You are on page 1of 27

Rapid Prototyping Journal

Quality control issues in 3D-printing manufacturing: a review


Hsin-Chieh Wu, Tin-Chih Toly Chen,
Article information:
To cite this document:
Hsin-Chieh Wu, Tin-Chih Toly Chen, "Quality control issues in 3D-printing manufacturing: a review", Rapid Prototyping
Journal, https://doi.org/10.1108/RPJ-02-2017-0031
Permanent link to this document:
https://doi.org/10.1108/RPJ-02-2017-0031
Downloaded on: 07 March 2018, At: 14:05 (PT)
References: this document contains references to 0 other documents.
To copy this document: permissions@emeraldinsight.com
Access to this document was granted through an Emerald subscription provided by emerald-srm:178665 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.


Quality control problems in 3D printing manufacturing: A review

Abstract

Purpose – This study investigated issues of quality and quality control (QC) in 3D

printing by reviewing past work and current practices. Possible future developments

are also discussed.

Design/methodology/approach – After a discussion of the major quality dimensions

of 3D-printed objects, the applications of some QC techniques at various stages of the

product life cycle (including product design, process planning, incoming QC,
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

in-process QC, and outgoing QC) are introduced.

Findings – The application of QC techniques to 3D printing is not uncommon. Some

techniques (e.g., cause-and-effect analysis) have been applied extensively; others,

such as design of experiments, have not been used accurately and completely and

therefore cannot optimize quality. Taguchi’s method and control charts can enhance

the quality of 3D-printed objects; however, these techniques require repetitive

experimentation, which may not fit the workflow of 3D printing.

Originality/value – Because quality issues may discourage customers from buying

3D-printed products, enhancing 3D printing quality is imperative. In addition, 3D

printing can be used to manufacture diverse products with a reduced investment in

machines, tools, assembly, and materials. Production economics issues can be

addressed by successfully implementing QC.

Keywords: quality; quality control; 3D printing

1. Introduction
1.1. Background

The quality of a product or process has several dimensions: performance,

reliability, durability, serviceability, aesthetics, features, perceived quality, and

conformance to specifications or standards (Montgomery, 2008). According to the

American Society for Quality (2016), quality control (QC) includes the observation

techniques and activities used to fulfill quality requirements. The seven basic tools of

QC are cause-and-effect diagrams, check sheets, control charts, histograms, Pareto

charts, scatter diagrams, and design of experiments (DOE) (Montgomery, 2008). QC


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

activities can also be classified into incoming QC (IQC), in-process QC (IPQC), and

outgoing QC (OQC) or quality assurance, depending on when these activities are

performed (Orient Semiconductor Electronics, 2016).

3D printing involves building a 3D object from a 3D model layer by layer with

resin or other materials (Berman, 2012). 3D printing has been used to fabricate

prototypes, mockups, replacement parts, dental crowns, artificial limbs, and even

bridges (Berman, 2012). With these successes, 3D printing is considered a convenient

tool for producing complex internal and external porous structures (Asadi-Eydivand et

al., 2016). 3D printing is also cheaper and more efficient than other rapid prototyping

technologies, such as selective laser sintering (SLS) and stereolithography (Silva et al.,

2008). As an extension, 3D printing has been incorporated into a convenience store

chain to form a ubiquitous manufacturing network (Lin and Chen, 2017).

From a manufacturing perspective, 3D printing is a special manufacturing process

in which there is no time gap between the research and development (R&D) stage and

the mass production stage. In conventional manufacturing processes, a time gap is

usually required for amassing the factory capacity and acquiring the raw materials.
This study investigated issues of quality and QC in 3D printing by reviewing past

work and current practices. In this paper, the current practices are mapped to the

stages of a QC cycle (i.e., product design, process planning, IQC, IPQC, and OQC),

and quality and QC activities that were either ignored or not actively performed in 3D

printing are listed (see Fig. 1). Finally, this paper discusses possible future

developments.

Product Design current practice 1


current practice 2
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Process Planning …
current practice n

IQC

IPQC to be performed

OQC

Fig. 1. Investigations performed in this study.

1.2. Importance of the topic

A drug made through 3D printing was approved by the FDA in 2015 (Norman et

al., 2016), which not only revolutionizes pharmaceutical manufacturing but also poses

a challenge for 3D printing because of the high quality requirements for making drugs.

