DEPARTMENT OF MECHANICAL AND

INDUSTRIAL ENGINEERING

TOMÁS MEALHA MARTINS RODRIGUES CUNHA

BSc in Science of Industrial Engineering and Management

A PROCESS-BASED COST MODEL FOR

WIRE AND ARC ADDITIVE
MANUFACTURING

INTEGRATED MASTER IN INDUSTRIAL ENGINEERING AND MANAGEMENT

MANAGEMENT ENGINEERING
NOVA University Lisbon
JUNE, 2022
2
 DEPARTMENT OF MECHANICAL AND INDUSTRIAL
ENGINEERING

A PROCESS-BASED COST MODEL FOR WIRE AND

ARC ADDITIVE MANUFACTURING

TOMÁS MEALHA MARTINS RODRIGUES CUNHA

BSc in Science of Industrial Engineering and Management

Adviser: Helena Maria Lourenço Carvalho Remigio,

Associate Professor NOVA University Lisbon
Co-adviser: Bruno Alexandre Rodrigues Simões Soares,
Assistant Professor NOVA University Lisbon

Examination Commitee:
Chair: Isabel Maria do Nascimento Lopes Nunes, Associate
Professor NOVA University Lisbon

Rapporteurs: Radu Godina, Assistant Professor NOVA University

Lisbon

Member: Helena Maria Lourenço Carvalho Remigio, Associate

Professor NOVA University Lisbon

INTEGRATED MASTER IN INDUSTRIAL ENGINEERING AND MANAGEMENT

NOVA University Lisbon

JUNE, 2022
3
4
A Process-based Cost Model for Wire and Arc Additive Manufacturing

Copyright © Tomás Mealha Martins Rodrigues Cunha, NOVA School of Science and
Technology, NOVA University Lisbon.
The NOVA School of Science and Technology and the NOVA University Lisbon have
the right, perpetual and without geographical boundaries, to file and publish this thesis
through printed copies reproduced on paper or on digital form, or by any other means known
or that may be invented, and to disseminate through scientific repositories and admit its cop-
ying and distribution for non-commercial, educational or research purposes, as long as credit
is given to the author and editor.

5
6
 ACKNOWLEDGEMENTS

First, I would like to thank my advisor Helena Carvalho and my co-adviser for the sup-
port, help, understanding, and motivation given throughout the development of this thesis.
I would also like to thank Valdemar Duarte for his help and support during the thesis
preparation.
I would also like to thank the Smart WAAM project: Microstructural Engineering and
Integrated Non-Destructive Testing (Project nº20144 - FPA2016/EIT/EIT RAW
MATERIALS), sponsored by EIT raw materials and the European Union, for allowing me to
develop this thesis.
I would also like to thank my family for motivating and supporting me in the most
difficult moments of the elaboration of this document.
Finally, I thank my friends for motivating and supporting me in the most difficult mo-
ments of the elaboration of this thesis.

i
ii
 ABSTRACT

In engineering one of the main criteria to evaluate new technology or product is its eco-
nomic viability. This can only be done by identifying the costs related to the process or prod-
uct. Within the Smart WAAM project, which aims to study the use of wire and arc additive
manufacturing (WAAM) technology to create, repair and expand the life of large industrial
components, it was necessary to develop a cost model to study the economic viability of this
technology.
This thesis's primary goal is to develop a cost model for the WAAM technology, consid-
ering a product life cycle approach. For this purpose, it was necessary to develop a model to
estimate the cost of the WAAM technology, as well as the main factors influencing the cost.
A process-based cost model (PBCM) was developed since it allows to analyse the costs
of the different life cycle phases of a product and estimates the production costs. The study
main steps were the objective and scope definition, using a cradle to gate approach, the pro-
cess description, and the cost model's development. The object of study was an experimental
WAAM machine developed at NOVA School of Science and Technology, and the functional
unit was a hollow stainless steel AISI316LSI cube of approximately 7x7x7 cm. The data collec-
tion process included the compilation of secondary data available in public websites, but also
primary data was collected through unstructured interviews with researchers who developed
and worked with the WAAM machine. The model was validated, and the factors influencing
the cost were identified.
It was possible to determine that the production of 500 cubes has a total cost of 259.95€
per piece. The WAAM process and the surface finishing process and substrate removal rep-
resent 84% of the total cost. The main factors that influence the total cost of the process are the
acquisition cost of the machines for the production and parts finishing, the cost of the tools,
namely the cutters, and the production overheads.

Key words: Wire and Arc Additive Manufacturing, Life Cycle Cost, Process-Based Cost
Model

iii
iv
 RESUMO

Um dos principais critérios para a avaliação de uma nova tecnologia ou produto, na

engenharia, é a sua viabilidade económica. Esta, só pode ser estudada através da identificação
dos custos inerentes ao processo ou produto. No âmbito do projeto Smart WAAM, que pre-
tende estudar a utilização da tecnologia fabrico aditivo usando fio consumível e arco elétrico
(WAAM) para criar, reparar e expandir a vida útil de grandes componentes industriais, foi
necessário desenvolver um modelo de custos que permita estudar a viabilidade económica da
utilização desta tecnologia.
O principal objetivo desta dissertação é o desenvolvimento de um modelo de custo para
a tecnologia WAAM, considerando uma perspetiva do ciclo de vida do produto. Para esta
finalidade, foi necessário desenvolver um modelo que permita estimar os custos da tecnologia
WAAM, assim como os principais fatores que influenciam o custo.
Foi desenvolvido um modelo de custo baseado no processo, pois permite analisar os
custos das diferentes fases do ciclo de vida do produto, bem como fazer estimativas dos custos
de produção. Inicialmente o estudo consistiu na definição do objetivo e do âmbito, utilizando
uma abordagem berço ao portão, a descrição do processo e o desenvolvimento do modelo de
custo. O objeto de estudo foi um equipamento experimental desenvolvido na NOVA School
of Science and Technology e a unidade de análise foi um cubo oco de aço inoxidável
AISI316LSI com aproximadamente 7x7x7 cm. Foram recolhidos dados secundários (por exem-
plo preços de energia e matéria-prima) e dados primários recorrendo a entrevistas não estru-
turadas a investigadores que desenvolveram e trabalham com o equipamento. Por fim, o mo-
delo foi validado e foram identificados quais os fatores que mais influenciam o custo.
Foi possível determinar que a produção de 500 cubos, tem o custo total de 259.95 € por
peça. O processo WAAM e o processo de acabamento e remoção do substrato representam
84% do custo total. Os principais fatores que influenciam o custo total do processo são: o custo
de aquisição das máquinas de produção e acabamento das peças, o custo das ferramentas,
nomeadamente as fresas, e as despesas gerais de produção.

Palavras-Chave: Fabrico Aditivo usando Fio Consumível e Arco Elétrico, Análise Económica
de Ciclo de Vida, Modelo de Custos Baseado no Processo

v
 CONTENTS

1. INTRODUCTION ......................................................................................................................... 1

2. STATE OF ART ............................................................................................................................. 4

2.2.1. Strengths and Limitations of WAAM ............................................................................ 8

2.3.1. Goal Definition ................................................................................................................ 10

2.3.2. Scope Definition .............................................................................................................. 10
2.3.3. Life Cycle Inventory ....................................................................................................... 15
2.3.4. Results and Discussion................................................................................................... 18
2.3.5. Conclusions ..................................................................................................................... 19

3. WAAM LIFE CYCLE COST MODEL DEVELOPMENT ........................................................ 23

3.4.1. Characterization of the WAAM Machine .................................................................... 28

3.4.2. WAAM Process ............................................................................................................... 29
3.4.3. Cost Model ....................................................................................................................... 31

3.6.1. Model Validation and Sensitivity Analysis ................................................................. 39

3.6.2. Final Results .................................................................................................................... 45

4. CONCLUSIONS AND FUTURE WORK ................................................................................. 51

BIBLIOGRAPHY..................................................................................................................................... 54

APPENDIX .............................................................................................................................................. 58
vi
vii
 List of Figure

FIGURE 2-1 SCHEMATIC OF GAS METAL ARC WELDING. ADAPTED FROM BEKKER & VERLINDEN, 2018............. 7
FIGURE 2-2 DRAGON BENCH SOFA, MADE IN STAINLESS STEEL BY WAAM. RETRIVED FROM BEKKER &
VERLINDEN, 2018 .......................................................................................................................................... 8
FIGURE 2-3 DIAGRAM OF UNIT PROCESSES. ADAPTED FROM PRIARONE ET AL (2019) ...................................... 12
FIGURE 2-4 DIAGRAM OF UNIT PROCESSES. ADAPTED FROM PRIARONE ET AL., 2020 ....................................... 12
FIGURE 3-1 WAAM MACHINE ............................................................................................................................ 24
FIGURE 3-2 LIFE CYCLE COST MODEL METHODOLOGY ...................................................................................... 24
FIGURE 3-3 HF-WAAM ILLUSTRATION.............................................................................................................. 29
FIGURE 3-4 DIAGRAM OF THE WAAM PROCESS ................................................................................................ 30
FIGURE 3-5 SENSITIVE ANALYSIS OF THE PART INFORMATION ........................................................................... 40
FIGURE 3-6 MATERIAL INFORMATION SENSITIVE ANALYSIS ............................................................................... 41
FIGURE 3-7 WAAM INPUTS SENSITIVE ANALYSIS............................................................................................... 41
FIGURE 3-8 WAAM PERCENTAGE INPUTS SENSITIVE ANALYSIS ........................................................................ 42
FIGURE 3-9 FINISH MACHINING AND SUBSTRATE REMOVAL PERCENTAGE INPUTS SENSITIVE ANALYSIS ......... 43
FIGURE 3-10 FINISH MACHINING AND SUBSTRATE REMOVAL INPUTS SENSITIVE ANALYSIS .............................. 43
FIGURE 3-11 HEAT TREATMENT PERCENTAGES INPUTS SENSITIVE ANALYSIS .................................................... 44
FIGURE 3-12 INSPECTION PERCENTAGE INPUTS SENSITIVE ANALYSIS ................................................................ 44
FIGURE 3-13 TOOLS SENSITIVE ANALYSIS ............................................................................................................ 44
FIGURE 3-14 COST PERCENTAGES ........................................................................................................................ 46
FIGURE 3-15 COST PER PROCESS WITH AND WITHOUT MATERIAL AND TOOLS COSTS........................................ 47
FIGURE 3-16 PROCESS COST PER HOUR ................................................................................................................ 48
FIGURE 3-17 MODEL COST BOUNDARIES DEPENDING ON PROCESSES ALLOCATION .......................................... 48
FIGURE 3-18 MODEL COST BOUNDARIES DEPENDING ON WAAM ALLOCATION .............................................. 49
FIGURE A2-1 SUBSTRATE CUTTING AND OXIDES REMOVAL PERCENTAGE INPUTS SENSITIVE ANALYSIS ........... 63
FIGURE A2-2 SUBSTRATE CUTTING AND OXIDES REMOVAL PERCENTAGE INPUTS SENSITIVE ANALYSIS ........... 64
FIGURE A2-3 HEAT TREATMENT INPUTS SENSITIVE ANALYSIS ........................................................................... 48
FIGURE A2-4 INSPECTION INPUTS SENSITIVE ANALYSIS ................................................................................. 4965

