Approximate Methods: Point Collocation Method
The statement of the equilibrium equations applied to a set  is as follows. Assuming that at equilibrium
 is as follows. Assuming that at equilibrium  is the symmetric Cauchy stress distribution on
 is the symmetric Cauchy stress distribution on  and that
 and that  is the displacement vector distribution and knowing the relationship
 is the displacement vector distribution and knowing the relationship  , then the equilibrium equation seeks to find
, then the equilibrium equation seeks to find  such that the associated
 such that the associated  satisfies the equation:
 satisfies the equation:
      ![Rendered by QuickLaTeX.com \[ \mathrm{div}\sigma+\rho b= 0 \]](https://engcourses-uofa.ca/wp-content/ql-cache/quicklatex.com-e526576fab30481b2c1132610584b9ac_l3.png)
where  is the body forces vector distribution on
 is the body forces vector distribution on  ,
,  is the mass density, and
 is the mass density, and  is the space of all possible displacement functions applied to
 is the space of all possible displacement functions applied to  , i.e.,
, i.e.,  . The term “Kinematically admissible” in
. The term “Kinematically admissible” in  indicates that the space of possible solutions must satisfy the boundary conditions imposed on
 indicates that the space of possible solutions must satisfy the boundary conditions imposed on  (as stated below) and any differentiability constraints.
 (as stated below) and any differentiability constraints. 
The boundary conditions for the equations of equilibrium are usually given on two parts of the boundary of  denoted
 denoted  . On the first part,
. On the first part,  , the external traction vectors
, the external traction vectors  are known so we have the boundary conditions for
 are known so we have the boundary conditions for  since
 since  (
 ( is the normal vector to the boundary). On the second part,
 is the normal vector to the boundary). On the second part,  , the displacement is given.
, the displacement is given. 
The point collocation method seeks an approximate solution to the above problem by assuming that the solution  has a particular form with a finite number of unknowns, i.e., by looking for
 has a particular form with a finite number of unknowns, i.e., by looking for  in a subset
 in a subset  that is finite dimensional but still able to approximate functions in
 that is finite dimensional but still able to approximate functions in  . The finite number of unknowns can be found by satisfying the differential equation of equilibrium at a number of chosen points equal to the number of unknowns.
. The finite number of unknowns can be found by satisfying the differential equation of equilibrium at a number of chosen points equal to the number of unknowns. 
Example
Using a polynomial function of the third degree, find the displacement function of the shown bar by satisfying the essential boundary conditions, the differential equation of equilibrium at  and at
 and at  ., and the nonessential boundary condition at
., and the nonessential boundary condition at  . Assume that the bar is linear elastic with Young’s modulus
. Assume that the bar is linear elastic with Young’s modulus  and cross-sectional area
 and cross-sectional area  and that the small strain tensor is the appropriate measure of strain. Ignore the effect of Poisson’s ratio.
 and that the small strain tensor is the appropriate measure of strain. Ignore the effect of Poisson’s ratio.

Solution
Exact Solution
The exact solution can be obtained by directly solving the differential equation of equilibrium utilizing  :
:
      ![Rendered by QuickLaTeX.com \[ \frac{\mathrm{d}^2u_1}{\mathrm{d}X_1^2}=-\frac{cX_1}{EA} \]](https://engcourses-uofa.ca/wp-content/ql-cache/quicklatex.com-e217ce8a05948eb7cad9ddbffe0fd596_l3.png)
with the boundary conditions:  and
 and  .
.
Therefore:
      ![Rendered by QuickLaTeX.com \[ u_1=\frac{cL^2}{2EA}X_1-\frac{c}{6EA}X_1^3 \]](https://engcourses-uofa.ca/wp-content/ql-cache/quicklatex.com-04850a8347c5fc3568692561add41bf3_l3.png)
DSolve[{u''[X1] == -c*X1/EA, u'[L] == 0, u[0] == 0}, u[X1], X1]
Approximate Solution
A polynomial of the third degree approximate solution has the form:
      ![Rendered by QuickLaTeX.com \[ u_{approx}=a_0+a_1X_1+a_2X_1^2+a_3X_1^3 \]](https://engcourses-uofa.ca/wp-content/ql-cache/quicklatex.com-7f7962616a945c3ce442eb459cdfc531_l3.png)
with
      ![Rendered by QuickLaTeX.com \[\begin{split} \frac{\mathrm{d}u_{approx}}{\mathrm{d}X_1}&=a_1+2a_2X_1+3a_3X_1^2\\ \frac{\mathrm{d}^2u_{approx}}{\mathrm{d}X_1^2}&=2a_2+6a_3X_1 \end{split} \]](https://engcourses-uofa.ca/wp-content/ql-cache/quicklatex.com-1a22992fa3e34263cf4a6b4e79be7521_l3.png)
The following four equations will be utilized to find the four unknowns:
      ![Rendered by QuickLaTeX.com \[\begin{split} @X_1=0&:u_{approx}=0\\ @X_1=L&:\frac{\mathrm{d}u_{approx}}{\mathrm{d}X_1}=0\\ @X_1=\frac{L}{2}&:\frac{\mathrm{d}^2u_{approx}}{\mathrm{d}X_1^2}\bigg|_{X_1=\frac{L}{2}}=-\frac{cX_1}{EA}\bigg|_{X_1=\frac{L}{2}}\\ @X_1=L&:\frac{\mathrm{d}^2u_{approx}}{\mathrm{d}X_1^2}\bigg|_{X_1=L}=-\frac{cX_1}{EA}\bigg|_{X_1=L} \end{split} \]](https://engcourses-uofa.ca/wp-content/ql-cache/quicklatex.com-a71f9ee3c68f6de8e689e9bfe5dee510_l3.png)
The approximate solution that satisfies the above four unknowns has the form:
      ![Rendered by QuickLaTeX.com \[ u_{approx}=\frac{cL^2}{2EA}X_1-\frac{c}{6EA}X_1^3 \]](https://engcourses-uofa.ca/wp-content/ql-cache/quicklatex.com-bc15831d4fe7a75a814fde470299a81e_l3.png)
Notice that since the exact solution is contained in the space of approximate functions which were chosen to satisfy both the essential and nonessential boundary conditions, the solution obtained is indeed the exact solution!
View Mathematica Code
u = a0 + a1*x1 + a2*x1^2 + a3*x1^3;
ux = D[u, x1]
uxx = D[ux, x1]
Eq1 = (uxx + c*x1/EA) /. x1 -> L/2;
Eq2 = (uxx + c*x1/EA) /. x1 -> L;
Eq3 = (ux) /. x1 -> L;
Eq4 = u /. x1 -> 0;
s = Solve[{Eq1 == 0, Eq2 == 0, Eq3 == 0, Eq4 == 0}, {a1, a2, a3, a0}]
u /. s[[1]]
