CHEE 222: CHEMICAL PROCESS DYNAMICS

AND NUMERICAL METHODS

Module 2: Algebraic Equations, Single Variable

Nonlinear Equations Part I
 Topics to be Covered in this Module
This module deals with various forms of algebraic equations that arise for
lumped-parameter system operating under steady state. The following
topics will be covered:

• Non-linear Single-Variable System

– Iterative Solution
• Bisection Method
• Newton’s Method
• Convergence criteria
– Computer-based solution
• Multivariable Systems
– Linear equations
– Non-Linear systems

2
 Introduction
• In Module 1 we developed models consisting of ordinary
differential equations

• To solve these equations we need initial conditions

• Initial conditions will often be steady-state value (ex

monitoring deviation from steady state with a disturbance)

3
 Solving Steady State Parameters
• To solve for steady state parameters the time derivative term is set to 0:

• Or in vector notation:

4
 Introduction
• For now we will deal with methods of solving
single variable non linear algebraic equations
– Graphical
– Direct Substitution/Fixed Point Iteration This
– Interval Halving (Bisection) Lecture
– False Position (Reguli Falsi)
– Newton’s Method

• Often iterative methods

5
Graphical Method
 Graphical Method
• Can find the values of roots by examining the
graph of a function:

-3

-5

7
 Graphical Method
• Advantages
– Can visually see the relationship between x and f(x)
– Can see all the roots in the plotted domain (other methods may only find one root, or another, depending on initial guess)

• Disadvantages
– Manual method, can be slow
– Can be imprecise (only an approximation, useful for initial guess)

8
Direct Substitution/Fixed Point
Iteration
 Direct Substitution/Fixed Point Iteration
• Relatively simple method for solving single algebraic equation
with single unknown:

• We rearrange this to the equivalent form:

• Our guess for x at iteration k+1 is based on the evaluation of g(x)

10
 Direct Substitution/Fixed Point Iteration
• Example:

• By factoring we know roots will be:

• We can rearrange to isolate for x in various ways. For

example:

11
 Direct Substitution/Fixed Point Iteration
• Example:

• By factoring we know roots will be:

• We can rearrange to isolate for x in various ways. For

example:

12
 Direct Substitution/Fixed Point Iteration
• If we start with an initial guess of 4:

• Monotonic convergence to root at x=3

13
 Direct Substitution/Fixed Point Iteration
• If we rearranged to get:

• And start with an initial guess of 4:

• Oscillatory convergence to root at x=-1

14
 Direct Substitution/Fixed Point Iteration
• If we rearranged to get:

• And start with an initial guess of 4:

• Oscillatory convergence to root at x=-1

15
 Direct Substitution/Fixed Point Iteration
• If we rearranged to get:

• And start with an initial guess of 4:

• Diverging

16
 Direct Substitution/Fixed Point Iteration
• Convergence depends on the initial guess used for a
particular rearrangement
• Additional information on direct substitution/fixed
point iteration convergence can be found in:
– Gerald, C.F., Wheatley, P.O. (1999). Applied Numerical
Analysis. New York: Addison-Wesley
– Chapra, S.C., Canale, R.P. (2006). Numerical Methods
for Engineers. New York: McGraw-Hill.

17
Interval Halving (Bisection)
 Interval Halving (Bisection)
• Conceptually simple method of finding roots
• Requires that the sign of the function is known
• Starts with two x values, a and b, that bracket a
root
– If f(a)f(b)<0 (opposite sign), and the function is
continuous, than the interval between a and b will
contain at least one root

19
 Interval Halving (Bisection)
• Conceptually simple method of finding roots
• Requires that the sign of the function is known

• Starts with two x values, x1 and x2, that bracket a root

– If f(x1)f(x2)<0 (opposite sign), and the function is
continuous, than the interval between x1 and x2 will contain
at least one root
• Evaluate the function at the midpoint
• Replace bracket limit that has the same sign as the
function value at the midpoint, with the midpoint value
• Repeat

20
 Interval Halving (Bisection)
• For example:
– Find x1, where f(x1) is negative and x2 where f(x2) is
positive
– Determine the midpoint (x3). If f(x3) is negative:
throw out x1 and make x3 lower bracket
– Determine midpoint between x3 and x2 (x4). If f(x4)
is positive: throw out x2 and make x4 the upper
bracket
– Find midpoint between x3 and x4 (x5). If the f(x5) is
negative: throw out x3 and make x5 the lower
bracket

21
 Interval Halving (Bisection)
• Graphically: Initial
Upper
Limit

Initial
Lower
Limit 22
 Interval Halving (Bisection)
• Advantages:
– Simple to understand
– Guaranteed to work if f(x) is continuous in [a,b] and
the values of a and b bracket a root (f(a)f(b)<0)

• Disadvantages:
– Slow to converge
– May be difficult to find an interval that brackets
roots when there is more than one root
– Only finds one root at a time

23
False Position (Reguli Falsi)
 False Position (Reguli Falsi)
• Referred to as False Position or Reguli Falsi
• Uses two previous iterations to determine next
iteration value
• Steps:
– Select xk and xk+1 to bracket solution
– Draw line between f(xk) and f(xk+1), the x-intercept
of this line is xk+2
– Evaluate f(xk+2).
– Replace bracket limit that has the same sign for its
function as f(xk+2)
– Continue until tolerance is met

25
 False Position (Reguli Falsi)
• Advantages:

Initial
Lower
Limit

Initial
Upper
Limit

26
 False Position (Reguli Falsi)
• In the above plot x3 can be found using the
following equation:

27
 False Position (Reguli Falsi)
• Advantages:
– Can be used when the derivative of the function is
unknown
– Generally converges more quickly than the
bisection method

• Disadvantages
– Potentially slower than other methods
– May be difficult to find an interval that brackets
roots when there is more than one root
– Only finds one root at a time

28