viii
ix
x
 List of Table

TABLE 2-1 COMPARISON OF VARIOUS WAAM TECHNIQUES. ADAPTED FROM WU ET AL., 2018....................... 7
TABLE 2-2 STRENGTHS AND LIMITATIONS OF WAAM ......................................................................................... 9
TABLE 2-3 DATA FOR LIFE CYCLE INVENTORY FOR THE WAAM UNIT PROCESS. RETRIEVED FROM PRIARONE
ET AL., 2020 ................................................................................................................................................. 17

TABLE 2-4 DATA FOR LIFE CYCLE INVENTORY OF THE WAAM UNIT PROCESS. . RETRIEVED FROM PRIARONE
ET AL., 2019 ................................................................................................................................................. 17

TABLE 2-5 DATA FOR LIFE CYCLE INVENTORY OF THE MACHINING UNIT PROCESS . RETRIEVED FROM
PRIARONE ET AL., 2019 ............................................................................................................................... 18
TABLE 2-6 INPUTS COMPARATION OF THE STUDIES FROM PRIARONE ET AL., 2019 AND CUNNINGHAM ET AL.,
2017 ............................................................................................................................................................. 20
TABLE 3-1 EXPERIENCE OF THE EXPERTS INTERVIEWED ...................................................................................... 31
TABLE 3-2 FIXED AND VARIABLE COSTS .............................................................................................................. 31
TABLE 3-3 SPECIFIC PROCESS INPUTS ................................................................................................................... 33
TABLE 3-4 GENERAL INPUTS ................................................................................................................................ 33
TABLE 3-5 MATERIAL INFORMATION .................................................................................................................. 37
TABLE 3-6 LINE UTILIZATION INPUTS .................................................................................................................. 38
TABLE 3-7 GENERAL INPUTS DATA...................................................................................................................... 38
TABLE 3-8 SPECIFIC PROCESS INPUT DATA .......................................................................................................... 39
TABLE 3-9 VARIABLE COSTS AND PERCENTAGES ................................................................................................ 45
TABLE 3-10 FIXED COSTS AND PERCENTAGES ..................................................................................................... 45
TABLE 3-11 PROCESSES CYCLE TIMES AND PERCENTAGES .................................................................................. 48

xi
xii
 Acronyms

𝐶0 : machine-specific coefficient
𝐶1 : machine-specific coefficient
𝑀
𝑃𝑠𝑡𝑏 : power demand of the machine tool in stand-by mode
𝑡𝑠𝑀 : start-up, set up, workpiece clamping/unloading time
SECiM: specific energy consumption during cutting
C Machining: machining total cost
EMT : electrical energy consumption of the machine tool
t ci : cutting time (h) per tool
𝑇𝑖 : tool life (h) per tool
qL : consumption rate of the cutting fluid (kg/h)
mc: mass of the material that has to be removed
𝑡𝑡𝑐 : tool change time
𝑡𝑐 : cutting time
𝑇: tool life
𝐶𝑀𝑇 : indirect cost rate for machining
𝑀
𝐶𝑀𝑂 : labour charge rate for machining
𝑡𝑀𝑇 : total machining time
𝛼: fraction of time of attendance of machine operator
𝐶𝐸𝐸 : cost of electric energy
𝐶𝑡𝑜𝑜𝑙 : cost of cutting tool
Club: cost of the cutting fluid
C WAAM : WAAM total cost
qG : consumption rate of the shielding gas (l/h)
CW : indirect cost rate for the WAAM system
CMO W: labour charge rate for the WAAM system
t WT: total time for the WAAM process

xiii
𝛽: fraction of time of attendance of the WAAM operator
𝐶𝐸𝐸 : cost of electric energy
𝑃𝑠𝑡𝑏 𝑊 : cost of cutting tool
𝑡𝑠 𝑊 : cost of the cutting fluid
𝑆𝐸𝐶 𝑊 : specific energy consumption during deposition
𝑚𝑤𝑖𝑟𝑒 : mass of the (wire) material that has to be deposited
𝑡𝑑𝑤 : dwell time
𝐶𝑔𝑎𝑠 : cost of the shielding gas
𝑡𝑑 : deposition time
𝑟 : opportunity cost of capital
𝑛 : machine

xiv
 Abbreviations

AM: additive manufacturing

WAAM: wire and arc additive manufacturing
HF-WAAM: hot-forging wire and arc additive manufacturing
LCA: life cycle assessment
LCC: life cycle cost
PBCM: process-based cost model
CNC: computerised numerical control
𝑆𝐸𝐶: specific energy consumption during cutting
𝑀𝑅𝑅: material removal rate

xv
xvi
 1.
INTRODUCTION

The first chapter begins with the context and motivation for conducting the study, fol-
lowed by the problems and objectives. Then, the approach and contributions will be presented
and, finally, the document organisation.

1.1. Context and Motivation

Additive manufacturing (AM) or 3D printing can be defined as the technology that pro-
duces a 3D figure by using deposition, solidification, or fusing methods (Bekker & Verlinden,
2018). AM is a disruptive technology with a broader product customization range when com-
pared to traditional methods. Not only this, but it also offers the possibility to manufacture a
product in one phase, decreasing the lead-time and minimizing material and resource waste
while also diminishing the supply chain steps (Bekker & Verlinden, 2018).
Huang et al. (2013) points out that the benefits of AM compared to traditional processes
can be summarized as follows:

• Material efficiency: Due to the layer upon layer production system, it is possible
to improve the efficiency of the raw material usage when compared with tradi-
tional methods that have a higher rate of waste material.
• Resource efficiency: AM methods do not need as many additional resources as
traditional technologies.
• Part flexibility: AM enables designing a complex part without focusing on as-
sembly or the number of parts.
• Production flexibility: Due to the reduction of bottlenecks, setup times and
number of parts, AM enables the possibility of producing smaller batches with
fewer costs than traditional methods.
There are several technologies or AM processes (Huang et al., 2013), among them wire
and arc additive manufacturing (WAAM) is a technology that prints a piece or product by
wielding layer upon layer (Bekker & Verlinden, 2018). In contrast to other AM technologies,
WAAM allows to build medium to large pieces due to the use of robotics arms and the high

1
deposition rate (Ding et al., 2015). However, compared with other AM technologies, WAAM
products have a problem with the finished surface, which causes a need to use a hybrid pro-
duction with machining to remove the extra material corresponding to the layered surface
(Bekker & Verlinden, 2018). WAAM is a technology that has increased industrial interest due
to the ability to build medium-to-large products with high deposition rates. However, more
studies are needed to understand the economic viability of this technology (Priarone et al.,
2020).
To develop this technology the project smart WAAM was sponsored by EIT raw mate-
rials and the European Union. This project is led by NOVA School of Science and Technology
has as partners the French National Centre for Scientific Research, Norwegian University of
Science and Technology, OCAS NV, and the Université de Bordeaux. The scope of the smart
WAAM is: the microstructural engineering and integration of non-destructive testing meth-
odologies. The objective is to improve the lifetime of industrial components through the im-
plementation of in-line multiparametric non-destructive testing considering a life cycle ap-
proach and contribution to the economic, social and environmental performance of the indus-
trial system.

1.2. Problem and Objectives

In the smart WAAM project, it was necessary to study the cost of WAAM technology.
Cost estimation is one of the essential engineering tools because it allows analysing the eco-
nomic viability of a given process or product. It will enable to understand how a process can
or cannot be implemented. Although this technology has been the subject of study in recent
years, it is still understudied. To fulfil this research gap, in this thesis a life cycle cost model
was developed to evaluate the economic viability of a WAAM machine developed at the Nova
School of Science and Technology of the Universidade Nova de Lisboa (NOVA SST).
A process-based cost model (PBCM) of the WAAM machine was developed to under-
stand the costs related to the technology implementation, considering all associated processes,
as it is a hybrid production model. This study intends is to contribute with a model that esti-
mates the costs of implementing this WAAM technology. The model can be used as a part of
an extended tool to enable a more informed decision making at an early stage.
In this way, a process-based cost model (PBCM) for a product manufactured by NOVA
SST WAAM machine was developed. This model was applied to the production of 500 stain-
less steel AISI316LSI hollow cubes with 7x7x7 cm in order to identify the main driver costs, as
well as the effect that their variation may cause on the implementation costs of WAAM tech-
nology.

1.3. Approach and Contributions

At first, bibliographical research was done in order to characterize and understand the
existing information about AM, particularly WAAM. This information also served as a basis

2
for the subsequent steps. Next, it was described how a life cycle assessment is carried out to
substantiate the necessary methodology. Finally, from the literature review, two works
(Priarone et al. (2019) and Priarone et al. (2020)) related to life cycle cost models for WAAM
were selected for a detailed analysis. Moreover, two cost models (Priarone et al. (2019) and
Cunningham et al. (2017)) for WAAM available in the literature were analysed and compared.
After the elaboration of the state of the art, the WAAM machine to be studied was char-
acterized, as well as the rationale for its existence. Next, the life cycle cost model was devel-
oped with the following steps:
• Goal definition.
• Scope definition: In this stage, the WAAM machine was characterised, as well as
the process to be analysed and the cost model to be used was developed
• Life cycle Inventory: Here, the data necessary to carry out the cost estimative
was obtained from interviews with researchers working at NOVA SST and from
public websites with current market prices.
• Results and discussion: At this stage, the model was validated, and the sensitiv-
ity of the inputs was analysed. Finally, costs were discussed, and the main cost
drivers of the analysed production process were identified.
At the end of the life cycle cost model development, a final analysis was made in order
to conclude the study.