Bose et al. (2013) considered product quality a major limitation in applying 3D printing

to biomedical applications. Wittbrodt et al. (2013) asserted that the quality of

3D-printed objects is a critical factor in the viability of widespread applications of

low-cost 3D printing.
The ultimate target of QC is to manufacture products economically by eliminating

defects and waste. Mironov et al. (2011) asserted that an automated QC system is

essential to the success of a 3D printing system for organ biofabrication. 3D printing

has been recognized as an effective means for economical manufacturing through

reducing the investment in machines, tools, assembly, and materials for diversified

products (Weller et al., 2015). The production economics issues of 3D printing can be

addressed by successfully implementing QC. Because quality issues may discourage

customers from buying products manufactured through 3D printing, creating pressure

to enhance the quality of 3D printing through QC activities, QC is a major concern for


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

3D printing (Weller et al., 2015). A QC program has been launched to guarantee that

lenses manufactured through 3D printing meet both industry and user requirements

(Shang, 2016).

1.3. Problems with existing methods

According to Silva et al. (2008), quality and QC issues in 3D printing or other rapid

prototyping systems have not been sufficiently addressed. In particular, quality and QC

standards for 3D printing lack a clear definition (Berman, 2012).

2. Quality of 3D-Printed Objects

The traditional quality dimensions or attributes are not equally emphasized in 3D

printing, but three of them are critical to 3D-printed objects: aesthetics, conformance

to specifications, and performance. These are discussed in the following sections.

2.1. Aesthetics
A 3D-printed object usually has a rough surface finish and the coloration may not

be as clear as expected, causing it to fail to meet aesthetic requirements. The

resolution of a 3D printer, or the layer thickness it can achieve, is a limiting factor in

this regard (Zavorotnitsienko, 2015). To solve this problem, Lanzetta and Sachs

(2003) used a bimodal powder (i.e., a mix of two types of powder) such that the finer

powder appeared on the upper surface of a 3D object, resulting in a finer surface

finish. The choice of an appropriate orientation can also help to improve the surface

finish of a 3D object (Campbell et al., 2002; Alfieri et al., 2017). In addition, various

power densities can be tested, and the one contributing to the most aesthetically
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

pleasing surface finish can be chosen (German, 1992); this requires some DOE

techniques. The type of printhead, such as the drop-on-demand type or the continuous

jet type (Lanzetta and Sachs, 2003), and the logic behind the slicing program

(Herrmann et al., 2014) also influence the surface finish of a 3D-printed object.

Otherwise, postprocessing treatments are required to polish the surface (Shang, 2016).

For example, Alfieri et al. (2017) proposed a laser processing approach for reducing

surface roughness that incorporated scanning optics and beam wobbling, and used it

to postprocess metal parts fabricated through the selective laser melting of stainless

steel. A recent review of pre-processing and post-processing techniques applied to

enhance the surface finish of a 3D-printed object refers to Chohan and Singh (2017).

2.2. Yield and conformance to specifications

Conformance to specifications is a quality attribute emphasized in many

3D-printing applications. If an object is 3D scanned and then duplicated by a 3D

printer, the shape of the 3D-printed object should be as close as possible to the original.

For example, in diagnosis and treatment planning, 3D biomedical models must be


accurately reproduced to be usable (Silva et al., 2008). Because the 3D model of a

product can easily be shared via the Internet, in theory the product can be manufactured

very similarly anywhere if the same printer, material, and printing conditions are used.

However, some studies have still noted limited reproducibility of products created

using 3D printing (Weller et al., 2015).

An extended concept of conformance to specifications is yield, which is the

percentage of jobs that are successful after production. A perfect yield (100%) is

required for human organ biofabrication using 3D printing (Mironov et al., 2011).

According to Weller et al. (2015), because 3D printing can prevent mistakes by


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

eliminating many manual operations, it increases yield. However, each product printed

with a 3D printer may be considerably different, which is not conducive to the

accumulation of knowledge about products. Consequently, the results of yield

calculations can fluctuate widely, and 3D printing may not have the same learning

process as a volume production case (Chen and Wang, 1999). Nevertheless, knowledge

about the use of a 3D printer and the control of 3D-printing processes can be

accumulated. According to Grieser (2015), the learning curve for yield in 3D printing is

steep, meaning that such knowledge or experience is easy to obtain.

2.3. Performance

The performance of a 3D-printed object depends on its purpose. If it is only a

prototype, it may not have all of the required functions, but otherwise its performance

should be comparable to that of the same product fabricated using traditional

manufacturing technologies. For example, a product traditionally created through

molding but now created through 3D printing is expected to have features such as a

high impact resistance and Young’s modulus (Hopkinson and Dickens, 2001). A
mechanical part, such as a build tray, should have a high tensile strength and modulus,

even if it is built through 3D printing (Barclift and Williams, 2012). Calì et al. (2012)

decomposed a 3D object into joints, and after the assembly of these joints, the 3D

object was posable and would not fall. Lenses constructed through 3D printing should

achieve high transparency and surface smoothness (Shang, 2016).

