl
2
 k 1 
p 
 l  45 EI  15  2 

 We can demonstrate by recurrence that:

 6 EI  Pl  2
3Pl q  2, 2k  kq11 2  x    kq 21  x   x kq 2  x    kq 21  x 
  5  l   16 EI    8
  
This family of polynomials is symmetric according to n,
The value of the determinant is calculated in the table 1 which means:
 5  780  . Thus, M1  0 is given by: k  0; n  nn 2k  x    nn 2k  x 
1 Indeed:
0  M1   5  4 4  153  56 2  2091   pk n p n!
780
Cn C p  k   n  p  k  !  n  p  !  2 p  k  n 
1  3Pl 2 Pl 13Pl 5 Pl  
  4  15  56 Pl  209  
780  8 15 15 8  C p C n  k  p  n!


328Pl  n p  n  p  ! n  p  k  ! 2 p  k  n 
2925  Cnp  k C pn  kp  Cnp C pn  k  p
The other hyperstatic moments M i are given by We give below the expressions of some polynomials:
q  n0 2  x   1
M q 1     m  q  m : 
m0  n1 2  x   nx

 5173Pl
 2 n  n  1 2
 M 2  1  4  0  120  n 2  x   n  x
  2
 M    4   15  56   587027 Pl
 4  3 n  n  1 n  2  3
 n 2  x   n  n  1 x 
3 2 1 0
120 x
 6
 M    4   15    2621 Pl 
 3 2 1 0
2  ...
  n
 M 5   4  4  3  15  2  56 1  209  0  
273858 Pl  nn 2  x    C pn  p x 2 p  n
 20   n 1 
The problem is therefore completely solved using  p
 2 


equations  7  ,  8  and  9  . Given that  nk 2  4  k  0;2n  is symmetric according

to n, we will therefore construct a triangle to simplify the
calculation of hyperstatic moments (below an example up
3. Second method: Trinomial method to n = 5):
Table 2. Staircase polynomial  nk 2  4  (n from 0 to 5)
This method is based on an approach similar to that 1
used for the determination of the expression of the 1 4 1
trinomial coefficients or in general the p-nominal 1 8 18 8 1
coefficients [5]. We write the three-moments equation in
1 12 51 88 51 12 1
the same form as equation  3  . By writing this linear
1 16 100 304 454 304 100 16 1
system in its matrix form, we obtain: 1 20 165 720 1770 2424 1770 720 165 20 1
3.1. Trinomial method statement the matrix N n p which is only the monomial of index
p
The hyperstatic moments at the supports of a continuous
q  p of 1  4    C pj 4 p  j . So, we can write that:
p

beam of n  1  n  1 spans of the same length l are j 0

given by the following expressions: min  k , q 

Lkq   kq 2  4    Ckp C pq  p 42 p  q q  0, 2k
 n2  q 1 
  n    m 1  4m  n 1m p
 2 


  0  M 1  m0
Using equation 10  , the coefficients of the first
  n  2  4 n 1 11 1
 q column of Tn are given by:
q  0, n  1 M q 1    m  q  m n 1
   1 Cnp 1 pq 2  4    q
p
m 0
n 1  q 1
p
  1 Cnp 1 pm 2  4  12  
Where m 
p
 2 

 m 1
p 
Which give us 13 :
 2 
q n 1 q
M q 1     1 Cnp 1 pm 2  4  q  m    m  q  m
p

3.2. Demonstration of Trinomial method m0  m 1

p
m0

 2 
First, we can demonstrate using the characteristic
Using expression 13 and the following equation
equation of the sequence of   n n that:
M n 1  4M n   n , the expression of the unknown of the
We easily notice that: Tn  Nn  I n . The matrix N n is problem 0  M1 is:
strictly triangular. Let calculate Nn p  p  n 1 . For the n2
 n    m 1  4 m  n 1 m
first column, we have p  1 zeros above followed by
 0  M1  m0
p  n  2  4 n 1
monomials of 1  4    C pj 4 p  j :
p

j 0

 0 0 0 0 0 3.3. Practice of Trinomial method

  According to practice of the determinant method, we
 0 0 0
 C p0 4 p  recommend following the steps:
  a) Calculate the  m for m  1; n
 