1.4. Structure

This thesis is organised into four chapters. The first chapter is the present one, where an
introduction of the thesis is made. The other chapters are described below:
• 2 State of Art: This chapter conducted a study and literature review on AM, fol-
lowed by a more in-depth literature review on WAAM technology. Another ob-
ject of study was life cycle assessment, where it was described and presented
how to perform it. Finally, two life cycle cost models and two cost models were
analysed. At last, a comparison between the two cost models was made in order
to compare the two approaches.
• 3 Life Cycle Cost Model: This chapter initially established the methodology to
be adopted for the elaboration of the model, as well as the scenario to be ana-
lysed. Next, the context was described, and the causes of the development of the
hot forging WAAM technology were presented. Finally, the life cycle cost model
was executed, where the machine was characterised, the cost model was devel-
oped and applied, and at the end, the results were analysed.
• 4 Conclusion: In this chapter, the last considerations are carried out in order to
conclude the study.

3
 2.
State of Art

Additive Manufacturing
Over the last decades, there has been a significant development in AM technologies.
AM or 3D printing is a technology that consists of producing a 3D figure by using deposition,
solidification, or fusing methods to build it (Bekker & Verlinden, 2018). These technologies
have a massive potential because, while traditional technologies use subtractive processes to
make an object, AM technologies build products layer by layer based on a 3D model data
(ASTM, 2021). These methods enable the flexibility and ability to produce complex compo-
nents with less material consumption (Priarone et al., 2019).
AM technology is composed of different types of processes that build products in dif-
ferent ways. Some processes use inkjet-type printing heads to spray binder or solvent into
powdered ceramics or polymer. While others use laser or electron beams to melt or sinter
metal or plastic powder together (Huang et al., 2013). Huang et al. (2013) enunciated some of
the most used AM technologies.

• Fused deposition modelling;

• Inkjet printing;
• Laminated object manufacturing;
• Laser engineered net shaping;
• Stereolithography;
• Selective laser sintering;
• Electron Beam Melting;
• Digital light processing.
AM gives enormous opportunities for the industry and society. It could be the next step
toward a revolution in the industry. Huang et al. (2013) identify the three top benefits of AM
to society as the following:

• Healthcare products customized to the needs of individual consumers which is

expected to improve population wellbeing significantly;
• Reduced raw material usage and energy consumption, which is a key contribu-
tion to environmental sustainability;
4
 • On-demand manufacturing, which presents an opportunity to reconfigure the
manufacturing supply chain to bring cheaper products to consumers faster
while utilizing fewer resources.
AM is a disruptive technology that gives an ample opportunity to the industry since it
offers a large customization of products, the possibility to produce a product in one phase,
decrease the lead time, reduce material and resources waste, and diminish the supply chain
steps (Bekker & Verlinden, 2018).
Design innovation is another opportunity linked to the utilization of AM technologies
since it enables the production of complex objects without the necessity of using several pro-
cesses. This allows designers to focus on the product design instead of focusing on design for
manufacturing and design for assembly. This new way of looking at design allows the
product's manufacture near the clients, which will reduce the supply chain since objects will
have fewer components. This will directly impact warehouses, transportation, and packaging,
which will not be needed like they are now (Huang et al., 2013).
The advantages of AM compared to traditional processes can be summarized as Huang
et al. (2013) proposed:

• Material efficiency: While subtractive methods have a high rate of wasted mate-
rial, AM, due to the layer upon layer production system, has a low rate of wasted
material and can use the leftover material to produce new products.
• Resource efficiency: AM methods do not need additional resources like tradi-
tional technologies do (ex., cutting tools).
• Part flexibility: As mentioned above, AM permits designers not to focus on the
manufacturing or assembly processes, enabling the production of complex prod-
ucts with less parts.
• Production flexibility: The AM machines do not need set up, enabling the effi-
cient production of small batches. This technology reduces the production bot-
tlenecks due to its diminishing of pieces. Another reason is the fact that the prod-
uct quality only depends on the process and not on the operator, like traditional
methods.
AM is a disruptive technology that needs to be developed since it has some disad-
vantages like Huang et al. (2013) proposed:

• Size limitations: In most cases of AM technologies, the size of the printer repre-
sents a constraint to the dimensions of the product to be built. Another problem
is the lack of strength of materials like liquid polymers and resin’s powders that
do not permit the production of bigger pieces. Another problem is the reduced
deposition rate, which would imply a significant amount of time to produce a
large piece.
• Imperfections: Most of the products produced by AM methods require a finish-
ing surface step, due to the layer-by-layer method, most of the time, the surface
seems to be unfinished.
5
 • Cost: The AM equipment is costly, but over the following years, due to the
growth of this market, it seems probable that the costs will decrease.

WAAM
WAAM is a technology that uses a robotic arm to print a piece or product by wielding
layer upon layer (Bekker & Verlinden, 2018). This method is based on an energy deposition
on a metal wire (raw material), when the metal is melted, it is deposited layer-by-layer, by a
robotic arm, on a substrate to build a component or product (Priarone et al., 2019).
In contrast to other AM technologies, WAAM, theoretically, allows to build medium to
large pieces with a high deposition rate (Ding et al., 2015). This technology presents novel
opportunities in fabrication, repair, and refurbishment of products (Kokare et al., 2022). An-
other important point is that WAAM has a sustainability potential due the lesser material
removal and shorter lead times, when compared to traditional methods (Seow et al., 2020).
However, WAAM technology has a problem with the finishing step caused by the layered
surface finish, and this requires a machining step to remove the excess material (Bekker &
Verlinden, 2018). A good example was the study where Bekker & Verlinden (2018) proposed
a hybrid approach of WAAM and Computerized Numerical Control (CNC) milling for build-
ing a complex part with low material utilization. In this case, the outline was produced with
extra thickness to be milled to guarantee a suitable surface finishing. Nevertheless, with the
development of this technology and the consequent improvement of parameters, equipment
and printing strategies, the problem related to the layered surface finish will be minimized
over time (Bekker & Verlinden, 2018).
Priarone et al. (2020) pointed out that a limited number of studies have compared
WAAM with other technologies regarding environmental and economic competitiveness. So,
the decision on what technology to use should be made by a comparison of the WAAM tech-
nology with traditional technologies. This comparison needs to consider the material charac-
teristics too. Material properties like isotropy and tensile strength may change with the uti-
lized material, like the tolerance and geometric possibilities (Bekker & Verlinden, 2018). How-
ever, investigations about the properties of WAAM products have shown promising results.
In the case of titanium, the ductility is similar to extruded titanium, and the strength is only
10% less than extruded products. In the case of fatigue life of titanium produced by WAAM,
it exceeded the titanium products produced by extrusion in most tested specimens (Wang et
al., 2013).
WAAM technology typically can be divided into three types, depending on the heating
source it uses (Wu et al., 2018): Gas Metal Arc Welding (GMAW)-based, Gas Tungsten Arc
Welding (GTAW)-based, and Plasma Arc Welding (PAW)-based. In table 2-1 the main char-
acteristics of each type of WAAM are described.
Figure 2-1 shows a WAAM process based on gas metal arc welding. This approach re-
sults in a metal deposition as an electric wire by creating an arc with DC power. During the
process, an inert gas is added to shield oxygen and pollutants from the weld pool (Bekker &
Verlinden, 2018).

6
Table 2-1 Comparison of various WAAM techniques. Adapted from Wu et al. (2018)
Energy
WAAM Features
source
GTAW-based GTAW Non-consumable electrode; Separate wire feed process;
Typical deposition rate: 1-2 kg/hour;
Wire and torch rotation are needed;
GMAW-based GMAW Consumable wire electrode;
Typical deposition rate 3-4 kg/hour;
Poor arc stability, spatter;
Cold metal Reciprocating consumable wire electrode;
transfer Typical deposition rate: 2-3 kg/hour;
(CMT) Low heat input process with zero spatter, high process tol-
erance;
Tandem Two consumable wires electrodes;
GMAW Typical deposition: 6-8 kg/hour;
Easy mixing to control composition for intermetallic mate-
rials manufacturing;
PAW-based Plasma Non-consumable electrode; Separate wire feed process;
Typical deposition rate 2-4 kg/hour;
Wire and torch rotation are needed;

Figure 2-1 Schematic of gas metal arc welding. Adapted from Bekker & Verlinden (2018)

WAAM techniques use metals that are commercially available in the form of wires.
These wires can be of various materials. Wu et al. (2018) state that the most used materials to
produce WAAM products are:

• Titanium alloys: thanks to the high strength-to-weight ratio and high cost of titanium
alloys, the aerospace industry has widely studied the application of AM in aerospace
components. The utilization of WAAM to produce titanium components has enor-
mous business potential, particularly for large-sized components with complex struc-
tures.

7
 • Aluminium alloys: although there have been several successful trials for manufactur-
ing aluminium components, there are some problems. Firstly, the production of this
material by WAAM is only suitable for large and complex thin-walled structures since
the cost of producing small and simple parts by traditional methods is lower. On the
other hand, some aluminium series, such as AL 7xxx and AL 6xxx, are difficult to cast
due to turbulent melt pools and weld defects, which occur during the deposition
phase. Finally, the mechanical properties of the components produced are inferior to
those manufactured by traditional methods.
• Nickel-based superalloys: this is the second most studied material to apply to WAAM
due to their high strength and outstanding oxidation resistance at high temperatures.
Nickel-based superalloys are widely used in aerospace, aeronautical, chemical, petro-
chemical, and marine industries.
• Other metals: other metals have been studied to produce WAAM components, such
as magnesium alloy AZ31 for automotive applications, Fe/Al intermetallic com-
pounds, Al/Ti compounds, bimetallic steel/nickel, and steel/bronze parts for the
aeronautic industry. To produce intermetallic components with accurate pre-designed
composition is still a challenge for WAAM.