Various QC programs have been launched to guarantee that products

manufactured using 3D printing meet industrial and user requirements (Shang, 2016).

However, some researchers have asserted that a product fabricated using 3D printing

is usually not comparable in functionality to its counterparts produced using


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

traditional manufacturing technologies (Hopkinson and Dickens, 2001). By contrast,

other researchers have noted that 3D printing provides opportunities for further

improving the quality of a product manufactured using traditional technologies. For

example, aircraft parts constructed through 3D printing are expected to lighten an

aircraft by 50% (Young, 2015) because of the improved precision in forming parts

and the elimination of assembly operations.

According to Mironov et al. (2011), focusing on a single performance (or

functionality) area of a 3D object can prevent failures in bioprinting human organs and

the subsequent downtime. Nevertheless, more functionalities and more flexible

functionalities are always being pursued.

3. QC in a 3D-Printing Process

QC is a challenging task for 3D printing, for the following reasons.

(1) Although some applications of 3D printing for mass production exist, 3D printing

is mostly used for prototyping during R&D, for which the volume of production is

low and previous experience cannot be consulted.


(2) QC issues vary with the materials used or products to be printed. For example,

warping is a serious problem in products with elongated or rectangular shapes, but

not for products with vertical structures. For this issue, heating the print table in

advance may be helpful (Herrmann et al., 2014).

(3) The development of 3D printing technologies is still underway, meaning that there

are multiple alternatives without an absolute rule for choosing among them.

Nevertheless, according to Norman et al. (2016), an understanding of products and

processes facilitates the development of a QC strategy for different 3D printing

methods.
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Grieser (2015) identifies ten actions that can improve the quality of a 3D-printed

object.

(1) Setting up the 3D printer according to the equipment vendor’s instructions.

(2) Regularly updating the printer’s hardware and software.

(3) Periodically maintaining and calibrating the printer.

(4) Cleaning the printbed before every printing.

(5) Leveling the printbed before every printing.

(6) Adjusting the distance between the printhead and the printbed.

(7) Printing only objects with moderate size and complexity.

(8) Choosing a filament with sufficient adhesion.

(9) Terminating the printing process when the results of the first few layers are poor.

(10) Being patient.

This advice, if not properly followed, can cause defects in a 3D-printed object (Fig.

2).
man machine

A large or complicated The 3D printer was


3D object was printed not set up properly
The hardware and software
The process was were not updated regularly
not early terminated

The use was not patient The 3D printer was not maintained
and calibrated periodically
the poor quality of
a 3D-printed product
The printbed was not
cleaned before each print
The adhesion of the filament
was not sufficient The printbed was not
leveled before each print
The distance between printhead
and printbed was not adjusted

material method

Fig. 2. Cause-and-effect analysis for the poor quality of a 3D-printed product.


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

4. Applications of QC Techniques to 3D Printing

The following subsections discuss the applications of some QC techniques to 3D

printing. These applications are classified according to the stages of the product life

cycle (including product design, process planning, IQC, IPQC, and OQC) to which

QC techniques can apply.

4.1. Product design

The quality of a 3D-printed object depends on the quality of the initial 3D model

of the object; therefore, the image acquisition step is essential to the quality of a

3D-printed object (Rengier et al., 2010). For this step to be successful, the spatial

resolution of the imaging system, or voxels, must be sufficiently high (e.g., >400 µm

for medical applications) (Rengier et al., 2010). However, no clear guidelines have

been established for determining the required spatial resolution. Furthermore, some

factors influencing the quality of a 3D-printed object must be determined at this stage,

such as the slice thickness. Rengier et al. (2010) recommended a slice thickness of less

than 1 mm for medical applications.


Considerable effort has been devoted to building .STL/.OBJ-compliant 3D

models from existing 3D databases such as medical image databases (Rengier et al.,

2010) and anthropometry databases (Straub et al., 2015). The effectiveness of the

employed algorithm strongly influences the quality of the 3D model. However,

designing an effective algorithm is difficult. Furthermore, many existing 3D databases

are heterogeneous, meaning that a dedicated algorithm must be designed for each

database.

4.2. QC tools for process planning


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Several QC tools, such as expert systems, classification and regression trees

(CARTs), and decision trees, can be applied to facilitate the planning of the 3D

printing process.

Compared with control charts, expert systems may be more effective and

practical for process planning in 3D printing. As increasingly more products are

printed, a user accumulates knowledge about how to optimize the printing conditions

for various products. This knowledge can be subjectively expressed by the user, or

objectively mined using tools such as CARTs (Wu and Chen, 2015). The extracted

knowledge can be stored in a knowledge base on which an expert system can be built,

or illustrated with CARTs or decision trees. Figure 3 depicts an example of such a

decision tree. However, neither expert systems nor decision trees have been widely

applied to 3D printing, and most relevant studies have only reported on the gained

knowledge and experience.


