Materials & Process Selection: Selecting a Process
What is Process Selection?
The problem of determining the best way of performing a task is one that is found in many
walks of life. Whatever the type of problem, the aim is usually similar - to
select the best way of fulfilling a set of requirements from the
options available.
The field of engineering design is no exception. For many reasons, such as
the pressure to reduce cost and time to market, systems are needed to help with the choices that
arise on the route from an initial idea for a product to its realisation in the marketplace. The
designer needs a systematic and robust way of evaluating the available options and
identifying (without bias) which might be the best.
In brief, manufacturing process selection is the task of choosing a method for transforming a
set of materials into a given shape using one or more
processes. The 'best' process is normally considered to be the most economic,
subject to it meeting the technical constraints. The aids to selection tend to fall into one of
three main categories:
- Methodologies - are general approaches to tackling a large variety of
problems. They are not aimed at ways of selecting a particular process, but at creating frameworks
within which certain classes of problem can be addressed. The intention is to guide the creation
of procedures, ensuring that all the relevant aspects of selection are addressed.
- Procedures - can be considered as algorithms for tackling a set of related
process selection problems. These will contain all the necessary steps and the appropriate tools
for selecting a process in a given situation. The way in which the procedure is actually used to
carry out a selection will depend upon its implementation. Implementations are usually
software based, although they can be as simple as a series of instructions and charts.
- Tools - used in selection procedures usually help to answer single specific
question. Examples include charts of properties, linear regression analyses, databases and all
types of modelling. At the implementation stage it becomes possible to utilise more generic tools
such as fuzzy logic and expert systems.
The choice of approach depends, to a large extent, on the degree of detail required of the
answers. At the preliminary stages of design generic selection tools that have a broad scope, but
use low precision data, are usually most appropriate. As the design proceeds, and the problem is
specified in increasing detail, more focused selection tools are required. At the most detailed
levels of design, the selection field becomes so constrained that design optimisation
becomes more relevant.
Currently, selecting a manufacturing process relies heavily on in-house experience. Whilst
this knowledge and expertise will never be completely superseded by design tools, there are
several major inherent problems with relying solely upon it:
- There will be a bias towards existing technologies and, as a result, new processes may be
overlooked.
- Existing knowledge may not always be helpful when new processing conditions arise. Using
selection tools, such as modelling, suitable processing conditions can be predicted more rapidly
and accurately.
- Expertise can be difficult to rationalise or quantify. This makes the design process difficult
to document and does not provide a route to reliable and repeatable decision making.
Process selection, then, is intended to provide a systematic and robust way of
evaluating the available processing options and identifying (without bias) which might be the
best. There are many approaches that can be taken, which will depend on the level of detail
required; some of these are discussed below.
A more in-depth discussion of some of the selection issues that arise in engineering design,
and in particular process selection issues, can be found in the introductory chapters of a PhD
thesis by AM Lovatt (Cambridge University Engineering Department, 1998).
For links on current research go to:
| Materials Group
| EDC
| Granta Design |
| MMD
| CUED
| University |
2001-2003 ©
CUED
www-mech@eng.cam.ac.uk