2.2.1. Strengths and Limitations of WAAM

WAAM is a technology with advantages when compared to other technologies, alt-
hough there are a limited number of studies on this technology (Priarone et al., 2020).
One of the strengths of AM and WAAM is the ability to build lightweight structures like
the sofa presented in figure 2-2, this could not be produced by CNC casting or other traditional
methods (Bekker & Verlinden, 2018).

Figure 2-2 Dragon Bench sofa, made in stainless steel by WAAM. Retrieved from Bekker
& Verlinden (2018)

8
 Another strength of WAAM technology is the high material efficiency and freedom of
shape, which decreases material consumption (Bekker & Verlinden, 2018). This ability to
make the process more efficient reduces the buy-to-fly ratio for high-cost materials, such as
titanium alloys and nickel alloys. On the other hand, it also reduces production times and
costs and increases the productivity of design and prototyping (Li et al., 2019).
A further limitation of WAAM is the reduced lifespan of the manufactured products.
This is due to the use of extremely high heat sources that cause a reduction in the effective
fatigue life and tensile strength (Li et al., 2019).
One of the advantages of WAAM, when compared to other AM technologies and tradi-
tional methods, is the ability to produce large components that theoretically could be of any
size (Bekker & Verlinden, 2018). On the other hand, one of the limitations of WAAM is the
low dimensional accuracy and a reduced feature resolution (Priarone et al., 2019). As it hap-
pens with tensile strength, the low dimensional accuracy is due to the use of extremely high
heat sources, which cause residual stress and distortion (Li et al., 2019). However, Colegrove
et al. (2013) developed a method of reducing these effects by using two rollers, a profile roller
and a slotted roller, which reduced surface roughness and distortion, with the slotted roller
being more effective.
Priarone et al. (2019) studied the comparisons between two methods: a standard milling
process and a combination of WAAM and milling. This study allowed to make a comparison
while the solid to cavity rate varied, this ratio is defined as the ratio between the mass of the
final piece and the mass of the product without a cavity. The conclusion was that when the
ratio increases, the WAAM technology turns into a less efficient option.
Due to the high temperatures used to melt the wire metal during the WAAM process, it
is necessary to pay attention to the cooling rate, as well as the existence of residual stress. This
leads to a limitation for WAAM technology, as it cannot deliver finished products, which have
to be post-processed to obtain the desired shape and size (Singh et al., 2021). A summary of
the strengths and limitations found in the literature are refereed in table 2-2.

Table 2-2 Strengths and limitations of WAAM

Strengths Limitations
Ability to build lightweight structures Decrease in the lifetime of the metal components
Raw materials efficiency Low dimensional accuracy
Produce parts of any size Low efficiency for high solid-to-cavity ratios
Freedom of shape Need for post-processing
High deposition rate

Life Cycle Cost (LCC)

Although this thesis's objective is to develop a life cycle cost model, in appendix 1 an
introduction to LCA is made to better understand the methodology. In recent decades, the
research based on the life cycle assessment has been growing, as well as the scientific publi-
cations in this area. Thus, life cycle assessment is a technique to analyse all the life cycle of a
process, product, or system, which are: raw materials, design, material preparation, produc-

9
tion, post-processing, use, maintenance and end of life (Ribeiro et al., 2020). This method con-
sists of 4 steps: definition of the goal and scope, inventory of the materials and resource flows,
assessing the environmental/social impact or cost of the system and interpretation of the re-
sults (Bekker & Verlinden, 2018).
When a disruptive technology is being studied, it is necessary to analyse its cost-effec-
tiveness compared with traditional concurrent methods. This necessity happens due to the
role that cost has as a factor in the decision making to analyse the economic viability of prod-
ucts (Ribeiro et al., 2020)

2.3.1. Goal Definition

The objective of a life cycle cost (LCC) is to build a methodology to track the economic
viability of a product manufactured using WAAM technology. To achieve the study's goal, an
existing model, either a single model or a combination of several, can be used to calculate all
costs. The life cycle may involve all phases from raw material to end-of-life, or it may focus
on a smaller number of life phases, like a cradle-to-gate approach. The cradle-to-gate approach
is more used in comparative studies when two similar products produced with different tech-
nologies are evaluated since, depending on the technology used, there are analogous phases,
such as obtaining the raw material, use, or maintenance.
In the case of a LCC that studies products manufactured by WAAM technology, the
motivation is usually related to the fact that WAAM is a recent and disruptive technology
since it is an additive technology that allows the construction of products from medium to
large sizes, but it has some limitations like the poor finish surface, that usually needs machin-
ing processes after WAAM. This leads to the necessity to compare this production approach
with the traditional ones. Another complementary aspect is that it is a technology whose eco-
nomic viability is not yet widely studied, thus requiring further investigations.
Another important point is the need to explicitly state the type of production used, as
this can impact the results since an integrated approach does not lead to the same results as a
single approach. In the case of the study done by Priarone et al. (2020), a WAAM based ap-
proach was used, with the substrate done by forming and machining, to compare it with a
traditional machining technique. In the case of the work done by Priarone et al. (2019), a
WAAM based approach was used with the substrate manufactured by traditional methods to
compare with a traditional method. In these two examples, it was also necessary to mention
that they were comparative studies. In the studied developed by Kokare et al. (2022), it was
proposed a digital platform were the LCA and LCC can be obtain for products manufactured
by WAAM. Once again it is stated the kind of production that is used, being a WAAM process
with post processing machining.

2.3.2. Scope Definition

The first step of the scope definition is the description of the product or products to be
studied, as well as the type of metal to be used and the WAAM machine that will produce the
product, and if it is a comparative study, it must enunciate the type of traditional process that

10
will serve as a comparison. For example, in Priarone et al. (2020), a comparative study is pro-
posed to produce three products:

• a cantilever beam made with the ER70s-6 steel, whose standard process is ma-
chined from billet,
• an aerospace bracket made with Ti-6Al-4V titanium, whose process is machined
from forging,
• and an aerospace frame with the AA2319 aluminium, whose traditional process
is machined from billet.
For those comparison Priarone et al. (2020) used two different types of WAAM config-
urations. A configuration was used for the first two products that rely on plasma-arc power
sources. The last one used a setup that relies on Cold Metal Transfer as a deposition process.
One of the fundamental points of the scope definition is the choice of the functional unit.
Usually, in LCCs that study WAAM technology, the functional unit is the production of a
single part. Another important point is the definition of productivity and economic metrics,
usually considered production time and cost.
Another critical step of the scope definition is the characterisation of the study bounda-
ries. This step depends on the life cycle model chosen to perform the research. In the case of
the two works analysed here, a cradle-to-gate approach is followed, and therefore the bound-
aries include the production of raw materials, the pre-manufacturing phases to produce the
incoming feedstock materials and all the manufacturing steps leading to the production of a
finished product.
After characterising the study boundaries, it is necessary to draw up a diagram repre-
senting the unit processes to be studied and their respective elementary flows. In the study
from Priarone et al. (2019), the unit processes are: raw material production, hot rolling for the
workpiece and the material that will feed the WAAM process, wire drawing, substrate ma-
chining, WAAM process, and finish machining. These processes can be observed in the dia-
gram of unit processes pictured in figure 2-3, where the elementary flows between each unit
process are also displayed. Since their study has an economic and environmental analysis, the
energy flows and emissions are also represented. In the case of the works Kokare et al. (2022)
and Kokare et al. (2022) the production system boundaries were Identified, being study only
the production phase.
In the study by Priarone et al. (2020), a diagram was drawn up, which can be seen in
figure 2-4. There are similarities with the diagram represented in figure 2-3. The main differ-
ence is the addition of the unit processes and their elementary flows, that are related to the
traditional process to be compared. In the case of the WAAM process, the unit processes are
raw material production, wire production, substrate production, WAAM and finish machin-
ing. The unit processes of the traditional process are raw material production, working piece
production and rough and finish machining.
The elaboration of diagrams such as the ones illustrated in Figure 2-3 and Figure 2-4 is
very important because it allows the conception of the production system. This step is bene-
ficial for the life cycle inventory phase. It provides a better understanding of the production
system, the resource flows that occur during the production of WAAM products, and the costs
11
related to each unit process. In this way, it will be easier to identify the data needed to carry
out the study in the life cycle inventory phase.

Figure 2-3 Diagram of unit processes. Adapted from Priarone et al. (2019)

The last step in the scope definition of the LCC is the presentation of the cost model used
to carry out the study. This step is very important as it is directly related to the life cycle in-
ventory and to the discussion of results. Two different cost models are presented below, used
to calculate the costs associated with the production of one or several WAAM products.

Figure 2-4 Diagram of unit processes. Adapted from Priarone et al. (2020)

12
 In the study from Priarone et al. (2019), two cost models were proposed, one for each
process unit, namely the substrate and finish machining and the WAAM process. For this,
several types of costs were considered, for the machining unit process, indirect costs were
considered, such as equipment depreciation costs and production and administration over-
heads. The labour costs associated with the production and the direct costs, such as electric
power and cutting tools, were also considered. To obtain the total machining time, the start-
up time, set up time, workpiece clamping/unloading time, the cutting time, and the time for
the tool changes were considered.

