Anda di halaman 1dari 27

ME 6000 COMPUTATIONAL METHODS IN ENGINEERING

3-0-0-3

Course Content:
Approximations: Accuracy and precision, definitions of round off and truncation errors, error propagation. Algebraic equations: Formulation and solution of linear algebraic equations, Gauss elimination, LU decomposition, iteration methods (Gauss-Seidal), convergence of iteration methods, eigen values and eigenvectors. Interpolation methods: Newtons divided difference, interpolation polynomials, Lagrange interpolation polynomials. Differentiation and Integration: High accuracy differentiation formulae, extrapolation, derivatives of unequally spaced data, Gauss quadrature and integration.

Transform techniques: Continuous Fourier series, frequency and time domains, Laplace transform, Fourier integral and transform, Discrete Fourier Transform (DFT), Fast Fourier Transform (FFT). Differential equations: Initial and boundary value problems, eigenvalue problems, solutions to elliptical and parabolic equations, partial differential equations. Regression methods: Linear and non-linear regression, multiple linear regression, general linear least squares. Statistical methods: Statistical representation of data, modeling and analysis of data, test of hypotheses. Introduction to optimization methods: Local and global minima, Line searches, Steepest descent method, Conjugate gradient method, Quasi Newton method, Penalty function. Solution to practical engineering problems using software tools.

Books:
1. Schilling R.J and Harris S L, Applied Numerical Methods for Engineering using MatLab and C, Brooks/Cole Publishing Co., 2000. 2. Chapra S C and Canale R P, Numerical Methods for Engineers, McGraw Hill, 1989. 3. Hines, W.W and Montrogmery, Probability and Statistics in Engineering and Management Studies, John Willey, 1990. 4.Santhosh K.Gupta, Numerical Methods for Engineers, New age international publishers, 2005.

1.Modeling and Error Analysis 1.0. Introduction:


Engineering problems require mathematical solutions. Some of the problems associated with these solutions are: Very large and tedious to calculate. Exact solution may not be available. With the development of inexpensive hardware and software, the role of numerical methods in engineering problem solving has increased dramatically in recent years.

Pre-Computer era approach to problem solving:


(i) Analytic or Exact method: Problems were
approximated with linear models, simple geometry and low dimensionality.

(ii) Graphical Solutions: These are usually in the form


of plots and nomograms. The results are not very precise. Often limited to the problems that can be described using three or fewer dimensions.

(iii)Calculators and Slide rules: Manual calculators


are slow and tedious. Consistent results are elusive because of simple blunders when numerous manual tasks are performed.

Computational Methods:
Computers and Numerical methods provide an alternative for such complicated calculations. The following are the advantages in using computational methods. (i)They are extremely powerful problem solving tools. Capable of handling large system of equations, non-linearities, complicated geometries. (ii) Computational methods reduce basic arithmetic operations. higher mathematics to

(iii) Results can be seen dynamically at the design stage and possible to control the errors due to various approximations. .

iv) Commercially available software packages involve


numerical method.

Caution while using commercial software:


a) Intelligent use of these programs requires knowledge of basic theory underlying the methods. the

b)

Many problems cannot be approached using the package programs. Due to this limitation, developing own programs to solve problems may be preferred. In some cases spending on expensive software can be avoided.

1.1 Mathematical Modeling and Engineering problem solving:


A mathematical model can be broadly defined as a formulation or equation that expresses the essential features of a physical system or process in mathematical terms. In general, it represents a functional relationship form.
Dependent variable = f {independent variable, parameters, forcing functions} (1.1)

Where, Dependent variable: reflects the behavior or state of the system. Independent variable: dimensions (Ex: Time and space). Parameters: reflective of systems properties. Forcing functions: External influences acting upon it.

Engineering problem solving process:

The mathematical expression of Second law of Newton F=ma (1.2) Eq. (1.2) can be recast in the form of Eq.(1.1) as a = F/m (1.3) Where a is dependent variable reflecting the systems behavior. F is forcing function. m is parameter representing a property of the system. Eq. (1.2) has a number of characteristics of a typical mathematical model. 1. It describes a natural process or system in the mathematical term. 2. It represents an idealization and simplification of reality. 3. It yields reproducible results.

Mathematical models of physical phenomena maybe more complex, and either cannot be solved exactly or require more sophisticated techniques than simple algebra for their solution. This is illustrated in the following example. Take a case of falling parachutist as shown in the Fig.1.2, and we are interested in the terminal velocity.

A model of this case can be derived from Eq. (1.2) by expressing the acceleration in terms of rate of change of velocity as dv m =F dt (1.4) F has two components FD : Downward pull of gravity. FU : Upward force of air resistance. F = FD + FU (1.5) With a +ve sign for downward force, using Newtons second law (1.6) FD = m g Assuming air resistance is linearly proportional to velocity and it acts in upward direction. FU = - c v (1.7) c is proportionality constant called drag co-efficient (kg/s).