Use printing conditions no. 1
Yes
Layer thickness ≤ c3
Yes No
Use printing conditions no. 2
Powder flowability ≤ c2
Use printing conditions no. 3
No Yes
Layer thickness ≤ c3
Yes No
Use printing conditions no. 4

Powder packing density ≤ c1


Use printing conditions no. 5
Yes
No
Layer thickness ≤ c3
Yes No
Use printing conditions no. 6
Powder flowability ≤ c2
No Use printing conditions no. 7
Yes
Layer thickness ≤ c3
No
Use printing conditions no. 8

Fig. 3. Example decision tree.


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Segmentation of the 3D model is an important step in the process planning stage

(Jacobs et al., 2008). Therefore, the effectiveness of the segmentation algorithm

critically influences the quality of the printed component. Common segmentation

algorithms include simple region growing, surface/volume rendering,

maximal/minimal intensity projection, and multiplanar reformation (Rengier et al.,

2010). However, low-resolution and nonenhanced 3D models require more enhanced

segmentation algorithms (Rengier et al., 2010).

4.3. DOE and Taguchi’s method

DOE, a basic tool in quality engineering, can be applied to optimize the settings

of a 3D printer as well as other factors in a 3D-printing process, especially when there

are interactions among the factors. However, DOE is rarely accurately and completely

applied; instead, the settings of a 3D printer are often determined subjectively, for

example, according to the limited experience of the user (Herrmann et al., 2014). A

possible reason is the high number of factors to consider. For example, to fabricate
scaffolds with 3D printing, the values of at least six major factors (powder packing

density, powder flowability, layer thickness, binder drop volume, binder saturation,

and powder wettability) must be set to optimize the quality of a printed scaffold, but

to discover the appropriate values requires repetitive and time-consuming

experimentation (Bose et al., 2013). To overcome this problem, Taguchi’s orthogonal

arrays can limit the replications of experiments to cover a wide range for each factor

(Yang and El-Haik, 2008). For example, for three factors (e.g., powder packing

density, powder flowability, and layer thickness) each with two levels, in theory, 8

(23) replications are required to consider all possible combinations to optimize the
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

performance; however, the L4 orthogonal array (see Table 1) can be used to achieve

the same goal with only four replications.

Table 1. L4 orthogonal array.


Powder
Replicatin Powder Layer
Packing
No. Flowability Thickness
Density
1 Level 1 Level 1 Level 1
2 Level 1 Level 2 Level 2
3 Level 2 Level 1 Level 2
4 Level 2 Level 2 Level 1

When each factor has three levels, central composite design (CCD) can be employed

to build a quadratic model for the response variable, thereby eliminating the need to

conduct a complete three-level factorial experiment (Mohammad et al., 2017).

Some examples of applying DOE and Taguchi’s method to 3D printing are

reviewed as follows. To analyze the effects of three factors on the mechanical

properties of a 3D-printed photopolymer part, Barclift and Williams (2012) performed

a three-factor, two-level full factorial DOE that included eight experiment runs. Hsiao

(2015) designed an experiment of nine runs according to the L9 (34) orthogonal array to
investigate the effects of four control factors (including nozzle height, printing speed,

and UV light exposure time), each with three levels, on two qualities (variation in the

droplet diameter and the maximum coverage) of a product fabricated by a photocurable

printing system. The results along the two quality dimensions were aggregated using

the grey relational analysis method (Kuo et al., 2008). The values of the control factors

giving the best grey relational grade were chosen.

4.4. IQC
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

An important goal of IQC is to ascertain the suitability of an incoming material

for the other materials or subassemblies. However, in 3D printing, various products

are commonly created from a single type of material. In addition, the whole

manufacturing process of a product can be completed by a 3D printer, eliminating

subassembly. An incoming material is not checked but simply replaced if it is unfit

for the use. These properties eliminate the need to conduct extensive IQC in 3D

printing, although the choice of material remains critical to the quality of a 3D-printed

object (Lanzetta and Sachs, 2003). IQC is generally a simple task in 3D printing

because the inputs to a 3D printer are raw materials. However, Hopkinson and

Dickens (2001) noted a contamination problem caused by manually scooping the

unused materials in a 3D printing process back into the feed tray.

4.5. IPQC

In-process quality control (IPQC) is a QC activity that is critical but may be

overlooked (Manufacturing Behavioral Science, 2010). IPQC is based on the premise

that some signals detected during the manufacturing process can be related to defects

in finished goods. As Grieser (2015) mentioned, it is important to terminate a 3D


printing process early if the first few layers are not satisfactory. This IPQC task can be

fulfilled if a 3D-printing process is closely and continuously monitored either

manually or automatically. However, IPQC is a challenging task for 3D printing

because of the considerable variations between the products manufactured by a 3D

printer. For this reason, visual inspection is a common IPQC practice for 3D printing.