To calculate the energy consumption of the substrate machining and finish machining
processes, the model proposed by Kara & Li (2011) was used, where it is presented a model
of unit process energy that correlates the specific energy consumption with the material re-
moval rate. This correlation is shown in equation 1. The electrical energy consumption of the
machine tool during the idle, cutting and tool change operational modes is presented in equa-
tion 2. The machining costs were calculated based on equation 3.
𝐶1
𝑆𝐸𝐶 = 𝐶0 + (1)
𝑀𝑅𝑅

Where:
𝑆𝐸𝐶: specific energy consumption during cutting (kWh/kg);
𝐶0 : machine-specific coefficient;
𝐶1 : machine-specific coefficient;
𝑀𝑅𝑅: material removal rate.
𝑛
𝑡𝑐𝑖
𝐸𝑀𝑇 = 𝑀
𝑃𝑠𝑡𝑏 ∙ 𝑡𝑠𝑀 + ∑ (𝑆𝐸𝐶𝑖 𝑊 ∙ 𝑚𝑐𝑖 + 𝑃𝑠𝑡𝑏 𝑀 ∙ 𝑡𝑡𝑐𝑖 ∙ ) (2)
𝑇𝑖
𝑖=1
Where:
𝐸𝑀𝑇 : electrical energy consumption of the machine tool;
𝑀
𝑃𝑠𝑡𝑏 : power demand of the machine tool in stand-by mode (kW);
𝑀
𝑡𝑠 : start-up, set up, workpiece clamping/unloading time (h);
𝑆𝐸𝐶𝑖𝑀 : specific energy consumption during cutting (kWh/kg);
𝑚𝑐 : mass of the material that has to be removed (kg);
𝑡𝑡𝑐 : tool change time (h);
𝑡𝑐 : cutting time (h);
𝑇: tool life (h).
𝑛
𝑀𝑎𝑐ℎ𝑖𝑛𝑖𝑛𝑔 𝑀 𝑡𝑐
𝐶 = 𝐶𝑀𝑇 ∙ 𝑡𝑀𝑇 + 𝐶𝑀𝑂 ∙ 𝛼 ∙ 𝑡𝑀𝑇 + 𝐶𝐸𝐸 ∙ 𝐸𝑀𝑇 + ∑ (𝐶𝑡𝑜𝑜𝑙𝑖 ∙ 𝑖 ) + 𝐶𝑙𝑢𝑏 ∙ 𝑞𝐿 ∙ 𝑡𝑐 (3)
𝑇𝑖
𝑖=1
Where:
𝐶𝑀𝑎𝑐ℎ𝑖𝑛𝑖𝑛𝑔 : machining total cost;
𝐸𝑀𝑇 : electrical energy consumption of the machine tool;
𝑡𝑐𝑖 : cutting time (h) per tool;
𝑇: tool life (h) per tool;
𝑞𝐿 : consumption rate of the cutting fluid (kg/h);
𝐶𝑀𝑇 : indirect cost rate for machining (€/h);
𝑀
𝐶𝑀𝑂 : labour charge rate for machining (€/h);
𝑡𝑀𝑇 : total machining time (h);

13
 𝛼: fraction of time of attendance of machine operator (𝛼 ≤ 1);
𝐶𝐸𝐸 : cost of electric energy (€/kWh);
𝐶𝑡𝑜𝑜𝑙 : cost of cutting tool (€);
𝐶𝑙𝑢𝑏 : cost of the cutting fluid (€/kg).

Still, Priarone et al. (2019) used a model similar to equation 3, where the same system of
boundaries is used to calculate the WAAM product's costs. This model also considered indi-
rect costs, the costs associated with energy consumption, for which idle time, deposition time
and dwell time were considered. The dwell time should be chosen carefully to avoid the col-
lapse or re-melting of the deposited material. To obtain the costs associated with producing
this type of product, the cost of shielding gas was also considered. This model is represented
in equation 4.
𝐶 𝑊𝐴𝐴𝑀 = 𝐶𝑊 ∙ 𝑡𝑊𝑇 + 𝐶𝑀𝑂 𝑊 ∙ 𝛽 ∙ 𝑡𝑊𝑇 + 𝐶𝐸𝐸 ∙ (𝑃𝑠𝑡𝑏 𝑊 ∙ 𝑡𝑠 𝑊 + 𝑆𝐸𝐶 𝑊 ∙ 𝑚𝑤𝑖𝑟𝑒 + 𝑃𝑠𝑡𝑏 𝑊 ∙ 𝑡𝑑𝑤 ) + 𝐶𝑔𝑎𝑠 (4)
∙ 𝑞𝐺 ∙ 𝑡𝑑

Where:
𝐶 𝑊𝐴𝐴𝑀 : WAAM total cost;
𝑞𝐺 : consumption rate of the shielding gas (l/h);
𝐶𝑊 : indirect cost rate for the WAAM system (€/h);
𝐶𝑀𝑂 𝑊 : labour charge rate for the WAAM system (€/h);
𝑡𝑊𝑇 : total time for the WAAM process (h);
𝛽: fraction of time of attendance of the WAAM operator (𝛽 ≤ 1);
𝐶𝐸𝐸 : cost of electric energy (€/kWh);
𝑃𝑠𝑡𝑏 𝑊 : cost of cutting tool (€);
𝑡𝑠 𝑊 : cost of the cutting fluid (€/kg).
𝑆𝐸𝐶 𝑊 : specific energy consumption during deposition (kWh/ kg);
𝑚𝑤𝑖𝑟𝑒 : mass of the (wire) material that has to be deposited (kg);
𝑡𝑑𝑤 : dwell time (h);
𝐶𝑔𝑎𝑠 : cost of the shielding gas (€/l);
𝑡𝑑 : deposition time (h).

Although it is not a LCC study, Cunningham et al. (2017) propose a cost model and
sensitivity analysis for WAAM technology. This cost model consists of a time-based activity-
based costing approach. For the elaboration of this model, the activities necessary to produce
WAAM products were considered, which are: substrate preparation, set up, deposition, cool-
ing, rolling, heat treatment, machining, substrate removal and inspection. The activities such
as wire deposition, cooling and rolling are repeated cyclically for each deposited layer. The
costs of the activity rates were based on the sum of direct and indirect costs for each activity,
however, costs associated with sales expenses, such as administration, management, and fa-
cility, were excluded from the model. This study was done on two different products pro-
duced using WAAM technology compared to two powder-based metal AM processes, which
are Electron Beam Melting and Direct Melting Laser Sintering and conventional CNC machin-
ing.
To implement the cost model proposed by Cunningham et al. (2017), machine data was
defined as fixed operational parameters between setups, user inputs were defined as variables

14
that change between setups, indirect costs were generated mainly from machine data, and
direct costs were generated from the information presented above and based on market prices.
To carry out Cunningham et al. (2017) study, data was generated both from material
removal rates, such as substrate preparation, as from software Autodesk, which was used to
generate the tool paths of deposition for both products. Data was also obtained from calcula-
tions, such as active deposition and motion times, and from literature, such as set up, rolling
and cooling times, which limited the study since the data was based on the use of Titanium
Alloy Ti6A14V, and therefore it was necessary to consider the use of this material. For this
study, some assumptions were made to simplify the model. One of these simplifications was
related to indirect costs, in which estimations were made based on labour and machine over-
heads. A 46% capacity utilisation at 16 operational hours for 252 days was assumed to calcu-
late the indirect depreciation costs. The labour cost was assumed as an indirect cost, and it
was considered that the WAAM machine required constant observation. The inspection costs
were not considered due to the inexistence of similar data for other products.
For the WAAM processes, a summary metrics approach developed by Baumers et al.
(2012) was used to provide an equivalent cost per kilogram of product. Naturally, the value
was updated to 2017, which is the date of Cunningham et al. (2017) article. The comparison
was made for WAAM BTF ratios of 5, 10, 15, and 20.
To find out which were the key cost drivers in the model developed by Cunningham et
al. (2017), a sensitivity analysis, more specifically a One-Factor-at-a-Time sensitivity analysis
approach, was used. For this, the maximum and minimum values for the machine and user
inputs were considered, considering realistic industry values and price fluctuations. Industry
indexes were used to obtain average values for the variables and rates. To calculate the sensi-
tivity indices, equation 5 was used.
𝐷𝑚𝑎𝑥 − 𝐷𝑚𝑖𝑛
𝑆𝑒𝑛𝑠𝑖𝑡𝑖𝑣𝑖𝑡𝑦 𝑖𝑛𝑑𝑒𝑥 = (5)
𝐷𝑚𝑎𝑥

In the case of the study carried out by Priarone et al. (2020), the same cost model devel-
oped by Priarone et al. (2019) was used. This cost model considered the cost of consumables,
production costs, including set up, work frame calibration, substrate preparation and part
building, labour costs, delivery costs and costs related to overheads and facility charges. This
cost model was used for both the WAAM process and the traditional process. As this study is
a Life Cycle assessment, cost is one of the categories of the decision model developed.

