Dies Manufacturing

ABSTRACT
To reduce the cost of laminated dies manufacturing, the amount of the extra
material and the number of slices must likewise be reduced. This is considered
an optimization problem, which can be solved by genetic algorithms (G.A.).
A new niching and weak elimination crowding method is presented in this
paper. Using the proposed method, results show not only a significant
improvement in the quality of the optimum solution but also a substantial
reduction in the processing time.

INTRODUCTION
 Laminated tooling is based on taking sheets of metal to produce the final product, after
cutting each layer profile using laser or other techniques. CNC machining removes the
extra material and brings the final product to specific tolerances.
 Usually CNC machining, to remove the extra material and finish the surface and bring it
to desirable tolerances.
 Laminated tooling, as a manufacturing method, uses Computer Aided Design (CAD) to

produce tools.
 This technology involves designing the tool within 3-D CAD software, slicing the model
into cross sections, cutting off the 2-D profiles for each slice, bonding the sheets
together and finally a post processing stage
 It is obvious that having smoother surface after joining the sheets means .
 Therefore, an optimization problem is defined to find the best sort of
thicknesses to minimize the number of slices and the volume deviation at
the same time.
 Rapid prototyping and laminated tooling are two manufacturing methods

that have almost the same process steps mentioned in the previous
paragraph.
 In rapid prototyping, the RP machine can create slices with any

thicknesses within a boundary, based on the RP machine capability.
 In laminated tooling, all the thicknesses must be selected from a set of

available sheet thicknesses.
Parallel genetic algorithm
Parallel genetic algorithms typically divide the solution space into sub populations.
This helps the system to consider different areas of the solution space.
Consequently, the searching performance is increased with this modification.
Moreover, the calculation process for each subpopulation can be done on different
processors
 Although parallel genetic algorithms improve searching
performance, having several powerful processors available at the
same time is not always possible.
 This is the disadvantage of the parallel method, which makes it

unsuitable for the aim of the current research.
NICHING TECHNIQUE
To maintain diversity in the population, there are other methods which are mostly
categorized as niching techniques.
 These methods divide all members of the population into groups based on their
fitness values.
 Each group, which is named as a niche, represents a group of individuals with almost
the same characteristics.
Some of the most important niching method is fitness method,

sequential niching method, restricted tournament.
 But this technique suffer from the timing problem, so we need some other
technique to solve this.
Weak elimination crowding method
The weak elimination crowding is the new crowding method which
has been designed for the purposes of this research.
It gives the system more options for creating new thickness vectors by
injecting standard thicknesses into each generation of the population.
In this crowding method, fitness values are calculated

for all members of the current population as in regular genetic
algorithms.
Then all members are ranked based on their fitness

values.
The N weakest members of the population are then
replaced by N new randomly created members
N = Cf .M
Where,
Cf = is the user-defined crowding factor value.
• This value represents the magnitude of the diversity in the population,with

M the population size.
 Schematic of the modified genetic algorithm procedure in
the current research.
Table 1.
Different cases in optimization
manufacturing.
 Above fig. shows the value of volume deviations for all thickness vectors in
the first generation and the last generation.
 As depicted in fig. a, in the first run, the optimization process can create thickness
vectors that reduce the amount of volume deviation
 This figure also shows that, at the time of termination (the 20th iteration), almost
all thickness vectors produce the same volume deviation.
 In the next run, the weak elimination crowding method is employed to improve the
search technique.
 Fig. (b) shows diversity in the population in the last generation.

 As depicted in Fig. c, there is another improvement in the quality of the solution in
this case, as the best set creates a lower amount of V compared to the first and
second runs.
 Fig. shows that, during the different generations of the three runs, the minimum
value of volume deviation is getting closer to the ideal value.
Comparison of fitness values
 As depicted in Fig. 1 the highest fitness values are improved through the
generations, in each run.
 since a larger population is used for the third run, as a result the maximum
fitness value is improved in the third run.
 The figure shows how the weak elimination crowding method helps the
optimization process reach better results.
Conclusion
 A modified genetic algorithm was used to find the best locations of slice planes in a
laminated manufacturing process.
 Premature convergence, one of the main problems in most genetic algorithm

applications can now be prevented by a newly developed crowding method.
 For the aim of this research, a weak elimination crowding method was designed.
This technique significantly reduces the processing time for finding the optimum
thickness vector.
 The modified genetic algorithm was applied to a CAD model to show the
performance of the proposed optimization method.
Thank you!!!

Dies Manufacturing

Uploaded by

Copyright:

Available Formats

Dies Manufacturing

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dies Manufacturing

Uploaded by

Copyright:

Available Formats

ABSTRACT

an optimization problem, which can be solved by genetic algorithms (G.A.).

A new niching and weak elimination crowding method is presented in this

improvement in the quality of the optimum solution but also a substantial

reduction in the processing time.

 Laminated tooling, as a manufacturing method, uses Computer Aided Design (CAD) to

 Rapid prototyping and laminated tooling are two manufacturing methods

 In rapid prototyping, the RP machine can create slices with any

 In laminated tooling, all the thicknesses must be selected from a set of

Consequently, the searching performance is increased with this modification.

 This is the disadvantage of the parallel method, which makes it

Some of the most important niching method is fitness method,

In this crowding method, fitness values are calculated

Then all members are ranked based on their fitness

• This value represents the magnitude of the diversity in the population,with

 Fig. (b) shows diversity in the population in the last generation.

 Premature convergence, one of the main problems in most genetic algorithm

You might also like