Substituting Eqs. (1.5 - 1.7) in Eq. (1.4) (1.8) Dividing by , (1.9)


Eq.(1.9) is a differential equation and the exact solution is obtained from

(1.10)
Eq.(1.10) is in the general form of Eq. (1.1). Where

v(t) is the dependent variable t is independent variable. c and m are parameters. g is the forcing function

Example 1.1: A parachutist with a mass of 70 kg jumps out of a stationary hot air balloon. Compute velocity prior to opening the chute. The drag coefficient is12.5 kg/s. Solution: Using the Eq. (1.10)
9.81(70) (12.5/ 70) t [1 e ] = v(t ) 12.5

Velocity at different time steps are tabulated below t, s 0 2 4 6 8 10 12 ..


v, m/s 0 16.50 28.04 36.12 41.77 45.72 48.49 ... 54.94

Velocity Vs Time

Eq. (1.10) is called analytical or exact solution. A numerical solution approximates the exact solution. The time rate of change of velocity can be approximated as (1.11) Where v and t are differences in velocity and time computed over finite interval. Substituting Eq.(1.11) in Eq.(1.9)

v(ti + 1) - v(ti ) c = g v(ti ) ti + 1 - ti m

This equation can be rearranged to yield (1.12) c v(ti + 1) = v(ti ) + [ g v(ti )](ti + 1 ti ) m Eqn (1.12) is a transformation of differential equation (1.9) into an algebraic equation. New value = Old value + Slope x Step Size.

Exact Vs Numerical Solution

Example 1.2 Develop a numerical solution to falling parachutist problem given in example 1.1. Employ a step size of 2 s. Solution: At Initial time ( t i = 0), velocity of parachutist = 0 Substituting in Eq. (1.12) to compute the velocity at ti+1 = 2s For the next interval (from t = 2 to 4 s) the computation is repeated, resulting

The calculations are continued and tabulated below.

t, s
v,m/s,Exact

0 2
0 0 16.50 19.62

4
28.04 32.23

6
36.12 40.34

8
41.77 45.55

10
45.72 48.90

12

54.94

48.49 . 51.06 .

v, m/s Numerical

Bungee Jumping: Conservation of Mechanical Energy


The jump height off of the Verzasca dam is 220 m. Given a jumper of mass 77 kg, determine the relationship between the stiffness k of the bungee cord and its unstretched length L0; i.e., find k as a function of L0, so that the jumper has zero speed at the bottom of the dam. In addition determine the unstretched length so that acceleration of the jumper does not exceed 4 g during the jump

Fig.1

SOLUTION
Referring to the FBD in Fig, we model the jumper as a particle subject to gravity and the force Fb of the bungee cord, which we model as a linear elastic spring. The jumper begins the jump at 1 with zero speed and ends the jump at 2 with zero speed (see Fig.1). Since we know the jumpers speed at 1 and 2 and we know all the forces doing work on the jumper, we can apply the work energy principle to the jumper to determine the relationship between the bungee stiffness and its unstretched length. We will then apply Newtons second law to the jumper to determine the maximum acceleration so that we can find k and L0 for the bungee cord. Note that all the forces acting on the jumper are conservative and that Fb is zero until the jumper falls a distance equal to the unstretched length of the cord.

Governing Equations Balance Principles Applying the work-energy principle between 1 and 2, we have T1+V1 = T2 + V2 (1)

1 2 1 2 T1 = mv1 T2 = mv2 (2) and 2 2 where m is the jumpers mass and v1 and v2 are the jumpers speed at 1 and 2, respectively. Since we also need to determine the jumpers acceleration, we will write Newtons second law for the jumper in the y direction as

: mg Fb = ma y

(3)

where we note that Fb = 0 until the bungee cord engages.

Force Laws: If we place the datum line for gravitational potential energy at 2, then

V1 = mgh

and

1 2 V2 = k (h L0 ) , 2

(4)

where h= 220 m, mg =756 N, and we have accounted for the potential energy of the bungee cord in V2. We will also need the force law for the bungee cord, which is given by

Fb = k = k ( y L0 ),
where we note that Fb = 0 when y L0,

(5)

Kinematic Equations Since the jumper starts and ends the jump with zero speed, we have

v1 = v 2 = 0.

(6)

Computation Substituting Eqs. (2), (4), and (6) into Eq. (1) and solving for k, we obtain 2mgh (7) k=

(h L0 )2

Fig.2. Eq. (7) gives the desired k as a function of L0, a plot of which is shown in Fig.2.

We now want to design the bungee system so that the maximum acceleration of a jumper of mass 77 kg does not exceed amax = 4 g. Referring to the FBD Fig.1, we know that until the spring engages, the jumper will be in free fall and his acceleration will be g downward. Once the spring engages, the acceleration is determined by solving Eqs. (3) and (5) for ay, which gives
k a y = g ( y L0 ). m

(8)

Recall that Eq. (7) provides a relation between k and L0 that ensures that the jumper has zero speed at 2. Therefore, using Eq. (7), Eq. (8) becomes
ay = g 2 gh( y L0 )

(h L0 )

or

ay g

=1

2h( y L0 )

(h L0 )

, y > L0 ,

(9)

Fig.3

We will choose the following value of the unstretched length of the bungee cord: (10) We now need to verify that the design criterion requiring ay < 4 g is always met. To do so, using the chosen L0 and referring to Fig.3, we plot ay/g versus y. The plot shows that the chosen value of L0 is such that our goal is met.

Anda mungkin juga menyukai