2.3.3. Life Cycle Inventory

The objective of the life cycle inventory is to collect and compile data regarding each
elementary flow of the processes required to produce WAAM products. In the case of a com-
parative study, it is necessary to collect and compile data regarding the elementary flows of
the processes that will be compared to the WAAM technology. The first step is to clarify where
the data used comes from. Usually, the data for the process to be studied (WAAM) is collected
on an experimental basis, while the data for the processes used as comparisons are taken from
databases. Although, according to Priarone et al. (2020), the data for the WAAM process was
15
obtained from the Welding Engineering and Laser Processing Centre at Cranfield University,
and the raw material production, pre-manufacturing phases, and machining data were col-
lected from the Cambridge engineering selector database.
In the study carried out by Priarone et al. (2019), the first step was the case study expla-
nation, where a product was produced with the combination of a substrate produced through
machining and a top produced through WAAM technology. This process was compared with
standard milling. One of the objectives of the study was to compare the two processes based
on the solid to cavity ratio, which can be defined as the ratio between the final mass of the
product and the mass that would be contained in the bounding volumetric envelope of the
part itself (Campatelli et al., 2020). The data concerning the manufacturing processes were
experimentally characterised, and the data concerning the materials production and the pre-
manufacturing phases were collected from the literature.
The next phase of a life cycle inventory is the material production and pre-manufactur-
ing characterisation. In the case of the work from Priarone et al. (2020), it was necessary to
quantify the wastes. As such, it was considered a rate of 1.14 for wire drawing and 1.05 for
hot rolling, this rate is the ratio between the material input and output. The WAAM process
waste was quantified based on laboratory experiments, and the leading causes were in-pro-
cess material vaporisation, small droplets of molten material and wire scraps. For the WAAM
process, it was assumed a material utilization efficiency of 98%. The costs of feedstock mate-
rials were obtained based on market values, with a 15% range of variation. In the study by
Priarone et al. (2019), the values used for the wastes were slightly different since the ratio
between input and output for hot rolling was assumed to be 1.05, and 1.12 for wire drawing
and the material losses for the WAAM process were neglected. The values used for the life
cycle inventory of the WAAM unit process, and the machining unit process are listed in table
2-4 and table 2-5, respectively.
The last life cycle step characterises the WAAM unit and machining unit processes. In
this phase, the values assumed for the model's realisation are inventoried. In the case of a LCC
study, it is necessary to state the costs necessary to produce the product, the machine param-
eters, and the process-related consumption, such as energy consumption.
In the study of Priarone et al. (2019), the life cycle inventory of the unit process begins
by identifying the machine that was used for the WAAM process, as well as the characterisa-
tion of the deposition, such as the wire diameter, the welding speed, the wire feed speed and
the flow rate during deposition. The specific energy consumption was measured as well. The
process, dwell time, start-up time, set up, and post-processing was considered. For the calcu-
lation of the indirect costs, it was assumed the methodology from (Baumers et al., 2012), for
which it was considered the administrative and production costs, a machine depreciation of
8 years, an annual operation time of 5000 h/y and a maintenance and consumables cost of 6%
of the machine cost, per year. For the cost of the WAAM machine, it was also considered the
energy source and the robotised handling system. The shielding gas flow rate was measured.
The annual (1616 working hours) labour cost was estimated, assuming that during set up and
post-processing, the operators, are working and the existence of supervisors. A similar ap-
proach to the one mentioned above was taken in the life cycle inventory for the machining
unit process. The specific electric energy was obtained through equation 1. The standby power
16
and the idle time were considered to calculate the energy consumption. The parameters were
chosen based on the recommendations of the supplier's catalogue. Due to economic and en-
vironmental feasibility, dry cutting conditions were chosen. The material removal rates for
roughing and face milling, semi-finishing, and finishing operations were considered. The cost
of the machine was evaluated, the indirect costs were estimated, and the purchase cost of the
endmill was considered. The tool life was assumed for roughing, semi-finishing and finishing
operations.
A similar methodology to the one used by Priarone et al. (2019) was followed in the
Priarone et al. (2020) study. For the life cycle inventory for the WAAM unit process, average
deposition rates and the specific electric energy were taken for aluminium, steel, and titanium
and are listed in table 2-4.
The energy consumption associated with the movements was neglected, and to correct
the electric energy demand to primary energy, a coefficient of 0.38 was used. The cost associ-
ated with tools was not considered since this is a per part study based on a series production.
The WAAM process and auxiliary equipment's energy consumption and the shielding gas
consumption were monitored during the productive and non-productive times. In the case of
the life cycle inventory for the machining, the unit process times were estimated based on the
material removal rates. The specific electric energy was presented for each process and for
each material used regarding the energy consumption. The methodology used to identify the
ranges of cost variation was the same presented in the study of Priarone et al. (2019).
Table 2-3 Data for life cycle inventory for the WAAM unit process. Retrieved from Priarone et al. (2020)
Data Aluminium Steel Titanium
Average deposition rates 2,40 kg/h 0,94 kg/h 0,66 kg/h
Specific electric energy 63 MJ/kg 23.7 MJ/kg 33.4 MJ/kg

Table 2-4 Data for life cycle inventory of the WAAM unit process. Retrieved from Priarone et al. (2019)
WAAM
Wire diameter 0,8 mm
Welding speed 300 mm/min
Wire feed speed 4,55 m/min
Flow rate 38.13 m3/s
Specific energy consumption 1.36 kWh/kg
Dwell time 100 s
Time for start-up, set up and post processing 30 min
Administrative and production cost 5.46 €/h
Indirect cost 8,61 €/h
WAAM machine cost 85 000 €
Shielding gas flow rate 14 l/min
Shielding gas cost 2,14 €/m3
Workpiece cost 0,50 €/kg
Wire cost 1,80 €/kg
Annual labour cost 35 000 €

17
Table 2-5 Data for life cycle inventory of the machining unit process. Retrieved from Priarone et al. (2019)
Machining
C0 3,524
C1 2066
Standby power 2,2 kW
Idle time 15 min
Material removal rates for roughing and face milling 149.2 mm3/s
Material removal rates for semi-finishing 35.8 mm3/s
Material removal rates for finishing 6.0 mm3/s
Machine cost 200 000 €
Indirect cost 12,86 €/h
Endmill cost 20 €/tool
Roughing tool life 30 min
Semi-finishing tool life 45 min
Finishing tool life 45 min

2.3.4. Results and Discussion

In this phase, calculations are made to obtain the product's life cycle costs. In the case of
a comparative study, the costs associated with the product or set of products produced by
other means of production are also computed. These costs are then presented. At the discus-
sion stage, the results and features are commented as well as the changes that they may cause,
since they are correlated with the performance indicators, such as the process efficiency when
changing the solid-to-cavity rate.
The results and discussion phase of the study carried out by Priarone et al. (2019) pre-
sented the energy consumption, production costs, and manufacturing times. These values are
presented for the two approaches studied, considering the data reported in the life cycle in-
ventory phase and using the cost and energy models presented in the scope definition phase.
Next, some considerations about the results are exhibited. One of the considerations is the
relationship between the efficiency of the process and the solid-to-cavity rate. Although the
integrated approach is generally more time-consuming than the traditional approach, the
solid-to-cavity rate increases the time consumption for the WAAM integrated process in-
creases as well. This happens because there is a linear relationship between the time spent for
the WAAM process and the increase of the material to be deposited when the deposition rate
is fixed. Regarding costs, it should be noted that manufacturing expenses have the most
weight in the total cost since the cost of feedstock material is low.
In the study conducted by Priarone et al. (2020), the results concerning the economic
perspective were also presented. The aluminium frame was the product that presented a
much lower cost for the WAAM approach when compared with the conventional approach.
The production cost of the titanium bracket was slightly higher for the WAAM approach, and
the manufactured steel beam through the WAAM approach had a much higher cost than the
conventional one. In terms of efficiency, measured through production time, both approaches
had similar values for aluminium frame and titanium bracket, but in the case of steel beam,

18
the production time of the WAAM approach was much higher than the traditional one. In
terms of the products’ mechanical characteristics, yield strength and tensile strength were
found to be slightly lower for the WAAM approach, and in some cases, it was verified a higher
elongation at break values than the expected values for the parental materials.

2.3.5. Conclusions
The conclusions are the last stage of a LCC study, in which it is verified if the goal and
scope were reached, and the conclusions of the study are presented. In the study conducted
by Priarone et al. (2019), the conclusions state that the goal of developing a framework that
can be used to assess and compare the performance of a WAAM based approach and machin-
ing has been achieved. It is also highlighted that the main factor considered for the study was
the solid-to-cavity ratio. As this study also has an environmental component, it was concluded
that the WAAM approach is favourable due to the efficient use of raw materials, however, it
still presents higher costs and production times than the conventional methods. Finally, the
fact that the breakeven point varies with the solid-to-cavity ratio proved the importance of
having a decision tool for the production phase.
In the study carried out by Priarone et al. (2020), the purpose of comparing WAAM
based additive/subtractive manufacturing approaches and machining was achieved. It was
further noted that the cost and time of production depend on the component to be produced
and the material to be used. This is due to the solid-to-cavity ratio's importance for the results.
Finally, although the results obtained in this study are only valid for the conditions assumed,
the model developed is useful as a decision support tool when considering environmental
and economic factors.
In the work of Kokare et al. (2022) It was concluded that the product produced by
WAAM was more economical to produce In batches of 4, being the cost of one piece 52.31 €
and 18.45 € for a batch of 4 pieces. In another study it was possible to Identify the main cost
drivers In the production of a specific component, which where In the first place labour cost
and followed by the machine cost (Kokare et al., 2022)

Cost Models a Comparative Analysis

To better understand the similarities and differences in the existing WAAM cost models,
table 2-6 shows the input data for the cost models proposed in the work of Priarone et al.
(2019) and Cunningham et al. (2017).
It is clear that the Priarone et al. (2019) model does not analyse the same activities as
Cunningham et al. (2017) model.
The difference between these two models happens because Cunningham et al. (2017)
model is an activity-based cost model that focuses on the costs of all the activities involved in
the production of WAAM products, considering the stages of heat treatment, substrate re-
moval and inspection, which are not present in Priarone et al. (2019) model. It should be noted
that the substrate removal stage is not considered because the substrate was produced
through traditional methods in order to be a part of the final part, thus facilitating the process.

19
 Another important fact is that Priarone et al. (2019) model lumps the substrate machin-
ing and finish machining stages with the same inputs to calculate the associated costs. This
happens because both processes are based on machining to remove material. Because the cost-
ing methodologies for each process are different also means that the data used In the two
models may be different, even if they are used to cost the same task. These differences are
noticeable since, in Priarone et al. (2019) model, costs are calculated for each unit process,
whereas in Cunningham et al. (2017) model, cost rates result from the sum of direct and indi-
rect costs for each activity.
Another difference between the two models is the way of quantifying energy consump-
tion. In Cunningham et al. (2017) model, energy consumption only appears in some activities,
excluding WAAM, and appears as a rate. In Priarone et al. (2019) model, energy consumption
is calculated and well described.
Table 2-6 Main cost model inputs
Priarone
Cunningham
Activity Input Data et al.
et al. (2017)
(2019)
Substrate size X
Substrate material X
Substrate mass X
Substrate cost/kg X X
Cutting tool cost X X
Number of parts per build X
Cutting tool life X X
Material Removal Rate X
Set up time X X
Loading time X X
Substrate Start-up, unloading time (h) X
Preparation Total machining time (h) X
Cost of electric energy (€/kWh) X
Cost of the cutting fluid (s/kg) X
Power demand of the machine tool in stand-by mode (kW) X
Specific energy consumption during cutting (kWh/kg) X
Mass of the material that has to be removed (kg) X
Tool change time (h) X
Cutting time (h) X
Coolant flow rate and cost/litre X
CNC power rating X
Non-cutting motion time X
Time for WAAM process (h) X
Power demand of the WAAM system in stand-by mode
X
(kW)
Time for start-up, set up and post-processing of WAAM (h) X
Specific energy consumption during deposition (kWh/ kg) X
Dwell time (h) X
WAAM
Autodesk motion and deposition tool paths X
deposition
Cooling and rolling time X
Wire mass X
Wire cost/kg X X
Inert gas cost/kg X X
Travel speed X
Wire feed Speed X X
Inert gas flow rate X X
20
This is due to the nature of the study where Priarone et al. (2019) model was proposed since
energy consumption of the WAAM production process is also one of the factors studied in
addition to process costs. This means that some of the inputs are not directly related to costs
but are related to energy consumption, which in turn is needed to calculate the cost of the
system. Some examples of these inputs are start-up, set up, workpiece clamping/unloading
time used to compute energy consumption, specific energy consumption during cutting and
mass of the material that must be removed, which serve the same purpose.
Table 2-6 Cost models comparative analysis. Cont.

