1. INTRODUCTION
When developing materials for new additive manufacturing
applications, it is necessary to use analysis tools to prioritize and to
discriminate the most influencing variables over the final properties,
throughout the studied process. From the quality assurance point of
view, the basic criterion to evaluate is customer satisfaction. Also it
is mandatory to introduce an optimization approach throughout the
process in order to strategically prevent waste and to create value at
every step of the process. Value at Materials Development for Additive
Manufacturing can come from satisfying customer requirements or being
producible at the selected AM technology, by reducing post processing
tasks or by improving characterization reliability.
The rapid expansion of additive manufacturing has not been
accompanied by the development of process standards that assure its
correct and reliable application (Grimm, 2004). In this sense, the lack
of standardization is one of the barriers this technology has to
overcome to be used as production alternative. This work is based upon
the development of a new Alumina powder based system intended to be used
at 3D printing technique, for the printing of refractory casting moulds.
The new material's development cycle has been analyzed from the
point of view of a product development process. In a first approach the
entire cycle has been deployed into its sub processes in order to
identify the influencing variables over the final part. We wanted to
base the qualification of the desired results over the customer demands
for the final application. For this purpose we have applied the Quality
Function Deployment method to deploy the quality requirements linked by
the sub processes correlations (Revelle, 1998). This analysis lead to a
higher understanding of the variables involved and showed the
convenience of implementing standard methods. By assuring controlled
conditions trough the entire fabrication process of the material,
variability of the resulting properties can be kept under minimum
levels.
2. CASE STUDY
This study has been developed with the additive layered deposition
of ceramic powder technique. This technology was developed at MIT in
1989 as a 3DPrinting technique for prototyping. The Z310 3DPrinter used
joins the particles together with a liquid binder that is selectively
deposited to the powder through a conventional ink-jet printing head.
The ejection of the liquid binder follows the sliced two dimensional
profile of a computer model. The subsequent stacking and printing of
powder layers on previously printed layers generates the complete three
dimensional structure of the desired object. The function of the binder
is to join adjacent powder particles of the same and neighbouring layers
(Khalyfa, 2007).
The materials system studied was required to be at one hand, a
mixture of powder components able to be agglomerated and which after
processing would exhibit properties of a refractory castable system. As
the substrate for this mixture, it was selected electro-fused alumina,
Alodur[R] WSK (Treibacher Schlaifmittel, Italy) at refractory grades:
F360 & F600 according to FEPA. For binding purposes at sintering, it
was chosen hydratable alumina (Alphabond 300, Almatis, Netherlands).
This component works as hydraulic binder at high temperatures and it is
calcium free, which enables a higher compaction of the ceramic matrix
when sintered. As far as it interacts with water, provides hydrophilic
properties to the mixture which contributes to the printability of the
system (Goberis, 2004). To ensure the low temperature layer-by-layer
consolidation, it was decided to add Dextrin from Potato Starch, a
widely used binder. On the other hand, the aqueous dissolution employed
to activate the binding mechanism was derived from the commercial
dissolution provided by machine manufacturer. It was composed of
Glycerol and anionic surfactant Triton X-100 (Sigma-Aldrich, Germany).
The objective of the study is to obtain and to validate a materials
system to be properly printed and to comply with final application
requirements of customer. Beginning with the raw materials described
above, the study came throughout the fabrication process until the
finished tool was ready. As far as it is not possible to evaluate
tooling performance until the tool is used, the evaluation integrates
all the steps of the process:
* Powder Material Preparation: Alphabond 300 (1%wt), Dextrin (9%wt)
and Alumina Powder (90%wt) at refractory grades (F360, F600).
* Proprietary Binder Preparation: Aqueous dissolution prepared from
surfactant Triton X-100 (1%wt) and Glycerol (9%wt).
* Layered Manufacturing: Prepared materials were charged into the
Z310 3Dprinter. STL file obtained from Solid Works software was charged
into controlling ZPrint software. Then required parts were printed, with
fixed layer thickness (0.2 mm). Saturation was an evaluated parameter.
The parts were left at 40 [degrees]C during 20 hours and then were blown
at 0.5 bar to eliminate residual powder.
* Infiltration. The parts were immersed into a liquid cianoacrylate
bath (ZCorp, USA) at ambient temperature during 20 seconds.
* High Temperature Sintering. The parts were sintered at a Hobersal
XG3-16 high temperature furnace. This is the final step to achieve final
mechanical properties of the refractory castable mould. Sintering
parameters were studied during this work.
