Assignment 1
Name-Evin Daniel Renji
Roll no.- 2018087
Introduction
Finite element methods are techniques used to solve and analyse physical
and mathematical models. They provide us with an approximation for the
governing differential equations of the model.This assignment consists of a
heat transfer problem which has been approximated by these methods.
The methods used to solve the problem have been listed below:
1.Point collocation
2.Least square method
3.Galerkin approach
4.Galerkin weak form
The methods selected provide an approximation for our model. Piecewise
formulation was omitted since their real power can only be seen when we
perform refinement of the mesh. To do this we could use commercial
softwares like ANSYS and ABAQUS.
Physical model
A heat transfer problem has been considered for our analysis.Inspiration
for the problem was drawn from the DS302 project where a sustainable
vegetable cart was to be made as shown below.
The container consists of a refrigeration system and a rod will run over the
entire chamber. This rod (made of wood) will support the mist sprinklers
and will be in our main interest.
(Air enters the model through holes which will be provided at the
front of the cart.It will pass through the cooling pad and a fan will be
placed on the wall shown to keep the vegetables fresh. Those details have
been left out here to deal with our interests.)
The refrigeration system will cool the divider and hence heat transfer
will occur across the entire rod in the form of convection to air and
conduction across the material. Only one dimension will be considered to
stick with the syllabus.2 images have been included for more context.
Derivation of pde
Assumptions considered:
1.The wall is at a constant temperature of 24°C considering the
refrigeration system provides a cooling of 3°C .
2.The end of the rod will be insulated i.e. No heat transfer occurs from the
end.
3.Air around the rod will be flowing at 1.5 m/s.
4.Thermal conductivity of the rod remains constant.
5.Room temperature (Tinf ) = 27°C .
6.Conduction occurs through the material and convection occurs through
surfaces in contact with air.
Derivation for our required equation has been provided in the image below
using energy conservation .
Problem statement
Q. Solve the following PDE with the boundary conditions using the 4
methods
Given:
h= 24.58 W/m2 °C Ac=0.006 m2
k=0.7 W/(m2 °C ) P=0.34 m
L=0.6 m
Tinf=27°C
Solution
1.Point collocation
Method overview:In this method we solve for the coefficients by
considering the residual to be equal to 0 at an arbitrary point.
Let us assume a polynomial trial solution:
Applying the boundary conditions:
Therefore trial solution in one parameter is:
Minimising the residual by equating with 0 gives:
Solving and substituting the given values and putting in trial solution
Putting an arbitrary value of x=L/2=0.3 gives
Therefore approximate solution is:
2.Least square method
Method overview-Our main aim in every method is to minimise the
residual as much as possible.In here we will square integrate the residual and
equate it with 0 to find the required coefficients
Let us assume a polynomial trial solution:
Applying the boundary conditions:
Therefore trial solution in two parameters is:
(After substituting L=0.6)
Now,
Squaring R gives:
Integrating from 0 to L(0.6) having weighting function as unity:
Differentiating this wrt c2 and c3 gives the following two equations:
Solving them gives
Therefore approximate solution found using this method:
3.Galerkin Approach
Method overview- Galerkin suggested the usage of the same basis or
shape functions for the trial and weight functions.This approach has produced
much better approximations and this idea had revolutionised the study of
approximating differential equations.
Let us assume the same polynomial trial solution from method 2:
The two weight functions will be:
The residual will again come out as found before
Next step is to integrate the residual with the weights from 0 to L(0.6)
--1
Again,
--2
Solving 1 and 2 gives:
Therefore approximate solution found using this method:
4.Galerkin Weak form
Method overview-The weak form of an equation is obtained when the
governing differential equation is integrated and multiplied with a weight .This
equation is then integrated by parts which results in lowering the differential
order.This method helps us to consider a lesser order polynomial trial function.
Governing differential equation:
Applying integration by parts:
From our 2nd boundary condition
Now let's consider our trial solution as:
From the first boundary condition:
The weighting functions and their derivatives will be:
Using W1=x in our weak form:
--1
Using W2=x2 in our weak form:
--2
Solving 1 and 2 gives:
Therefore approximate solution found using this method:
Conclusion
The methods have been studied and used effectively on a physical model of a
heat transfer problem.