Priarone
Cunningham et
Activity Input Data et al.
al. (2017)
(2019)
Heat treatment time X
Heat Treat-
ment Ramping Time X
Furnace Power Rating X
BTF Ratios X
Cutting tool life X X
Material Removal Rate X
Set up time X X
Loading time X X
Start-up, unloading time (h) X
Coolant flow rate X
CNC power rating X
Machining Non-cutting motion time X
Total machining time (h) X
Cost of electric energy (€/kWh) X
Cost of the cutting fluid (s/kg) X
Power demand of the machine tool in stand-by mode (kW) X
Specific energy consumption during cutting (kWh/kg) X
Mass of the material that has to be removed (kg) X
Tool change time (h) X
Cutting time (h) X
EDM time X
Substrate EDM wire consumption rate and cost/kg X
Removal EDM feed rate X
EDM Power Rating X
Set up X
Inspection
Inspection time X
Labour cost for each activity X X
Machine Cost for each activity X X
Labour input/ fraction of time of attendance of the machine
X X
Indirect operator for each activity
costs Machine depreciation cost X X
WAAM depreciation cost X X
Maintenance X
Production and administration overheads X
21
 Although the material removal rate only appears as input for cost in Cunningham et al.
(2017) model, it should be noted that this parameter was also used in Priarone et al. (2019)
model to calculate the specific energy consumption during cutting.
There are some similarities in the calculation of indirect costs between the two models.
However, it should be noted that different approaches were considered for the labour force
since, in the case of the Priarone et al. (2019) model, the need for labour is considered as a
fraction of the time that the worker was allocated to the machine, whereas in the case of Cun-
ningham et al. (2017) model a conservative approach is followed, where it is considered
that the WAAM machine needs constant supervision. Another difference in the calcula-
tion of the indirect cost is that production and administration overheads and maintenance
costs were considered in Priarone et al. (2019) model and were not used in Cunningham et al.
(2017) model.
The use of different WAAM processes also influences the type of inputs that are used.
This happens in the WAAM activity, where Cunningham et al. (2017) model considers cooling
and rolling time. These times are associated with the cyclically cooling and rolling of the de-
posited material, improving the final product's mechanical characteristics. In the Priarone et
al. (2019) model, this does not happen.

22
 3.
WAAM Life Cycle Cost Model Development

3.1. WAAM Life Cycle Cost Model Proposal

After understanding the advantages that WAAM technology has on different indus-
tries, from aeronautic to the automotive industry, it is necessary to understand the economic
viability of this technology. Although there are already some studies in the area, as shown in
the sub-chapter 2.4, the economic viability of the WAAM technology has been not much stud-
ied, and therefore, a cost model proposal will be made to calculate the costs involved.
For the development of the cost model, a PBCM was chosen since it allows the calcula-
tion of the costs associated with the production of WAAM parts and allows the estimation of
costs for future WAAM implementations. Considering that the WAAM technology is a tech-
nology under study, this cost model approach is the most indicated because it allows an esti-
mate of the costs associated with adopting this technology, allowing a more informed decision
making.
A LCC model was developed to study better the use of WAAM technology, whose ob-
jective is to develop a cost model that allows estimating the costs associated with the produc-
tion of WAAM parts using a cradle-to-gate approach. On the other hand, the LCC model also
aims to estimate the costs related to a given part. For this, a scenario of producing 500 hollow
cubes with approximately 7x7x7 cm was considered. It should be noted that all the processes
required to produce this part were considered.
The object of study is a new variant of HF-WAAM designed, manufactured, and tested
at the NOVA School of Sciences and Technology. The section 3.2 will provide an overview of
this new technology. The WAAM machine is shown in figure 3-1.
Figure 3-2 shows the scheme that expresses the study methodology. Firstly, a theoretical
framework of the HF-WAAM technology and the reasons that lead to its necessity will be
made. Next, the goal of the study is defined. Here the focus of the study will be defined, as
well as some limitations of the study and the approach to the life cycle model.
After the goal definition, the scope of the study will be defined. In this step, firstly, the
WAAM machine will be characterised, and then the productive system to be studied will be
presented, where the study boundaries are defined, as well as some of the limitations of the
study. The unit processes and the function unit are also defined and explained. Moreover, it
is also at this point that the processes where there is a waste of material are defined. Finally,
the cost model is developed. As already mentioned, this is where the PBCM is developed and
explained so that it can be used.

23
 Figure 3-1 WAAM machine

Figure 3-2 Life Cycle Cost Model methodology

Once the cost model is elaborated, it is necessary to collect primary data from interviews
with researchers working at NOVA SST and secondary from public websites that will be used
on the model. Table 3-1 shows the experience of the experts interviewed. At this stage, simi-
larly to the scope definition, some limitations and assumptions of the study are also presented.

24
Table 3-1 Experience of the experts interviewed
Academic Experience with Experience with Experience with
Expert
Degree AM WAAM LCC
1 PHD > 5 years > 5 years -
2 MSc 2 years 1 year 2 years
3 PHD > 5 years 2 years > 5 years
Finally, the results will be presented by applying the cost model to the described pro-
duction scenario. In this stage, there will be two approaches. Firstly, a sensitivity analysis will
be carried out to validate the model and verify which inputs cause the most variation in the
product's final cost. The other phase will be the cost discussion to identify the main produc-
tion cost drivers and understand the cost distribution in the total cost by identifying the value
of the variable and fixed costs. On the other hand, this phase also includes the identification
of total cost frontiers.
An important point that is represented in figure 3-2 is the influence that the various
stages of the LCC model have on each other. As the identification of limitations and assump-
tions may force changes in other stages of the study.

3.2. Hot Forging Wire Arc Additive Manufacturing

This section was written considering as main reference the work developed by Duarte
et al. (2020), where a new variant of WAAM, based on the addition of hot forging to the pro-
cess, was proposed.
The WAAM machine for which the cost model will be used is a technology based on hot
forging and therefore has the name hot forging wire arc additive manufacturing (HF-
WAAM). This WAAM machine uses the Metal Inert Gas (MIG) welding technique to melt the
metal that will be deposited and form the final part. This technique uses an electric arc to melt
the metal and inert gas to protect the metal from the atmosphere.
While other AM technologies have a slow deposition rate and a good surface finish,
WAAM has a high deposition rate but a low surface finish. Due to successive re-melting and
thermal solidification cycles, the parts' microstructure shows coarse columnar grain struc-
tures. Other problems with WAAM are that the mechanical strength is low and anisotropic,
which can cause premature failure of the components. Furthermore, the formation of pores
represents a potential detrimental feature. The existence of pores may cause a significant re-
duction of the mechanical properties of the parts depending on their volume fraction and
location. Another problem with pore formation is the difficulty that may exist in detecting
these defects in non-destructive testing.
The reduction or elimination of large grain structures and pores has been an area that
researchers have focused on. Several mechanisms have been developed and tested to improve
the microstructure of the produced parts as well as the mechanical properties of the compo-
nents produced by WAAM. Colegrove et al. (2014) have developed a mechanism that applies
interpass cold rolling to reduce grain size and porosity. This process, although effective, is too

25
time-consuming due to the process that takes place at a low temperature, which implies that
the time intervals between each layer deposited are high.
Hai-ou et al. (2016) presented another alternative to decrease the grain size and porosity
that consists of a hot micro-rolling tool. (Xie et al., 2016) showed a similar technique involving
a hot-rolling tool with three rolls that simultaneously deform the three sides of the layers.
Although these two approaches have quite satisfactory results, their application involves pro-
ducing and using different AM equipment. It prevents using a normal welding source and a
robot or XYZ table, making the process less economically competitive. Other possibilities that
have been studied are the utilization of laser shock peening and ultrasonic peening, but alt-
hough they can reduce residual stress and refine microstructures, rolling methods are more
efficient.
The HF-WAAM variant was developed to decrease or eliminate coarse columnar grain
structures and porosity without increasing the time between depositions. For this purpose, a
new variant was developed, consisting of a multi-feed device where the different inputs are
separate and can be controlled independently. These inputs are electric power, shielding gas,
material feed, and mechanical work. The latter uses a hammer activated by a vibrating actua-
tor placed inside the shaping gas nozzle. The existence of this hammer allows viscoplastic
deformation at high temperatures to reduce residual stress, eliminate post-heat treatments,
homogenize the grain structure, and increase ductility. Using a technology where the inputs
are independent and where the viscoplastic deformation is done at high temperature allows
the forging load to be reduced, the time between depositions does not increase, and there is
greater agility in tooling and equipment utilization.
During the manufacturing process, the hammer travels along the same path as the torch
and is in operation at the same time as the material is being deposited, causing the hammer
to deform the part at temperatures well above the recrystallisation temperature. Due to the
high temperature at which the process is carried out, it is possible to induce dynamic recrys-
tallization of the deposited material as well as cause the pores to collapse with a much lower
load than would be necessary if deformation were carried out at room temperature.
The recrystallisation of the grains in the WAAM processes has some advantages. Be-
sides increasing the mechanical strength, it decreases the susceptibility to large grains due to
the high density of nucleation sites on top of the deposited layer due to the existence of refined
grains. Finally, the deformation of the deposited layers also reduces material waste in subse-
quent machining processes, as deformation at high temperatures reduces waviness and sur-
face roughness, thus increasing precision.
One of the findings of the results obtained from the use of HF-WAAM is that each layer
has a decrease in height and an increase in width, which becomes more pronounced as the
hot forging increases, this occurs due to volume conservation and is accentuated in the case
where a rectangular hammer is used instead of a cylindrical hammer as the contact area is
smaller. This is because reducing the contact area increases stress and increases deformation.
Another positive result is that when the hammer is in operation, it has a much lower
temperature than the melted material, which means that the characteristics of the material
from which it is made do not change. Another positive point is the fact that there is no con-
tamination of the deposited material by the hammer.
26
 Regarding the hot forging process's effect on the grain structure, the results show that
plastic deformation at high temperatures causes a grain refinement, increasing the produced
component's strength. On the other hand, it was also visible in the results that using this
WAAM variant allows the reduction of the number of pores, which is advantageous for the
mechanical properties of the produced component.
The mechanical properties of products produced by the HF-WAAM variant were stud-
ied and compared with those of products produced by WAAM, and promising results were
observed. In the experiments performed, an average increase in ultimate tensile strength of
8.9% and an increase in yield strength from 360 to 450 were observed. On the other hand, a
reduction of 13.9 % in elongation was observed. This is because the grain size's reduction has
on the structure of the component since with the decrease of the grain size, the number of
boundaries between grains increases, which causes an increase in the difficulty of displace-
ment. Naturally, the result of this effect is a slight reduction in elongation. On the other hand,
as previously mentioned, reducing the grain size increases the strength of the produced com-
ponent. However, it is important to note that reduced ductility is not a problem that precludes
the use of AISI316L steel in most industries that require ductility.
The application of the HF-WAAM variant presents some advantages for production in
addition to the advantages described above, which concern the characteristics of the materials
produced. These advantages concern the efficiency of production and its costs. Since defor-
mation occurs at high temperatures, three advantages can be highlighted. Firstly, the fact de-
scribed above allows the forces exerted by the hammer to be much lower than those required
for an identical process at low temperatures, thus making energy consumption almost negli-
gible. On the other hand, this variant allows the use of traditional motion equipment, namely
XYZ tables and 6-axis robots, this fact is crucial because it allows the system to be more flexible
and allows the new variant not to be a constraint for the chosen equipment. Finally, this vari-
ant does not imply an increase in time between each deposition, allowing the system to main-
tain the same productivity as a conventional system without a hammer.
About the degree of development of the HF-WAAM variant, it is between TRL 4 and
TRL 5, according to the technology readiness levels scale. This happens because the technol-
ogy in question has been validated in the laboratory but has not yet been validated in a rele-
vant environment.