Nn p   Cp 4
k p k  b) Calculate the m for m  0; n 1 using
 
  equation 12  . To further simplify the method,
 Cp 0 0 
 p we propose the use Table 2.
 0 0 c) Calculate the hyperstatic moments
  M q 1  q  1, n  1  .
 0 C p
p C 4k
p
pk 0
C 4
p
p
0 
k
Using Newton's binomial, we have: Tn k   Ckp N n p . 3.4. Expressions of forces in continuous beam
p 0
k
The matrix Tn is also lower triangular of which it is 3.4.1. Bending moment in continuous beam
enough to know its first column to calculate it (the The bending moment [3] in the different spans of the
coefficients of each diagonal are the same). Thus, we can
continuous beam is given by: q  0, n  1
see that the coefficient of index q  q  0 of the first
 x x
column of the matrix Tn k , noted Lkq , contains the M q 1  x   mq 1  x   M q 1    M q 1
coefficient of index q  q  0 of the first column of the  l l
 x
q
matrix N n p if p  q  p  p  2 p , that is to say  mq 1  x     m 1    m   m 1   q  m
m0  l
q  q  1
 p  q either the index p begins of p    .
2  2  Where mq 1 is the static moment in the span q  1 taken
Also p  q , but p does not also exceed k , therefore independent and subjected to the external loads which it
supports.
 q  1
p  min  k , q  . So, we have    p  min  k , q  .
 2  3.4.2. Shear force in continuous beam
The first column of the matrix N n k contains k  1 The shearing force [3] in the different spans of the
nonzero coefficients and k zeros above, so the last continuous beam is given by: q  0, n  1
nonzero coefficient of the matrix Tn k is of index q such
M q 1  M q
that q  0  1  k  k  1 or more better q  2k , therefore Vq 1  x   vq 1  x  
l
0  q  2k . From the expression of Tn k , Lkq contains Ckp 1 q 1
and the coefficient of index q  p of the first column of  vq 1  x      m 1   m   q 1 m
l m 1
Where vq 1 is the static shear force in the span q  1 taken 4.1. Kernels method statement
independent and subjected to the external loads which it The hyperstatic moments at the supports of a continuous
supports. beam of n  1  n  1 spans of the same length l are
3.4.3. Support reactions of continuous beam given by the following expressions:
n
The support reactions of the continuous beam [3] are M k   mkj  j
given by the following equation: j 1

The m pq are called arms and given by:

q  0, n  1
M q  2M q 1  M q  2  kp   kq 
Rq 1  x   rqr1  rql1  sin   sin  
 1
pq
 n 1  n 1
n
l
1 q 1
m pq  
n  1 k 1  k 
15 
 rqr1  rql1    m  2 m1   m 2   q 1 m
l m 2
2  cos  
 n 1
Where rqr1 and rql 1 are respectively the static right and
left support reactions in the span q  1 taken independent 4.2. Demonstration of Kernels method
and subjected to the external loads which it supports.
First, we can demonstrate using the characteristic
equation of the sequence of   n n that:
3.5. Example of application of Trinomial
method By putting U  4  X and n U   n  4  U  in

We keep the same example treated in paragraph 2.6. system 14  , we will get:
We will therefore have the same values of  i . Thus, we
 0 U   1 ; 1 U   U

calculate 0 using system 10  : 
 n U   U  n 1 U    n  2 U 
 n2
 1  4 0  1
 Using characteristic equation r  ur  1  0 , we prove
2

 0  41  15 that the expression of the polynomial function is given by:

 1  4 2  56
  2  43  209   u  u2  4  
n 1 n 1
u  u2  4 
1
1  3Pl 5 Pl  n u        
0  M1 
2 Pl 13Pl      
 4  15  56 Pl  209 u  4  2 2
2
    