Nevertheless, sensors can be installed to facilitate automatic and quick measurement

and inspection of the printing results (Mendibil et al., 2016).

4.6. Control charts


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Control charts, a crucial tool in statistical QC, can be classified as control charts for

variables and control charts for attributes (Montgomery, 2008). The first category

contains control charts for individual measurements that are common in 3D printing.

However, such control charts are designed for identical products, which seldom happen

in 3D printing. The second category contains control charts for attributes that have

great potential to be applied to 3D printing. For example, the number of defects on a

3D-printed object is a critical problem. Under the assumption that the number of

defects on a unit of surface follows a Poisson distribution, control charts for

nonconformities (defects) can be constructed as follows to minimize the number of

defects (Montgomery, 2008):

Upper control limit = c + 3 c (1)

Center line = c (2)

Lower control limit = c − 3 c (3)

where c is the average number defects on a unit of surface according to collected

samples. However, such control charts are suitable only for products with

homogeneous properties.
4.7. Reliability

Human mistakes and neglect are the two main causes of low reliability, and 3D

printing has been considered effective for enhancing reliability because it eliminates

human intervention (Wittbrodt et al., 2013).

Compared with traditional manufacturing machines, a 3D printer is less reliable

and breaks down easily. To address this problem, 3D printer vendors continually

improve the hardware and software. For example, the printhead is a critical component;
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

the mean time between failures of a printhead is therefore a meaningful index of the

reliability of a 3D printer (Benchoff, 2015). To enhance thermal management, sensor

points have been added to the printhead (Tarantola, 2016). However, the reliability of a

3D printer is seldom monitored in practice (Benchoff, 2015).

The printing material is another cause of the low reliability of a 3D printer. For

example, some materials can clog the nozzle by drying within it. To address this

problem, Lanzetta and Sachs (2003) used ethylene glycol-based colloidal silica instead

of aqueous colloidal silica. However, the most critical criterion for choosing the

printing material is the resulting performance of the printed objects rather than the

reliability of the printer. Reliability may therefore not be the top priority in 3D

printing.

4.8. Cause-and-effect diagram

One of the seven basic tools of QC is the cause-and-effect diagram. Barclift and

Williams (2012) used a cause-and-effect diagram to categorize the factors influencing

the mechanical properties of a photopolymer part fabricated using 3D printing. The

factors were classified into six categories: man, machine, method, material,
measurement, and environment. The classification results helped group the efforts to

improve the quality of the 3D-printed object, thereby eliminating redundant or

contradictory actions and enhancing efficiency and effectiveness. Fig. 1 provides

another example.

4.9. OQC

OQC is the QC stage wherein the product is validated against the customer’s

requirements. Finished goods are usually sampled before inspection. However, each
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

product is usually printed only once, rendering sampling impossible; that is, all

products must be inspected. The features that are typically checked include product

appearance, performance, service life, and packaging, with some features (e.g.,

appearance and performance) having higher emphasis in 3D printing. In addition,

when a 3D-printed object is used for prototyping, there is no external customer but

only an internal customer (i.e., the R&D or product engineer), and the requirements

for the 3D-printed object differ from those for a commercial product. Moreover,

whether a 3D-printed object can meet the requirements for a commercial product is

often questioned.

4. Standards and Requirements

4.1. Quality and QC standards for 3D printing

Both .STL and .OBJ are standard file formats for 3D printing, and adhering to

them ensures ubiquitous printability for 3D objects (Lin and Chen, 2017). Some

standards and guidelines, such as the 3D Printing & Additive Manufacturing

Equipment Guideline by UL (UL, 2015), can be followed while developing


3D-printing equipment. Following such standards is believed to yield safe and

high-quality equipment.

Some standards and guidelines for the quality of a 3D-printed object and the

execution of QC activities within the 3D printing process have recently been proposed

by various international organizations. For example, the American Society for Testing

and Materials International (ASTM International) proposed some guidelines for

fabricating safe and high-quality components by using powder-bed fusion methods

involving laser and electron-beam sources (Orr, 2014), such as electron beam melting,

SLS, selective laser melting, and direct metal laser sintering. For example, the ASTM
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

F3091 standard specifies the quality requirements for the mechanical, tolerance,

surface finishing, and postprocessing properties of polyamide components fabricated

through SLS. The ASTM F3049 standard is a version of the ASTM F3091 standard

that is tailored for the automotive, aerospace, and medical industries. The WK46188

standard instructs how to determine the values of the process parameters for

powder-bed fusion. However, further efforts for standardization in 3D printing are

necessary and underway.

