ChE – 422 PROCESS ANALYSIS & OPTIMIZATION

Introduction

Saeed GUL, PhD

Professor
Department of Chemical Engineering,
University of Engineering & Technology Peshawar, PAKISTAN
 Course Objective

To impart knowledge of various models used

for process and performance analysis for
optimization in process industry

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 2
 Introduction

What is a process model?

A process model is a system of mathematical equations and
constants that are usually solved on a computer to make
quantitative predictions about some aspect(s) of a real process.

The specific variables required as input data and generated as

output predictions are important features of the model. The
equations often stem from a numerical solution to one or more
differential equations and their boundary conditions.

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 3
 Introduction
TYPES OF MODELS
A fully-mechanistic, or phenomenological, model solves equations
based solely on the fundamental laws which govern natural
phenomena. These laws include the differential equations governing
the conservation and transport of mass, momentum, mechanical
force, electromagnetic force and energy, in addition to
thermodynamics, phase equilibria, kinetics, and other relations.
A fully-empirical model is created by performing a curve-fitting
procedure on the results of a statistical study with no attempt to
understand the reasons for the relationships. They are good in
situations where fundamental theory is difficult to apply. However,
they usually become very inaccurate if extrapolated beyond the
specific range of processing conditions for which they were
developed.

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 4
 Introduction

A stochastic model is a tool for estimating probability distributions of

potential outcomes by allowing for random variation in one or more
inputs over time. The random variation is usually based on fluctuations
observed in historical data for a selected period using standard time-
series techniques.

Familiar examples of stochastic processes include stock market and

exchange rate fluctuations; signals such as speech; audio and video;
medical data such as a patient's EKG, EEG, blood pressure or
temperature; and random movement such as Brownian motion or
random walks.

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 5
 Introduction

What is optimization?
Finding the best candidate from a set of
alternatives without explicitly enumerating
and evaluating all possible alternatives.
In all engineering areas, optimization has a
wide range of applications due to the large
number of decision-making situations
involved in an engineering environment.
Due to the complexity of the equations
governing chemical processes, robust
optimization methods are required to find
the best designs in terms of the objectives
of the designer

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 6
 Introduction

WHY OPTIMIZE?
 In initial design: To improve the design of
equipment to get maximum production,
maximize profit, minimize cost, improve
efficiency,
 In plant operations: To improve the
plant performance, such as improved
yields of valuable products (or reduced
yields of contaminants), reduced energy
consumption, higher processing rates,
and longer times between shutdowns.
 In maintenance: To reduce maintenance
costs, less equipment wear, and better
staff utilization
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 7
 Introduction
SCOPE AND HIERARCHY OF OPTIMIZATION

Optimization can take place at many

levels in a company, ranging:
 from a complex combination of plants
and distribution facilities down
through individual plants,
 combinations of units,
 individual pieces of equipment,
 subsystems in a piece of equipment,
or even smaller entities

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 8
 Introduction
EXAMPLES OF APPLICATIONS OF OPTIMIZATION
Optimization can be applied in numerous ways to chemical processes and
plants. Typical projects in which optimization has been used include:
 Determining the best sites for plant location.
 Routing tankers for the distribution of crude and refined products.
 Sizing and layout of a pipeline.
 Designing equipment and an entire plant.
 Scheduling maintenance and equipment replacement.
 Operating equipment, such as tubular reactors, columns, and absorbers.
 Evaluating plant data to construct a model of a process.
 Minimizing inventory charges.
 Allocating resources or services among several processes.
 Planning and scheduling construction.

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 9
 Introduction
GENERAL PROCEDURE FOR SOLVING OPTIMIZATION PROBLEMS
No single method or algorithm of
optimization can be applied efficiently
to all problems. The method chosen for
any particular case depends primarily
on:
1. The character of the objective
function and whether it is known
explicitly,

2. the nature of the constraints, and

3. the number of independent and

dependent variables.

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 10
 The analysis and solution of optimization problem

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 11
 Statement of an Optimization Problem

All optimization problems are stated in some standard format.

Essential information needs to be extracted from the problem
statement and write them in mathematical form.