780  8 15 15 8 
u

328 Pl Now, using  cos   , we can write n  u  in a
2925 2
Then, we calculate the  m and we deduce M q 1 . The trigonometric form and then determinate the roots of  n .
problem is therefore resolved. After a classical mathematical calculation, we get the
factorization as follow:
4. Third method: Kernels method n
  k 
 n  X    U  2 cos  
k 1   n  1 
This method is based on an approach using polynomials n
  k 
and diagonalization. We write the three-moments equation    4  X  2 cos  
in the following form: k 1   n  1 
 M 0 =M n 1  0

 It is then easy to determine the spectrum of the matrix of
 M p  2  4M p 1  M p   p 1
 p2
system bellow:
Then, we write this linear system in matrix form:
  k  
Sp  MTn   4  2 cos   / k  1; n 
4 1 0 0   M 1   1    n  1  
 M   
1 4 1 0  2   2  Now, we are going to look for the eigenvectors to
0    
     diagonalize and reverse this matrix. Let
0 0 1 4 0   M k    k  X k  x1  n, k  ,.., xn  n, k   the eigenvectors of the matrix
0 0 1 0   
     in question associated with the eigenvalues
 0 1 4 1   M n 1    n 1 
0 4   M n    n   k 
 0 0 0 1 n , k  4  2 cos   . we consider the numerical
 n 1
The characteristic polynomial of matrix of this system x p  n, k 
verify the following recurrent relationship: sequence defined by u p  n, k   . We get the
x1  n, k 
0  X   1 ; 1  X   4  X

 14 following system:
n  2  n  X    4  X   n 1  X    n  2  X 

   k  2
u1  n, k   1 et u2  n, k   2 cos  n  1 
On 1  
n 1
        2   k   n  
 1  1
k 1 n 1
 sin    sin   sin   sin  
  k    n 1  n 1  n 1  n 1 
u p  n, k   u p  2  n, k   2 cos   u p 1  n, k    2   4  2k  2n  
 n 1  1 sin   1 sin 
k 1 n 1
 sin   sin  
 
  n 1

 n 1

 n 1  1 
n
  k   
un 1  n, k   2 cos   u n  n, k 

  k   2k    k 2  kn  

  n 1  sin    sin    1
k 1
sin    1 sin 
n 1

  n 1  n 1  n 1   n 1 
 
The characteristic equation of this sequence is given by  
 
k   
  0 : r 2  2 cos   r  1  0 . This sequence is
  n 
 sin  
 2n 
 sin    1
k 1  nk 
sin    1 sin 
n 1  n 2
 


 n 1    n 1  n 1  n 1  n 1  
given by: We can thus invert the matrix of the linear system using
pk pk
p  i  the mathematical formula for diagonalization:
u p  n, k    1  an, k e n 1  bn , k e n 1 
i

  MTn1  On Dn1On1 
Using initial conditions, we easily find:  n 1 p 2 n 1 p 2 p n 1 p kp 1 p np 
n

 an, k       
i  1  p 1  n , p p 1  n , p p 1  n , p p 1  n , p 
      n 
 bn, k  2 sin  k    1   2 p1 p 2 p 2
n  2 p kp
n n  2 p np 
   
 
 n 1 p 1  n , p p 1  n , p p 1  n , p p 1  n , p

And then, we get:  
 
 1 i  i pkn 1 i pkn 1   n 
p

u p  n, k   e e    kp1 p n  kp 2 p n  pk 2 n  
kp np 
 k          
2 sin  p 1  n , p p 1  n , p
  p 1 n , p p 1 n, p

 n 1  
 pk    
sin    
 n 1   n   n  2 
  1  np np  np np
p 1 n n
  np 1 p    np 
 k   p 1  n , p p 1  n , p vp 
p 1  n , p 
sin    p 1
 n 1 We can notice that the coefficients of this matrix are
exactly the arms:
Then we divide by the norm . 2 in order to normalize the
eigenvector using the mathematical formula:  kp   kq 
sin   sin  
 1
pq
n  pk qk  n 1  n 1
n
n
 kq 
2
n 1  n, k
 
n  1 k 1  k 
 m pq
q  1; n  sin 
 n  1   2

k 1
2  cos  
k 1  n 1
And:
We find in this case the expressions of the hyperstatic
 pk  
sin  
u p  n, k   n 1 
moments as follows:
  1
p 1
 pk  n
u p  n, k  n
 qk  
2 M k   mkj  j
2

q 1
sin 
 n 1
 j 1

We give some examples of arms for n  1 and n  2 :