4.2. Quality and QC requirements in various industries

Quality and QC requirements for 3D-printing applications vary across industries.

For example, appearance (e.g., surface finishing and colors) is a critical characteristic

for 3D-printed objects used in the education industry to teach students new concepts

(Rengier et al., 2010). The requirement also applies to the medical industry, where

3D-printed tissues, organs, and scaffolds must be vivid to facilitate the training of

doctors, the diagnosis of anomalies, and the planning of treatments. By contrast, in

mechanical industries, the primary concern is whether a component can be precisely


assembled along with other parts and whether the assembly functions as designed

(Barclift and Williams, 2012), and the appearance of the component is not a crucial

concern. Similarly, in the pharmaceutical industry, where customized drugs are

synthesized through 3D printing (Acosta-Vélez et al., 2017), the appearance of the

drug is largely irrelevant; the successful printing of the customized drug and it

efficacy are of prime importance.

By contrast, in all industries that apply 3D printing, conformance to the product

specifications and designed performance of the 3D-printed object must be ensured.


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

5. Future R&D Perspectives

Table 2 summarizes the investigation results and lists both problems that have

been solved and those that remain to be addressed.

Table 2. Summary of the findings.


Problems Yet to be
Stage Problems Solved
Investigated
 Guidelines for
determining the minimal
 Standard 3D model file spatial resolution of a 3D
formats scanning system
Product design  Algorithms for  Algorithms for converting
converting specific 3D other types of 3D
databases databases
 Enhancing low-resolution
3D models
 Segmentation algorithms  Systematic accumulation
for high-resolution and of acquired knowledge
enhanced 3D models and experience
Process planning
 Application of full  Segmentation algorithms
factorial or subjective for low-resolution and
DOE nonenhanced 3D models
 Applications of Taguchi
method and objective
DOE
 More types of materials
 Replacing materials  More methods of using
unsuitable for use the materials
IQC
 No need to conduct  Application of the
extensive IQC Taguchi method and
objective DOE
 Guidelines for early
termination
 Visual inspection
 Managing variation
 SQC guidelines for
IPQC between products
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

specific 3D-printing
 Dedicated control charts
processes
 SQC guidelines for other
3D-printing processes
 Selective and subjective
 Purpose-oriented OQC
OQC OQC
 Clear guidelines for OQC
 Internal customers

On the basis of these findings, we recommend that the following R&D directions

be explored in the near future:

(1) Product design: Guidelines for the minimal spatial resolution of a 3D scanning

system should be established, and these guidelines should account for the purpose

of 3D printing. In addition, effective algorithms must be designed for converting

the numerous heterogeneous 3D databases. Furthermore, low-resolution 3D

models must be enhanced to facilitate 3D printing.

(2) Process planning: There are already several online hubs for gathering and sharing

3D models (e.g., myminifactory.com). However, knowledge bases or expert

systems are still required for guiding the setup of 3D printers and the preparation

of the printing conditions. Such knowledge bases or expert systems can be

integrated with the software of a 3D printer. In addition, algorithms for


segmenting low-resolution and nonenhanced 3D models must be proposed.

Furthermore, the Taguchi method and objective DOE can be applied to enhance

the correctness and efficiency of experimentation to improve the quality of

3D-printed objects by optimizing the production setting.

(3) IQC: As researchers and practitioners desire more types of usable materials and

more methods to use these materials, Taguchi method and objective DOE can be

applied to optimize the method of using materials (e.g., choosing an optimal

combination of materials).

(4) IPQC: Guidelines must be established to determine whether a 3D printing process


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

should be terminated early. In addition, control charts dedicated to 3D printing as

well as the corresponding SQC guidelines should be devised.

(5) OQC: The features to be checked in OQC are clearly purpose-oriented; therefore,

relevant clear guidelines must be formulated.

6. Conclusions

From the previous review and discussion, the following can be concluded about

the current research on and practice of QC in 3D printing:

(1) 3D printer vendors continue to upgrade hardware and software to improve the

quality of 3D-printed objects. However, the preparation of the printing

environment and the setup, monitoring, and maintenance of a 3D printer may be

more critical, all of which are the responsibility of the user.

(2) The application of QC techniques to 3D printing is not uncommon.

Cause-and-effect analysis and DOE, in particular, are extensively applied.

However, the DOEs in many studies were subjective and incomplete, and
therefore could not guarantee the optimization of the quality of the 3D-printed

objects.

(3) QC techniques such as Taguchi’s method and control charts can be applied to

further enhance the quality of 3D-printed objects. However, these QC techniques

require repetitive experimentation, whereas an object may be 3D-printed only

once or a few times. This limits the applicability of these two techniques.