Essential parts of an optimization problem statement:

 The Design/Decision variables
 The Constraints
 The Objective function
 The Process Model

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 12
 The Design (Decision) variables

 The first task of the optimization process is to identify the

underlying design variables that influence the system being
optimized.
 The design or decision variables are varied during
optimization process in order to achieve the optimization
goal

There must be some degree of freedom among the design

variables. If all the design variables are fixed, there is no scope
for optimization. Thus one or more variables must be relaxed
so that the system becomes an underdetermined system which
has, at least in principle, infinite number of solutions

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 13
 The Objective function

A mathematical optimization problem is one in which

some function is either maximized or minimized relative to a
given set of alternatives. The function to be minimized or
maximized is called the objective function

The objective function may be technical or economic

Economic Objective functions Technical Objective functions

 Maximize Profit  Maximize reactor yield
 Minimize costs of production  Minimize size of an equipment

The technical objectives are ultimately related to economics

In all cases The objective function is a function of design variables

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 14
 The Objective function
Travel plan for a long trip
Minimize ticket price, f1
Minimize travel time, f2
Single Objective? Multiple Objectives?
Minimize f = f1(x) Minimize f = w1 f1(x) + w2 f2(x)
Minimize f = f2(x) w1 +w2 = 1.0
Weighting method
Specialize algorithm
are available for
multi-objective
optimization such
as Genetic
algorithms

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 15
 The Constraints

The constraints represent some additional functional relationship among the

decision variables and process parameters. The constraints originate as
decision variables must satisfy certain physical phenomena and certain
resource limitations.

Examples:
Variable bounds: 0 ≤ x ≤ 1

Inequality constraints Equality constraints

In packed bed reactor, sum of mole fractions should be
temperature should be less unity
than maximum allowable Component balance equation in
stress . distillation column, blending
Acidic condition: pH < 7 process.

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 16
 The Process Model

A process model is required that describes the manner in which the

decision variables are related. The process model also tells us how
the objective function is influenced by the decision variables.

A model is a mathematical equation or is a collection of several

equations that define how the decision variables are related and the
acceptable values these variables can take

Optimization studies are carried out using a simplified (but

reasonably acceptable) model of a real system. Working with real
system is time consuming , expensive, risky.

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 17
 Problem Statement
How the problem statement should be written

Given a design vector: x = [x1, x2, …,xn]T

An objective function, f(x) The optimization problem
is to:
A set of inequality constraints, g(x) ≥ 0
“find the values of the
A set of equality constraints, h(x) = 0 decision variables that
The general problem statement: min f (x) minimize the objective
function while satisfying
Subject to the constraints
G (x) ≥ 0
h(x) = 0
LB ≤ x ≤ UB

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 18
 Problem statement

Given a design vector: x = [x1, x2, …,xn]T

min f (x) max f - (x)

Subject to Subject to
G (x) ≥ 0 G (x) ≥ 0
h(x) = 0 h(x) = 0
LB ≤ x ≤ UB LB ≤ x ≤ UB

Equivalent

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 19
 Classification of Optimization problem
Classification based on present / absence of constraints:
 Constrained optimization problems: which are subject to one or
more constraints
 Unconstrained optimization problems: in which no
constraint exist

Classification based on the nature of equation involved:

Linear programming (LP), Nonlinear programming (NLP), Quadratic
programming

Classification based on the nature of the decision variables:

Continuous optimization, integer programming (IP), MILP, MINLP

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 20
 Classification of optimization methods
Classification based on search space:
Local search methods, Global search Methods

Classification based on the number of objective function:

Single objective optimization, Multi-objective optimization

Classification based on the type of model:

Deterministic optimization, Stochastic optimization

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 21
 Examples of engineering applications

The theory of optimization finds application in all branches of

engineering. The areas of applications may be broadly divided
as follows:

 Design of system component or entire system

 Planning and analysis of existing operations

 Engineering analysis and data reduction

 Control of dynamic systems: optimal control/dynamic optimization

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 22
 Optimal Design of a can

Design a can which will hold at least 500 ml of liquid.

Height = [7 12] cm , Radius = [3 7] cm

What dimensions for the cylindrical can will use

the least amount of material?
The material used can be minimized by
minimizing the area, A? h
Objective function: Constraints:
A = 2πr2 + 2πrh V = πr2h ≥ 500 ml
Lateral
r
Area of
two ends Area Bonds:
3 ≤ r ≤ 7, 7 ≤ h ≤ 12

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 23
 The optimal insulation thickness
Economic Criteria:
The addition of insulation should save
money through reduced heat losses.
But the insulation material can be expansive
The optimal thickness of insulation
needed can be determined by an
optimization technique.
The critical insulation thickness:
RTotal = Rconduction + RConvection
The conduction resistance (first term)increases
with the addition of insulation. However, the
convection resistance (2nd term) decreases due
to increasing outer surface area.
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 24
 Chemical Reactor Design: Series Reaction
𝑘𝑘10 𝑒𝑒 −𝐸𝐸1 𝑟𝑟𝑡𝑡
A→B →C How to maximize , CB (tf)
k1 k2
𝑘𝑘20 𝑒𝑒 −𝐸𝐸2 𝑟𝑟𝑡𝑡

