Civil Engineering Reference
In-Depth Information
From a first standpoint, the reasons for this can probably be found in the different economic stakes which
are more oriented toward mass production rather than security parts, in the internal structure of the
companies and their scientific background, as these companies have moved more recently toward com-
puter-aided techniques. From another standpoint, this delay is not amazing. In fact, non-steady-state metal
forming problems are much more complex. Until the beginning of the nineties, there was no actual robust
and reliable method to study them. Efficient simulation softwares have emerged only recently, making it
difficult to study the design optimization problems before that time. However, this presentation would
not be fair without mentioning the pioneer work of Kobayashi et al. [57], who were the very few scientists
handling these issues in the eighties. So, in a certain way, for non-steady-state metal forming processes,
optimal design techniques is actually a new field of study.
Presentation
The main objectives of this contribution are to first give a survey of the state of the art of the research
work which have been carried out in computer aided design techniques for forming processes. Then the
more promising techniques will be described, with a special focus on the gradient based optimization
methods. Finally, some of the preliminary and very encouraging optimization results will be shown,
along with what can be expected in a near future.
As the field of metal forming processes covers a wide domain, it would be impossible to address all the
applications in a single chapter of this topic. We shall then focus on a rather specific process, forging.
However, it must be pointed out that forging encompasses a lot of different processes and a wide range of
alloys: from unit parts to mass production, from very small and light parts to parts of several tons, from
economic rough preforms to elaborated security parts. Moreover, the forging process can be considered as
one of the most complex process, at least from the numerical standpoint. So, it is believed that the techniques
that are described in this chapter can be easily extended to most of the other forming processes.
5.2
State of the Art of Design Techniques for Non-Steady-State
Metal Forming Processes
Basic Design Techniques
The forging problem can be summarized as to find the best process design, which makes it possible to
form the required part. We do not consider the preliminary problem of dressing. We assume that the
forged part has already been designed out of the shape of the final part, which will often be obtained
after machining. The problem is just to design a deformation path for the material in order to obtain a
prescribed shape with satisfactory properties and for the lower cost. Of course, the deformation paths
are not unique, and that can include several forging operations. So the problem is to find the best shapes
of the preforming tools, the best forming temperatures (both the initial temperature of the billet and the
temperature of the dies), the best forging velocities, the best lubricant and lubrication conditions, etc.
The optimality conditions regard several parameters which may depend on the process itself. However,
most of the time, the optimal forging design has to obtain the right final part without major defects such
as folds, to minimize the number of operations, to minimize the size of the initial billet, to reduce the
maximum forging force during the operations, and so on.
This is a complex design problem as often the material flow is three-dimensional and difficult to
foresee. It is not possible to simplify it into a less complex problem which could be more easily studied.
In fact, in this area, there is a deep lack of simple mathematical models. Just a few problems can be analytically
solved, such as the upsetting of a cylinder (or a bar) between flat dies without friction. Although it provides
some interesting tendencies, it is far too simple to be useful for the design of a close-die forging process. So,
the industrial practice was mainly based on thumb rules and empirical knowledge which have been obtained
either by actual experiences, by reduced scale or simulation experiments, or by more complex mathe-
matical models [51, 2].