(4) 3D printing has both positive and negative effects on product quality. A product

fabricated using 3D printing is usually cruder than counterparts fabricated using

conventional manufacturing technologies. However, 3D printing can fabricate


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

parts with complicated shapes exactly according to the specifications, which can

improve the quality of a product composed of 3D-printed parts.

Regarding future R&D, QC will play a more important role in 3D printing as an

increasing number of products can be created or even mass produced through 3D

printing. In addition, to implement IPQC, a 3D printer can be equipped with sensors

to detect any abnormal conditions that require early termination of the printing

process.

References

[1] G. F. Acosta-Vélez, C. S. Linsley, M. C. Craig, and B. M. Wu (2017) Photocurable

bioink for the inkjet 3D pharming of hydrophilic drugs. Bioengineering, 4, 11.

[2] V. Alfieri, P. Argenio, F. Caiazzo, and V. Sergi (2017) Reduction of surface

roughness by means of laser processing over additive manufacturing metal parts.

Materials, 10, 30.


[3] American Society for Quality (2016) Quality assurance vs. quality control.

http://asq.org/learn-about-quality/quality-assurance-quality-control/overview/ove

rview.html

[4] M. Asadi-Eydivand, M. Solati-Hashjin, A. Farzad, and N. A. A. Osman (2016)

Effect of technical parameters on porous structure and strength of 3D printed

calcium sulfate prototypes. Robotics and Computer-Integrated Manufacturing, 37,

57-67.

[5] M. W. Barclift, and C. B. Williams (2012) Examining variability in the mechanical

properties of parts manufactured via polyjet direct 3d printing. International Solid


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Freeform Fabrication Symposium, pp. 6-8.

[6] B. Benchoff (2015) 3D printering: Hobbs meters and 3D printer reliability.

http://hackaday.com/2015/02/10/3d-printering-hobbs-meters-and-3d-printer-relia

bility/

[7] B. Berman (2012) 3-D printing: The new industrial revolution. Business Horizons,

55(2), 155-162.

[8] S. Bose, S. Vahabzadeh, and A. Bandyopadhyay (2013) Bone tissue engineering

using 3D printing. Materials Today, 16(12), 496-504.

[9] J. Calì,D. A. Calian,C. Amati,R. Kleinberger,A. Steed,J. Kautz,and T. Weyrich

(2012) 3D-printing of non-assembly, articulated models. ACM Transactions on

Graphics, 31(6), 130.

[10] R. I. Campbell, M. Martorelli, and H. S. Lee (2002) Surface roughness

visualisation for rapid prototyping models, CAD Computer Aided Design, 34(10),

717-725.

[11] T. Chen, and M.-J. J. Wang (1999) A fuzzy set approach for yield learning

modeling in wafer manufacturing. IEEE Transactions on Semiconductor

Manufacturing, 12(2), 252-258.


[12] J. S. Chohan, and R. Singh (2017) Pre and post processing techniques to improve

surface characteristics of FDM parts: A state of art review and future applications.

Rapid Prototyping Journal, 23(3), 4.

[13] R.M. German (1992) Prediction of sintered density for bimodal powder mixtures.

Metallurgical Transactions A (Physical Metallurgy and Materials Science),

23A(5), 1445-1465.

[14] F. Grieser (2015) 3D printing quality issues: 10 tricks to avoid them.

https://all3dp.com/3d-printing-quality/

[15] K. H. Herrmann, C. Gärtner, D. Güllmar, M. Krämer, and J. R. Reichenbach


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

(2014) 3D printing of MRI compatible components: Why every MRI research

group should have a low-budget 3D printer. Medical Engineering &Physics,

36(10), 1373-1380.

[16] N.Hopkinson, and P. Dickens (2001) Rapid prototyping for direct manufacture.

Rapid Prototyping Journal, 7(4), 197-202.

[17] S.-Y. Hsiao (2015) Study of Parameter Optimization of Gradient Color Process for

Phto-curable 3D Printing System. Ph.D. thesis, Institute of Automation and

Control, National Taiwan University of Science and Technology.

[18] S. Jacobs, R. Grunert, F. W. Mohr, and V. Falk (2008) 3D-Imaging of cardiac

structures using 3D heart models for planning in heart surgery: A preliminary

study. Interactive Cardiovascular and Thoracic Surgery, 7(1), 6-9.

[19] Y. Kuo, T. Yang, and G. W. Huang (2008) The use of grey relational analysis in

solving multiple attribute decision-making problems. Computers & Industrial

Engineering, 55(1), 80-93.

[20] Y.-C. Lin, and T. Chen (2017) A ubiquitous manufacturing network system.

Robotics and Computer-Integrated Manufacturing, 45, 157-167.