2  pk   
  1
p 1 1
sin  
n 1  n 1  n  1: m11  4

We call the following matrix Normalized kernels matrix: n  2 : m  m  4 m12  m21  
1
 11 22
15 15
 11  u  12  u  1k  u  1n  u  
  4.3. Practice of Kernels method
21  u  22  u  2 k  u  2 n  u  
  According to practice of the determinant method, we
  recommend following the steps:
On  u   k1  u  k 2  u  kk  u  kn  u  
  a) Calculate the  m for m  1; n
 
k, j  
2
  b) Calculate the mkj for 0; n using
  equation 15 
n1  u  n 2  u  nk  u  nn  u  
c) Calculate the hyperstatic moments
Consequence. Normalized kernels matrix is an M q 1  q  0, n  1 
orthogonal matrix [6] (it is the passage matrix of a
symmetric matrix). Therefore, its inverse matrix is given
by:
4.4. Expressions of forces in continuous beam
4.4.1. Bending moment in continuous beam
 The bending moment [3] in the different spans of the Then, we calculate the  m and we deduce M q 1 . The
continuous beam is given by: problem is therefore resolved.
q  0, n  1
5. Fourth method: Binomial Difference
 x x
M q 1  x   mq 1  x   M q 1    M q 1 method
 l l
 x
 
n
 mq 1  x     mqj  m q 1 j  mqj   j
This method is the most difficult method to establish its
j 1  l generalized theoretical formulas and its proof. For that, we
will be satisfied to present its theoretical framework, then
Where mq 1 is the static moment in the span q  1 taken we will treat in detail example already solved by
independent and subjected to the external loads which it preceding methods. We consider following linear system:
supports.  M 0  M n 1  0


 M p  2  4M p 1  M p   p 1
 2  p  n 1
4.4.2. Shear force in continuous beam
We can easily write MTn  Sn  4I n , where:
The shearing force [3] in the different spans of the
continuous beam is given by: 0 1 0 0
 
q  0, n  1 1 0 1 0 
0 1 0 
M q 1  M q  
Vq 1  x   vq 1  x   Sn   0 0 
l  0 1 0
 
  m 
n

q 1 j
 mqj  j  0 1 0 1
0 0 
 vq 1  x  
j 1
 0 1
l
Where vq 1 is the static shear force in the span q  1 taken 5.1. Theorical context of binomial difference
independent and subjected to the external loads which it method
supports. It is easy to show by induction that we can write  n
4.4.3. Support reactions of continuous beam defined by 14  in the following form:
The support reactions of the continuous beam [3] are  n
    1 Cnnkk  4  X 
nk
 
2k
given by the following equation:  2n X
 k 0

 n 2k
q  0, n  1    1
nk  p 2k  p
 4 Cnnkk C2pk X p
M q  2M q 1  M q  2  k 0 p 0

Rq 1  x   rqr1  rql1  
n

2 n 1  X     1 Cnnkk1  4  X 
nk 2 k 1
l
 k 0

 m  
n
 2m q 1 j  m q 1 j  j 
n 2 k 1
    1
n  k  p 2 k 1 p
Cnnkk1C2pk 1 X p
qj
4
r r
r l

j 1

q 1 q 1
l  k 0 p 0

Using Cayley-Hamilton theorem, we can easily get:

 1 n 2 k 1 p
    1
Where rqr1 and rql 1 are respectively the static right and 1 n  k  p 2 k  q 1 n  k
 MT2 n  4 Cn  k C2pk1C pq S2 n q
left support reactions in the span q  1 taken independent   2 n k 1 p  0 q  0
 n 2k p
 MT 1  1
and subjected to the external loads which it supports.
  1
nk  p 2k q
4 Cnnkk1C2pk11C pq S2 n 1q
 2 n 1
 2 n 1 k  0 p  0 q  0

4.5. Example of application of Kernels
method Where  2n and  2 n1 are given by equation  6 . By
We keep the same example treated in paragraph 2.6. simplifications, we find:
We will therefore have the same values of  i . Thus, we
calculate 0 using system 10  :  1 2 n 1


MT 2 n
1

 2n m0
  m,2n S2n m
 1  4 0  1
 
 0  41  15
2n
 MT 1  1
  2 n 1 
 2 n 1 m  0
 m ,2 n 1 S2 n 1m
 1  4 2  56
  2  43  209  n 2 k 1

   1
n  p  k 2 k 1 m
  m ,2 n  4 Cnnkk C2pk1C pm
1  3Pl 2 Pl 13Pl 5 Pl    m  pm
k   1
0  M1   4  15  56 Pl  209  2 
780  8 15 15 8   n 2k

   1
n p k 2k m
328 Pl  4 Cnnkk1C2pk11C pm
  m ,2 n 1  m 1  p  m
2925  k 
 2 

Now, we should calculate S n p to establish the expression  0 1  0, 1 0 0 0 
 
1
of MTn and then invert the linear system. For that, we 
 1  0, 1 0 1  0, 1 0 0 
introduce the differential binomial coefficient defined by: S51   0 1  0, 1 0 1  0, 1 0 
 
 0 0 1  0, 1 0 1  0, 1 
  p, q   2 and q  p  n n  p, q   Cnp  Cnq  
 0 0 0 1  0, 1 0 
We round the binomial differences in a triangle that looks
like Pascal's triangle. We call Trp the triangle of order p 0 1 0 0 0
 
of differential binomial coefficients. The anti-diagonal of 1 0 1 0 0
the triangle Trp is made up of the binomial coefficients  0 1 0 1 0
 
C pk . It’s easy to establish this triangle. We give some 0 0 1 0 1
0 0 
examples.  0 0 1
 2 1,0  0 2  0, 1 0 0 
 For p  3 :  
 0  2 1, 1 0  2  0, 1 0 
2 1 S5   2 1, 1
2
0 2 1, 1 0 2  0, 1 
 
2 3  0 2  0, 1 0 2 1, 1 0 
  
3  0 0  2 0, 1 0  21,0
1 1 0 1 0 0
 
The first anti-diagonal is made up of C 3k : 1, 3, 3, 1. Now, 0 2 0 1 0
 1 0 2 0 1
to get the other coefficients, we just use the subtraction  
between the binomial coefficient in the same diagonal of 0 1 0 2 0
0 1 
this coefficient and the first element of the anti-diagonal:  0 1 0
2  3 1 .  0 3 1,0  0 3  0, 1 0 
 
 For p  4 :  3 1,0  0 3 1, 1 0 3  0, 1 
The first anti-diagonal is made up of C 4k : 1, 4, 6, 4, 1. The S5 3   0 3 1, 1 0 3 1, 1 0 
 
second anti-diagonal is made up of 3  4  1; 5  6  1 . The  3  0, 1 0 3 1, 1 0 3 1,0  
 0 3  0, 1 0 3 1,0  0 
last anti-diagonal is made up of 2  5  3 .  
0 2 0 1 0
2 3 1  
2 0 3 0 1
5 4  0 3 0 3 0
 
3 6 1 0 3 0 2
0 0 
4  1 0 2

1
 4  2,1 0 4 1,0  0 4  0, 1 
 
 0 4  2,0  0 4 1, 1 0 
 For p  5 : S5  4 1,0 
4  0 4  2, 1 0 4 1,0  
 
similarly, we find:  0 4 1, 1 0 4  2,0  0 
   0, 1 0 4 1,0  0 4  2,1 
 4
5 4 1 2 0 3 0 1
 
