Scribd
Upload
3K views

Applications of Numerical Methods in Civil Engineering

Numerical methods are widely used in civil engineering to analyze systems that cannot be solved through closed-form analytical solutions. Some key applications of numerical methods in civil engineering include structural analysis to determine member forces, fluid dynamics simulations, and geotechnical modeling of groundwater movement. Numerical methods allow engineers to create mathematical models of complex real-world systems and use computational techniques like matrix methods, numerical integration, and numerical differentiation to obtain approximate solutions. This reduces the need for physical prototypes and experiments and allows optimization of designs.

Uploaded by

Asniah M. Rataban
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
3K views

Applications of Numerical Methods in Civil Engineering

Numerical methods are widely used in civil engineering to analyze systems that cannot be solved through closed-form analytical solutions. Some key applications of numerical methods in civil engineering include structural analysis to determine member forces, fluid dynamics simulations, and geotechnical modeling of groundwater movement. Numerical methods allow engineers to create mathematical models of complex real-world systems and use computational techniques like matrix methods, numerical integration, and numerical differentiation to obtain approximate solutions. This reduces the need for physical prototypes and experiments and allows optimization of designs.

Uploaded by

Asniah M. Rataban
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 9

Applications of Numerical Methods in Civil Engineering

Why are Numerical Methods so widely used in Engineering?


 Engineers use mathematical modeling (equations and data) to describe and predict the behavior
of systems.
 Closed-form (analytical) solutions are only possible and complete for simple problems
(geometry, properties, etc.).
 Computers are widely available, powerful, and (relatively) cheap.
 Powerful software packages are available (special or general purpose).

A few applications of Numerical Methods in Engineering:


• Structural/mechanical analysis, design, and behavior.
• Communication/power
 Network simulation
 Train and traffic networks
• Computational Fluid Dynamics (CFD):
 Weather prediction
 Groundwater & pollutant movement

Why numerical modelling?


 To reduce the number of prototypes constructed.
 To use as the forward solver in an optimization routine to gain the best possible design.

What are the alternatives?


 Design and build lots of physical models
 Time consuming
 Expensive
 Measurement difficulties (cables, calibration, mechanical precision, interference,
mutual coupling, field confinement, etc)
 Analytical solutions
 Limited to simple models because of the computational complexity
Numerical methods in Civil Engineering are now used routinely in structural analysis to
determine the member forces and moments in structural systems, prior to design. ... Such a beam
can be easily analyzed using the finite element method - one of the most commonly used
numerical technique in civil engineering.

Underlying any engineering application is the use of Numerical Methods.  Numerical Methods is
a manner in which 'discretization' of solutions can be achieved rather than analytical solutions
(eg. integration, differentiation, ordinary differential equations and partial differential equations).
Numerical Methods are also all the techniques encompassing iterative solutions, matrix
problems, interpolation and curve fitting. 

1. Matrix Methods (solving systems of equations)


1. Simultaneous Linear Equations
1. Naive Gauss Elimination
2. Numerical Integration
1. Rectangle Method
2. Trapezoid method
3. Simpson's 1/3 Method
3. Numerical Differentiation

Matrix Methods

What's the point of using matrices?  Well, every engineering problem is represented as an


equation or a system of equations.  When we have a system of equations, we have a system of
variables that need to be solve.  The 'solved variables' represent the solution to our problem.  We
may have something like this:

ach equation can be written in the matrix form as follows:


A represents the matrix of coefficients in equations 1, 2, and 3.  The linear vector or matrix of
variables (x) is shown  above as well.  Let's move forward to solving this system of linear
equations.

Simultaneous Linear Equations

Naïve Gauss Elimination The idea behind solving a system of equations is to eliminate
unknowns.  The goal is to get the [A] Matrix into a upper or lower triangular matrix so
that number of equations and variables are equal.  The following picture demonstrates what we
are trying to accomplish.

Here we can see that a values in the [A] matrix are being eliminated in each iteration (step) of
'upper triangularizing' the [A] matrix.  The final matrix is the upper triangular matrix we want.
So, how is this accomplished using Naive Gauss Elimination?

Well, this method is called naive because it does not precondition the matrix my pivoting row or
columns, it also doesn't allow for 'selective harvesting' or eliminating of individual entries of [A]
to make our life easier.  The first step is to eliminate all x1  entries below the first row (as shown
in the above figure).  This is accomplished by first doing forward iteration as follows.

For equation 2 (second row) divide equation 1 (row one) by a11 and multiply by a21.  

Now, subtract this from row 2 to eliminate the first entry.  For the [3x3] case, the only remaining
entries will be a22  anda23.  This process must be repeated for each respective row until the
matrix is fully reduced to the upper-triangular form. Note that entries that operate as a
denominator must not be zero.  In other words, we should try to avoid reducing a matrix with a
determinate equal to or close to zero.
EXAMPLE:
Back Substitution is used in this example to solve for the unknowns in the reduced matrix.  The
example demonstrates the ability for this method to be performed by a machine as it requires no
external knowledge regarding row/column switching.  The method is a bit cumbersome and falls
apart when entries are zero.

Numerical Integration

Rectangle Method Integration by means of separating a space by 2D or 3D rectangles is


extremely straight forward.  I picture is worth a thousand words:
 
Clearly, we can evaluate the areas of individual triangles from the left, right or center of the
function to be integrated. The sum of the areas of the individual rectangles approximated the area
under the function.  If we use the midpoint rule, we end up with the following approximation:

The overall error of the integration estimation is a function of the step size taken.  Smaller step
sizes result in less error.
Trapezoidal Integration
In an attempt to reduce the error of integration estimation, trapezoidal integration may be used. 
Similar to rectangular integration:

 
The error is slightly reduced and the integration formula is as follows:

Simpson's Rule Simpson's rule is based off a polynomial area approximation of any order larger
than one (only whole integers).  The order of the polynomial is the highest exponent observed in
the polynomial function. Between two points a polynomial of some order is assumes and an
algebraic expression for the area is formed on the basis of the polynomial order.  The below
picture illustrates the approximation:
Again, the step size is critical here. As in the rectangular and trapezoidal integration
approximations, if the step size is minimized, the error is lessened.

For an order two Simpson approximation, the following formula may be used:

Remember, it is the summations of piecewise integrals that evaluates your function between
desired bounds.

Numerical Differentiation

Derivatives may be viewed as the slope between corresponding points in a function (1D case). 
Therefore, if we want to approximate the slope we simply need to know the step size and change
in function values at corresponding points.  The following depicts an approximate derivative:

Above we see an approximation of the derivative of a function between the function variable
values of x(k-1) and x(k+1); this approximation is the derivative at xk. Therefore the basic
formula for a two point approximate derivative in 1D is:

For a three point approximation in 1D, the derivative approximation is:


 
This method applies for partial derivatives as well.  The variable is simply taken in the function
values for the respective dependent variable and evaluated.  The numerical derivative is the
computed using the above equation(s).  Again, the accuracy of these methods is dependent on the
step size.

Conclusion 

Numerical Methods in Engineering  is an extensive topic to be covered.  Almost everything


done relies on basic principles of applied mathematics.  Vibrational and impact problems in
structural mechanics often require use of Fourier Series Analysis, heat transfer may require
advanced techniques in solving elliptic integrals.
Philippine Engineering and Agro-Industrial College Inc.,
Datu Gonsi St., Brgy. Lomidong, Marawi City, LDS

Application of Numerical Methods in Civil Engineering


ES 19 (Numerical Methods)
Engr. Abdul Aziz Imam Omar

RATABAN, Juma-a M.

August 02, 2021

You might also like