Quality in Design - Medical Device Design: Innovation from Concept to Market - page 122

Target:

995V15liter 00 0 . .

M minimized
If we create a table of possible combinations then we can start to build a picture. If we start
with four values of diameter we will have corresponding lengths (from the volume
criteria/objective function), and as a consequence corresponding cylinder masses (cylinder mass
objective function).
Table 7.1 shows that as the diameter of the cylinder increases the length required decreases (as
expected). A quick scan of the table illustrates that the optimum solution is to have a diameter of
50 mm and a length of 50.7-51.2 mm, which yields the correct volume and the minimum mass of
about 0.47 kg.

Table 7.1: Optimization Table

Diameter D Volume V Length L Cylinder Mass M

(mm) (liter) (mm) (kg)

25 0.995 202.8 0.66

25 1.005 204.8 0.67

50 0.995 50.7 0.47

50 1.005 51.2 0.47

75 0.995 22.5 0.55

75 1.005 22.8 0.56

100 0.995 12.7 0.77

100 1.005 12.8 0.77

Clearly this was a very simple problem. But it illustrates the power of optimization. For
more complex systems this method will not work and you will need to use one of the many
techniques available (such as linear programming, Routh-Hurwitz, Monte Carlo method,
etc.). Most modern computer-based mathematics programs contain optimization routines, but
you must define the parameters. There is no need to purchase a program - many open-source
programs can be found on the web. If you have access to Microsoft Office® (Microsoft,
2011) then you have their optimization routine called solver (under Tools - Solver in Excel).
Use of this routine reveals the result shown in Table 7.2 .

Most computer-aided design (CAD) packages, such as Solidworks® and ProEngineer®,
come with built-in analysis that enables you to perform design optimization from the very
solid models you are drawing. There is really no excuse not to undertake some form of
optimization.