5 9 5 0 5 0 4 0
 3 0 6 0 3
9 10  
0 4 0 5 0
4 10 1 0 3 2 
 0
5
1 Now, we can calculate MT5 1 using the following
expression:
Since S n is symmetrical then S n p is too. In addition, S n p 1  4 
contains Trp by filling the empty boxes (in gray) of Trp MT5 1   
5  m 0
 m,5 S5 m 

with zeros. Then, we keep same values of the first anti- 2 2k

   1
pk
diagonal of Trp by following an anti-diagonal translation. where  m,5  42 k  m C32kk C2pk11C pm
 m 1  p  m
k  
Then, we fill in the rest with zeros. Thus, we will get S n p .  2 

Let doing the example already treated so that the principle It’s easy to calculate  m ,5 for m 0, 4 . The expressions
of this method is quite clear. We have n  5 , so we will
of hyperstatic moments are given by:
calculate S5 p for p  0, 4 :
    0,5   2,5  2 4,5  1  1,5  2 3,5   2  
M1  1  
  5  2,5  3 4,5   3   3,5  4   4,5  5 
 
   2 3,5  1   0,5  2 2,5  5 4,5   2  
 M 2  1  1,5 
  5 1,5  3 3,5   3   2,5  4 4,5   4   3,5  5 
  
  2,5  3 4,5  1  1,5  3 3,5   2  
  

 0,5  2 2,5  6 4,5   3  1,5  3 3,5   4  
1
M 3 
 5  
  2,5  3 4,5   5 

 1  3,5 1   2,5  4 4,5   2  1,5  3 3,5   3 
M 4   
  5    0,5  2 2,5  5 4,5   4  1,5  2 3,5   5 
 

 1  4,5 1   3,5  2   2,5  3 4,5   3  
M 5   
  5  1,5  2 3,5   4   0,5   2,5  2 4,5   5 
 
Therefore, the problem is solved.

6. Conclusion
We have well developed the four methods for solving
the three moments equation in the case of identical spans.
These methods can be used to facilitate the calculations
and the resolution of the longest linear systems. Also, we
can program and use them. We notice that the first method
remains the simplest and most convenient method and the
last method is complicated to apply it. We hope that these
methods can be generalized in the general case of non-
identical spans.

These methods can be established in mathematics by

combining between matrix calculus, polynomials,
recurrent sequences as well as combinatorial in order to
use them in different scientific problems that can be
transformed in form of recurrent sequences with second
member.

Acknowledgements
The heading of the Acknowledgment section and the
References section must not be numbered.

References
[1] Stephen P. Timoshenko. “HISTORY OF STRENGTH OF
MATERIALS, With a brief account of the history of theory of
elasticity and theory of structures” (DOVER PUBLICATIONS,
INC., NEW YORK, 1983). ISBN 0-486-611 87- 6.
[2] HENRY T. EDDY, 0. E, PH. D. “The theorem of three moments”
American Journal of Mathematics, 1878, Vol. 1, No. 1 (1878), pp.
27-31.
[3] Arnaud Delaplace, Fabrice Gatuingt, Frédéric Ragueneau. “Aide-
mémoire -Mécanique des structures, Résistance des matériaux”
(Dunod, Paris, 2008). ISBN 978-2-10-053958-1.
[4] Jean COURBON, Jean-Noël THEILLOUT. “Résistance des
matériaux, formulaire”. © Techniques de l’Ingénieur, traité
Construction.
[5] Aziz Atta. “p-nomial Coefficients and p-nomial Theorem”.
Turkish Journal of Analysis and Number Theory, vol. 8, no. 1
(2020): 6-15. doi: 10.12691/tjant-8-1-2.
[6] V. N. Jha, Study on Hermitian, Skew-Hermitian and Unitary
Matrices as a part of Normal Matrices, International Journal of
Open Information Technologies ISSN: 2307-8162 vol. 4, no. 11,
2016.