3. CASE ANALISYS
In order to study the development process in a structured way that
enabled us to focus on the desired results, and to introduce an
optimization approach which creates value through the entire process, it
was applied the Quality Function Deployment Method. The main approach of
QFD is to include the voice of customer as the cornerstone to design our
process parameters and quality assurance metrics. For this issue it was
launched a survey addressed to target customers whom gave their point of
view about their actual needs in relation to the refractory casting
moulds. In order to promote a structured insight of the different
factors and variables affecting the different sub processes, and to
obtain the input data to feed QFD analysis there has been applied
several tools inspired at the Design for Six Sigma Methodology--which
purpose is to identify, quantify and eliminate or control the sources of
variation of the key properties measured. First, it was developed a
workflow chart in order to have a schematic insight about the functional
steps of the process. It was found that four principal sub-processes had
to be considered: preparation of the printing materials, printing,
post-processing of the part and final performance of the part. As a
control step it was applied a characterization of the material obtained,
through test specimens and standard testing measurements, before the
final 3D solid was printed. If the material obtained complied with
required quality, then it would be printed and applied for a definitive
investment casting mould. Once the process has been mapped, the workflow
chart obtained was used as the basis for a deeper analysis of the
variables involved. Taking inspiration in the lean approach present at
Six Sigma philosophy, it was decided to continue the analysis of the
process with the SIPOC diagram. Here it is considered the specific
information related to Suppliers, Inputs of the process, Sub process
Steps Deployment, Outputs required and specially the voice of customer
is the pulling criteria explicitly expressed at this tool. The insights
obtained were multiple because they allowed observing the process from
the objectives point of view and this is a key factor to discover,
eliminate and prevent further waste. It were also deployed the sub
processes and it was applied the SIPOC tool to analyze sub processes
requirements and variable relationships. The deployment of every sub
process obtained by SIPOC was the base for standards creation. Affinity
diagram was a powerful tool to promote deep understanding about the
factors that can be generating undesired effects. Within this study it
was used to deploy and to organize the possible causes of trouble
observed through the development: heterogeneously agglomerated mixture
appearance, poor powder packing, bleeding, etc. The possible reasons
listed were useful to identify noise factors that can be minimized by
following standard procedures. Finally the chosen properties to be
measured were related to the independent variables possibly affecting
the results through the use of the cause effect diagram.
The use of Quality Function Deployment (QFD) enabled to determine
specific measurable parameters coming from two important inputs, the
final part user specifications listed as desired quality and the layered
manufacturing specific process requirements. In the case of materials
development, there is the particularity that the raw materials have to
be processed in several steps before the final application can be
achieved. To apply the deployment in this case it must be used a cascade
series of matrices which are linked by the dependent relationships of
the variables through the subsequent sub processes.
At the first matrix were introduced the customer requirements in
the "customer words", then it was deployed the first level of
process properties related to mechanical characterization of the
sintered part, weight, fabrication method, lineal dimensional variation.
After, it was deployed the next level of process properties where it was
proposed how to achieve that mechanical properties through process
controlled conditions like sintering temperature, sintering time,
density of sintered part, linear dimensional variation, printability of
the material, etc. The next step was to deploy how to achieve that
sintering temperature, density of sintered part, printability of the
material system, etc. It resulted on the measurement of prepared
materials properties like BET specific surface for the powder mixture,
viscosity and surface tension for liquid binder mixture. This deployment
is related to the quality control of 3D printing materials system
preparation and process conditions of the layered manufacturing. Finally
the last matrix reflected the need to determine standard procedures to
control variability over the process since the former steps. The
proposed standards procedures were as follows: Infiltration, BET Surface
measurement of the mixture, preparation of liquid binder, preparation of
powder mixture, measurement of linear dimensions variations, within
several other procedures.
Finally it has been selected seven independent variables to be
studied on an experimental basis in order to determine quantitatively
their influence over mechanical properties, dimensional variation and
density of the materials system developed. It was proposed a
Plackett-Burman Design of Experiments to obtain a statistical indication
of the main variables influencing the results.
4. CONCLUSION
It is possible to reduce variability and to optimize the final
properties required for a new materials system, through the adoption of
a structured analysis of the development process. In this study it was
decided to implement standard procedures to eliminate variability and to
reduce possible noise over the performance properties of the layered
manufactured part. The analysis has been conducted using Quality
Function Deployment approach and quality engineering tools like process
mapping, SIPOC diagram, affinity diagram. The analysis provides a
reliable base of inputs for quantitative optimization of final
properties to be studied on further research.
5. REFERENCES
A. Khalyfa, S. Vogt, J. Weisser (2007). Development of a new
calcium phosphate powder-binder system for the 3D printing of patient
specific implants, Journal of Materials Science: Materials in Medicine,
vol. 18, pp. (909-916). ISSN 0957-4530
T. Szucs, D. Brabazon. (2009). Effect of saturation and post
processing on 3D printed calcium phosphate scaffolds, Bioceramics21,
vol.396, pp. (663-669). ISSN 1013-9826.
S. Goberis, V. Antonovich. (2004). Improving the structure and
properties of a refractory castable containing porous chamotte fillers.
Refractories and Industrial Ceramics, vol.45, No.5 pp. (360-363). ISSN
1573-9139
J. ReVelle, J. Moran. (1998). The QFD Handbook, Jonh Wiley &
Sons, ISBN 0-471-17381-9, USA
Y. Akao. (1990). Quality Function Deployment: Integrating Customer
Requirements into Product Design. Productivity Press, ISBN
1-56327-313-6, New York, USA
M. Castrillon, C. Garcia, C Paucar. (2007). Evaluation of the
influence of particle size and time of thermal treatment on the
phisycal-mechanical characteristics of a composite of sintered alumina
infiltrated with lantanum glass, DYNA, National University of Colombia
Magazine, vol. 152, pp. (159-165). ISSN 0012-7353