Material Balance:
𝑑𝑑𝑑𝑑𝐴𝐴
=-𝑘𝑘1 𝐶𝐶𝐴𝐴
𝑑𝑑𝑡𝑡
𝑑𝑑𝑑𝑑𝐵𝐵
=-𝑘𝑘1 𝐶𝐶𝐴𝐴 - 𝑘𝑘2 𝐶𝐶𝐵𝐵
𝑑𝑑𝑡𝑡
𝑑𝑑𝑑𝑑𝐶𝐶
=-𝑘𝑘2 𝐶𝐶𝐵𝐵
𝑑𝑑𝑡𝑡

Energy Balance:
𝑑𝑑𝑇𝑇
ρvCp = (-ΔH1)K1CAV + (-ΔH2)K2CBV – UA (T - TJ)
𝑑𝑑𝑡𝑡

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 25
 Chemical Reactor: Parallel Reaction

Let’s look at the reaction, which produce a desired product B

B and undesired product C in a parallel reactions. A
C
Undesired side-product usually needs to be separated
B

C
Higher conversion to B will reduce the
separation cost, however increases the cost
of reactor. What is the optimum solution

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 26
 Refinery Blending Problem

Petrol Octane Price How much of

Type rating (Rs/liter) “Premium
petrol” and
Premium 93 85
how much of
Regular 85 80 “Regular
Petrol” should
Raw Octane Price Arability (Kilo be produced
Oil rating (Rs/liter) liter to maximize
the profit?
1 70 40 200
2 80 50 450 This is an
3 85 60 450 example of
4 90 70 500 Linear
Programming
5 99 75 350
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 27
 Nonlinear Regression

Redlich kwong equation

From a series of available experimental P-V-T data for a particular

gas, find the unknown constants a and b by nonlinear regression for
that gas

Exp.No T (K) V(CC/g mol) P (atm)

Minimize the objective
1 273 500 32 function to find a and b:
2 325 500 42 Error =
3 350 600 45 ∑𝑁𝑁
𝑖𝑖=1(𝑃𝑃𝑖𝑖,𝑒𝑒𝑒𝑒𝑒𝑒 - 𝑃𝑃𝑖𝑖,𝑐𝑐𝑐𝑐𝑐𝑐 )
2

N 370 700 40

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 28
 Review of Linear Algebra
In linear optimization problems, we encounter systems of linear
equations, therefore, a short review of linear algebra is necessary
before starting the course of optimization
For example, consider the problem of solving the following system of
three linear equations in the three variables x1; x2; x3 Є R:

One way to approach this problem is introduce the “vectors" u, v,w,

and b, given by

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 29
 VECTORS
Vector is directed line segment in N-Dimensions it has both “length” and
“directions”.
Vectors in RN is an ordered set of n real numbers.

V = (1,6,3,4) is in R4
1
Column vector = 63 Row vector = (1 6 3 4 )
4
The transpose of a column
vector is a row vector
T 𝑎𝑎
𝑎𝑎
=(a b) v = 𝑏𝑏
𝑏𝑏
𝑐𝑐

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 30
 Vector addition and vector subtraction

𝑢𝑢1 𝑣𝑣1 𝑢𝑢1 + 𝑣𝑣𝑣

U+v= + =
𝑢𝑢𝑢 𝑣𝑣𝑣 𝑢𝑢𝑢 + 𝑣𝑣𝑣

𝑢𝑢1 𝑣𝑣1 𝑢𝑢1 − 𝑣𝑣𝑣

U-v= - =
𝑢𝑢𝑢 𝑣𝑣𝑣 𝑢𝑢𝑢 − 𝑣𝑣𝑣

The difference of two vectors is result of adding a negative

vector
A – B = A + (-B)

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 31
 Properties of Vector Addition

1.Commutative property = A+B = B+A

2.Associative ( A+B) +C = A + (B+C)

3.There is a ZERO vector 0 = [ 0,0,……….,0]T Such that

A+0=0+A=A

Note That:
1. B + (A - B) = A
2. - (- B) = B
3. - (A - B) = (B - A)

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 32
 Vector Multiplication by Scalar

αv = α (x1, x2) = (α x1, + α x2)

v
αv Properties:
1. Distributive : α(A+B) = αA + αB
(α+β)A = αA + βA
2 . Associative : α (β A) = (α β) A

The scalar 1, α, 0, -1 satisfies:

1A = A α0=0
(-1)A = -A 0 α = 0

31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 33
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 34
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 35
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 36
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 37
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 38
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 39
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 40
31 May 2020 Dr. Saeed GUL, Department of Chemical Engineering, UET Peshawar, Pakistan 41