29-12-2012, 06:07 PM
TRADEOFF ANALYSIS FOR HELICAL GEAR REDUCTION UNITS
TRADEOFF ANALYSIS FOR HELICAL GEAR REDUCTION UNITS.pdf (Size: 556.72 KB / Downloads: 34)
Abstract
Multicriteria optimization plays a very important role in product development. During
the product concept phase, the designer decides a set of objectives that needs to be optimized.
Every designer strives to achieve a utopian design that will optimize all the objectives. In real
life, such a design is impossible to achieve. One of the reasons is the conflicting nature of
objectives that need to be optimized. The designer has to arrive at a compromised design that
will satisfy the requirement to the greatest extent. Compromise involves tradeoffs between
efficiency, cost, quality and other attributes.
The method presented here shows the implementation of a multi-criteria optimization
technique. Various techniques that are used are presented and later one of them is illustrated
using a gear train optimization problem. Numerous solutions are obtained by the analysis and
they all are optimal solutions. To choose between the designs is one of the difficult tasks and it
depends on the designer’s judgment. There are available methods that provide valuable
information about each design, and it helps the designer in making correct decision. The
proposed study incorporates one such decision technique.
Design Process
A design process is a process of determining structural properties which includes geometry and
material that will satisfy all the functional requirement of a prototype. There are numerous
processes available that help the decision-maker in deciding upon a good design that will fulfill
most of the requirements. The basic steps in a design process are shown in Table 1-1. The
starting step, Conceptualize the requirements (step 1) usually consists of some imprecise
statements of requirements. Research (step 2) on the possibility of developing such a product is
necessary, after which it is possible to restate the Objective (step 3). Constraints (step 4) are
formulated to bind the design process from generating infeasible designs. The Formulation and
Solution (step 5) of the problem is done using the objective and constraints and various
alternative solutions are obtained. In the next step, all the possible solutions are Analyzed (step
6) and weighed against some reference to either accept, reject or modify them. The preferred
solution is Selected at step 7. An Engineering Drawing (step 8) of the resulting solution is done.
The next step is Prototyping (step 9) followed by Production (step 10) [1].
Optimization
Optimization is a process of finding the best solution which is subjected to specific
circumstances. This is an integral part of design process. During the design process, the
decision-maker sets up an objective which is decided according to the functions to be performed
by the product, e.g., an objective can be to maximize the heat flow through the pipes of a heat
exchanger. This objective is then expressed in terms of some parameters, the results are
evaluated against some criteria, and a solution is obtained that will achieve the desired
objectives. In other words, optimization is a process of obtaining the design parameters that will
maximize or minimize the objectives set by the designer.
Single Objective Optimization
Introduction
After the design of the product is conceptualized, a mathematical model is formulated to find
out various design alternatives. Using the model, we can understand the performance
requirements and also the extent to which the product will satisfy these requirements. The
solutions that satisfy these requirements are called the feasible set of solutions. The design
problem enters the next phase of choosing between alternative solutions. The merits and
demerits of all the feasible solutions are analyzed and most suitable solution is chosen [3]. Since
in case of single objective optimization problem, there is only one criterion against which all the
solutions are evaluated, so there is only one solution if the problem is convex else there is one
local minimum.