[21] M. Lanzetta, and E. Sachs (2003) Improved surface finish in 3D printing using

bimodal powder distribution. Rapid Prototyping Journal, 9(3), 157-166.

[22] Manufacturing Behavioral Sceince (2010) In-process quality control: An

introduction. http://www.manufacturingbs.com/CB/CB-01.pdf

[23] X. Mendibil, I. Llanos, H. Urreta, and I. Quintana (2016) In process quality control

on micro-injection moulding: the role of sensor location. International Journal of

Advanced Manufacturing Technology, in press.

[24] V. Mironov, V. Kasyanov, and R. R. Markwald (2011) Organ printing: from

bioprinter to organ biofabrication line. Current Opinion in Biotechnology, 22(5),


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

667-673.

[25] A. Mohammad, A. M. Al-Ahmari, A. AlFaify, and M. K. Mohammed (2017)

Effect of melt parameters on density and surface roughness in electron beam

melting of gamma titanium aluminide alloy. Rapid Prototyping Journal, 23(3), 2.

[26] D. C. Montgomery (2008) Introduction to Statistical Quality Control. New York:

John Wiley & Sons.

[27] J. Norman, R. D. Madurawe, C. M. Moore, M. A. Khan, and A. Khairuzzaman

(2016) A new chapter in pharmaceutical manufacturing: 3D-printed drug products.

Advanced Drug Delivery Reviews, in press.

[28] Orient Semiconductor Electronics (2016) Quality control process.

http://www.ose.com.tw/electronics-manufacturing/commitment-to-quality/quality

-control-process

[29] T. Orr (2014) ASTM International proposes new 3D printing guidelines for

powder bed fusion machines.

https://3dprint.com/13367/astm-international-standards/

[30] F. Rengier, A. Mehndiratta, H. von Tengg-Kobligk, C. M. Zechmann, R.

Unterhinninghofen, H. U. Kauczor, andF. L.Giesel (2010) 3D printing based on


imaging data: review of medical applications. International Journal of Computer

Assisted Radiology and Surgery, 5(4), 335-341.

[31] J. Shang (2016) Luxexcel and Automation&Robotics collaborate on a quality

control program.

https://3dprintingindustry.com/news/luxexcel-automationrobotics-collaborate-qua

lity-control-program-95826/

[32] D. N. Silva, M. G. De Oliveira, E. Meurer, M. I. Meurer, J. V. L. da Silva, and A.

Santa-Bárbara (2008) Dimensional error in selective laser sintering and

3D-printing of models for craniomaxillary anatomy reconstruction. Journal of


Downloaded by INSEAD At 14:05 07 March 2018 (PT)

Cranio-Maxillofacial Surgery, 36(8), 443-449.

[33] J. Straub, B. Kading, A. Mohammad, and S. Kerlin (2015) Characterization of a

large, low-cost 3D scanner. Technologies, 3, 19-36.

[34] A. Tarantola (2016) Makerbot just made a more reliable 3D printer head.

https://www.engadget.com/2016/01/04/makerbot-built-a-longer-lasting-and-more

-reliable-extruder/

[35] UL (2015) 3D printing & additive manufacturing equipment guideline.

http://www.ul.com/wp-content/themes/countries/downloads/am/3D-PRINTING-

EQUIP-SAFETY-GUIDELINE_EDITION2.pdf

[36] C. Weller, R. Kleer, and F. T. Piller (2015) Economic implications of 3D printing:

market structure models in light of additive manufacturing revisited. International

Journal of Production Economics, 164, 43-56.

[37] B. T. Wittbrodt, A. G. Glover, J. Laureto, G. C. Anzalone, D. Oppliger, J. L.Irwin,

and J. M. Pearce (2013) Life-cycle economic analysis of distributed manufacturing

with open-source 3-D printers. Mechatronics, 23(6), 713-726.


[38] H.-C. Wu, and T. Chen (2015) CART–BPN approach for estimating cycle time in

wafer fabrication. Journal of Ambient Intelligence and Humanized Computing, 6,

57-67.

[39] K. Yang, and B. El-Haik (2008) Taguchi’s orthogonal array experiment. Design

for Six Sigma: A Roadmap for Product Development, pp. 469-497.

[40] J. Young (2015) 3D Printed Aircraft Parts and Engines Could Lighten Aircrafts

by 50%.

http://3dprinting.com/aviation/3d-printed-aircraft-parts-could-lighten-aircrafts-by

-fifty-percent/
Downloaded by INSEAD At 14:05 07 March 2018 (PT)

[41] D. Zavorotnitsienko (2015) Understanding 3D printer quality &resolution.

http://www.ilios3d.com/en/product-documentation/ilios-documentation-3dprint-q

uality

You might also like