CHAPTER ONE DESIGNING A WORKABLE SYSTEM It is often said that there are many possible answers to a design problem

and the idea conveyed is that all solutions are equally desirable. Actually, only one solution is the optimum based on a defined criterion (e.g. cost, size or weight). However, it is important to note that a workable system is more preferable to a non workable one and the cost and or time expended to move from a workable system to an optimum one may not be justifiable. A workable system is one that meets the requirements of the

purposes of the system, have satisfactory life and maintenance costs and abides by all constraints, such as size, weight, temperature, pressure, noise, pollution etc.

1.1 -

Steps in arriving at a workable system Select the concept to be used and Fix the parameters necessary to select the components of the system (so that design requirements and constraint are satisfied)

The two major steps in achieving a workable system are

1.2

Designing a workable system

In selecting the design concept to be used to solve a problem, the engineer should not just settle for the concept used on the last similar job but should be creative, review alternative concepts in a manner appropriate for the scope of the project and also review old ideas that were hitherto considered impractical or uneconomical. In designing a

workable system, the objective of the design is meeting the specified design criterion and not that of considering other alternative designs and choosing the optimum. In this instance, the initial cost, lifetime and maintenance costs for the design (which will be considered for an optimum design) are not considered. electricity is For instance, if 1500watts of thermal power required to satisfy a communitys power requirement,

then a workable design would be that of building a

plant (gas powered) of about 1800watts output for the community. But an optimum design would be considering the initial, maintenance and operating cost of this solution and comparing it with other alternatives like, hydropower plant, geothermal power plant, solar power generation etc, and choosing the one that is more economical and environmentally friendly. From the foregoing, there is a strong preference towards, designing optimum systems but other considerations like the size of the job, the cost of increased engineering time required for optimization, the pressure of calendar time, etc, may not permit the design to proceed beyond a workable design. Every engineering design embarked upon should at least be a workable one.

CHAPTER TWO OPTIMIZATION 2.0 Introduction

Optimization is the process of finding the conditions that give maximum or minimum values of a function. In the design of workable systems, often certain parameters are arbitrarily assumed and individual components are selected around these assumptions but in design optimization, the parameters are free to float until the combination of parameters is reached which optimizes the design. The high cost of engineering time required to optimize small projects may not justify the effort. However, for complex systems, it may be possible to optimize sub systems and then choose the optimum combination of them. Central to any optimization process is deciding on which criterion is to be optimized. Minimum cost is the most common criterion but in some instances, the manufacturer because of human, social and aesthetic reasons may turn aside from the most economical solution. For instance, in an air craft or space vehicle, minimum weight may be the criterion whereas in an automobile, the size of the system may be the criterion.

Optimization activities are often practiced under the name of operations research. Many developments in operations research emerged from attempts to optimize mathematical models of economic systems, but recently mechanical and chemical engineers have used it to optimize fluid and energy-flow systems. Optimization has two levels-

comparison of alternate concepts and optimization within a concept. However, a complete optimization procedure consists of proposing all reasonable alternate concepts, optimizing the design of each concept and then choosing the best of the optimized designs.

2.1

Mathematical representation of optimization problems

The mathematical statement of optimization entails specification of the function (the objective function) and the constraints. If y represents the function to be optimized and y is a function of
x1 , x 2 , ... , x n , which are called the

independent variables. The objective

function, then is
y = y ( x1 , x 2 , ... , x n ) optimize

(2.1)

In many physical situations, there are constraints some of which may be equality constraints
1 = 1 ( x1 , x 2 , ... , x n ) = 0 m = m ( x1 , x 2 , ... , x n ) = 0

(2.2) (2.3)

as well as inequality constraints

(2.4) (2.5)

The physical conditions dictates the sense of the inequalities in equations (2.4) and (2.5).

An additive constant appearing in the objective function does not affect the values of the independent variables at which the optimum occurs. That is, if
y = a + y ( x1 , ... , x n )

where a is a constant, the minimum of y can be written

(2.6)

A further property of the optimum is that the maximum of a function occurs at the same state point at which the minimum of the negative of the function occurs, thus
max[ y ( x1 , ..., x n )] =

[ y( x1 , ... , xn )]

(2.7)

In the optimization of thermal systems, establishing the objective function is often simple usually in the form of equation (2.1.) to (2.5). The challenge usually arises in writing the constraints. A strategy that is often used is to specify all the direct constraints, such as requirements of capacity, limitations of temperature and pressure etc. describe in equation from the component characteristics and properties of working substances, and write mass and energy balances.

Operations (1) and (3) usually provide a set of equations containing more variables than exist in the objective function. The set of constraint equations is then reduced in number by eliminating the

variable that do not exist in the objective function. It should be pointed out however that some optimization techniques permit additional variables in the constraints that do not exist in the objective function. Steps (2) and (3) in setting up the constraints are similar to the process used in structuring a system- simulation problem.

2.2

Optimization Procedure

In the optimization of systems, it is almost axiomatic that the objective function is dependent upon more than one variable. In fact, some thermal systems may have dozens or even hundreds of variables which demand sophisticated optimization methods presented is not exhaustive.

2.2.1 Calculus methods- Lagrange Multipliers The basis for using this method is the ability to extract derivatives of the objective function and constraints. The method of Lagrange multipliers performs an optimization where equality constraints exist but cannot directly accommodate in equality constraints. 2.2.2 Search Methods These methods involve examining a number of combinations of values of the independent variables and drawing conclusions from the magnitude of the objective function at these combinations search methods generally fall into elimination and hill-climbing techniques and are the ultimate approach when other methods of optimization fail. When applying these method to continuous functions, since only

discrete points are examined, the exact optimum can only be approached, not reached, by a finite number of trials.

2.2.3 Dynamic Programming This method is unique in that the result is an optimum function rather than an optimum state point. The problem handled by dynamic programming is one where the desired result is a path, e.g.; the best route of a gas pipeline. The result is therefore a function relating several variables. Dynamics programming does in a series of discrete processes what the calculus of variation does continuously.

2.2.4 Geometric Programming This is the youngest member of the optimization or programming family. It optimizes a function that consists of a sum of polynomials wherein the variables may appear raised to integer and non-integer exponents. A feature of this method is that the first stage of the solution is to find the optimum value of the function, rather than first to determine the values of the independent variable that give the optimum.

2.2.5 Linear Programming This is a widely used and well-developed optimization method applicable when all the equations (as in (2.1) to (2.5)) are linear. The constraint equations may be equalities or inequalities. It is used in

solving the blending, machine allocation, inventory and production planning and transportation problems.