3.3. Goal Definition

Over the last decades, the AM industry has been developed and studied because of the
way it has revolutionised the production of different types of products and materials. One of
these techniques is WAAM, a recent and disruptive technology that theoretically allows me-
tallic printing structures that can have any dimension. However, this technology still presents
some limitations, such as the lack of precision, the need for a substrate and the need for post-
processing machining. Besides these limitations, WAAM is still a production technique not
well studied, which leads to the need of this study, which aims to develop a cost model that
allows knowing the costs involved in the production of WAAM products.
27
 This study aims to develop a cost model that allows tracking the costs involved in the
life cycle of WAAM products and study its economic viability. This LCC object of study is a
WAAM machine designed, produced, and tested at the NOVA School of Science and Tech-
nology. The material to be studied will be stainless steel AISI316LSI. This study will be ori-
ented to the production of 500 hollow cubs with 7x7x7 cm. The production scenario to be
studied is a hybrid WAAM process.
In this study, the life cycle approach will be the cradle-to-gate since the study's main
objective is to develop a cost model that enables the estimation of the costs associated with
the manufacture of the parts.

3.4. Scope Definition

3.4.1. Characterization of the WAAM Machine

The following subchapter was based on the work developed by Duarte et al. (2020),
where a new variant of WAAM, based on the addition of hot forging to the process, was pro-
posed.
The developed WAAM machine, like other WAAM machines, has enormous potential
in some industrial sectors. Although the HF-WAAM has not been developed for a specific
industry like other WAAM machines, its application is of particular interest for heavy indus-
tries such as aeronautics, shipbuilding, automotive and mould making. One of the main rea-
sons for the interest in this technology is the fact that these industries use materials whose
acquisition cost is very high, such as the use of titanium in the aeronautical industry. The fact
that WAAM technology is based on the deposition of material to create a structure generates
an enormous potential since it is no longer necessary to use a large amount of material to
achieve the desired shape of the component through subtractive methods, as is the case of
traditional processes. This causes a reduction in the waste of raw materials.
Another great advantage that WAAM technology has for this type of industry is creat-
ing large parts since, theoretically, the part produced can be of any size. In the case of the
studied machine, the torch was intruded into a three-axis position system with a working
envelope of 2760x1960x2000 mm. This means that the system to be studied will be able to
produce parts up to these dimensions, although, as mentioned before, the torch can be placed
in another type of structure.
Another important point about the type of industry where this technology can be ap-
plied, and the type of product produced is the poor surface finish of the deposited material.
This causes a lower precision of the WAAM technology in relation to other types of additive
technologies. This limitation implies that these technologies can only produce components
with geometries of medium complexity.

28
 The WAAM machine to be analysed is equipped with a PRO MIG 3200 power source
from KEMPY that deposits 1 mm diameter AISI316L stainless steel on the substrate, it also
has connections to the tool steel hammer by an M6 leadscrew, two pneumatic cylinders as
vibrating actuators, namely a FESTO ADVC-6-5-A-P that produces a force of 17 N and 55 N
at a pressure of 0.6 MPa and a SMCCU10-10T that causes a force of 55 N at 0.6 MPa. A 5/2-
way bistable solenoid valve Festo VUVS-LK20 was used to actuate the pneumatic cylinders,
which was controlled by a data acquisition device. In figure 3-3, the torch and its components
are illustrated.

Figure 3-3 HF-WAAM illustration

3.4.2. WAAM Process

This sub-chapter will explain the WAAM process, from the material reception to the
storage of the produced parts. For this purpose, a diagram was elaborated to illustrate the
process and the operations involved in this process. The study's boundaries included the unit
processes: substrate cutting and oxide removal, WAAM process, finish machining and sub-
strate removal, heat treatment and inspection. The operations where material waste can be
found were pointed out in the diagram. The diagram of the WAAM process is presented in
figure 3-4. The process showed below was developed after interviews with experts working
in NOVA SST.
The first step is the raw material reception and storage, which are the wire and the sub-
strate. When the production starts, the set-up of the first operation is set, and next to the sub-
strate is cut with a bandsaw, and then the oxides are removed from the substrate using a steel
brush, at last, the substrate is washed with alcohol to guarantee there are not fats or other
29
contaminates or substances that could disturb the production. After the substrate preparation,
the material is checked, and if the substrate is in good condition, it goes to the WAAM process,
if it is not, the substrate returns to the previous operation.
When the substrate arrives to the WAAM process, the setup of the WAAM process is
done. The setup embraces preparing the torch, the wire, the shielding gas and placing the
substrate. Then the deposition starts. During this step, there is material waste, but it is ne-
glectable due to its small amount, and the welding nozzles used in the WAAM process will
be considered consumable/tools in the cost model. After the WAAM process, the part is in-
spected, and if it satisfies the requirements, it will be forwarded to the next process, but if it
has a fault, the part is scrapped due to the difficulty in reworking it.

Figure 3-4 Diagram of the WAAM process

In the finish machining operation, a CNC machining is used to refine the part's surface.
This stage consists of setup and machining. During this stage, consumables/tools are used,
such as the machine's milling cutters. A significant amount of material is removed, considered
waste material in the diagram. Then the part is inspected to know if the part needs to be re-
worked in the previous step or if it can go to the next operation.
After machining, it is necessary to remove the part from the substrate unless the sub-
strate is a massive part of the final product. This will depend on the kind of structure and
application of the final product. As has been mentioned before, the WAAM process is not
efficient in producing massive structures, so if the product has a massive part, it is a good
practice to manufacture it by traditional methods. In the diagram, the substrate removal is
considered a different operation from finish machining, although, in this study, it is consid-
ered that this stage is integrated into the finishing machining as the same unit process to sim-
plify the cost model since the machine used is the same. At last, the diagram shows material
waste at the substrate removal corresponding to the substrate since it cannot be reused due to
the surface's roughness and the difficulty of reworking it.

30
 After the substrate removal, the component needs a heat treatment to normalize the
structure. So, when the part arrives to this stage, an operator puts it in the oven. After the heat
treatment has concluded, the component is sent to the last inspection. The inspection will de-
pend on the kind of application since that will condition the requirements in this stage. If there
are special requirements related to the piece's structure, the inspection must be done with x-
rays or ultrasounds, but a visual inspection will be enough in case of no structural require-
ments. Except for the last inspection, the previous inspections will be part of the previous unit
processes for simplifications. The functional unit considered is one hollow cube with approx-
imately 7x7x7 cm.

3.4.3. Cost Model

When developing a cost model, the first step is to identify the most significant costs
associated with producing a given part while maintaining the model's reliability and credibil-
ity. The time horizon used for the calculation of costs will be one year.
The production costs to be calculated can be separated into two types of costs: variable
costs and fixed costs. Variable costs are costs dependent on the volume of production, which
means that they vary in absolute terms but are constant in terms of unit cost. Fixed costs do
not vary with production, which means that if there are no changes in the system, costs do
not change. However, fixed costs become more diluted in unit costs as the volume increases.
In table 3-2 are presented the division between variable and fixed costs. This division allows
a better understanding and analysis of the results.
Table 3-2 Fixed and variable costs

Fixed Costs Variable Costs

Machine Materials
Overheads Consumable
Maintenance Energy
Labour
Tools

It is important to define the concepts associated with time and utilisation in a manufac-
turing context. For such, time can be divided into available time and unavailable time, where
the first comprises uptime and idle time. The uptime is the time in which the workstation is
working, while the idle time is all the time in which the workstation could be working but has
no work to be carried. Unavailable time is associated with all types of unforeseen breaks, such
as scheduled breaks, maintenance, and the time when the factory is closed. Finally, downtime
is the idle time added to the unavailable time. In relation to factory costs, these must be asso-
ciated with the time when the stations are producing parts, so these expenses are allocated to

31
uptime. Another critical concept is dedication. Since a machine can be dedicated or non-ded-
icated. When a machine is dedicated, it means that the uptime is equal to the time that the
machine is being used to produce the part being studied.
When developing a PBCM, it is necessary to identify each process's inputs and outputs
since the cost model is based on the division of the production cycle into several processes,
and their inputs and outputs are used to calculate the production costs.
The inputs of each process can be divided into two categories: general inputs and spe-
cific process inputs. General inputs coincide with all processes and are influenced by the com-
pany or laboratory where the products are produced, such as wages or working days per year,
while specific process Inputs are related to the process in question. This means that these in-
puts are relative to a process and can vary. Therefore, some processes can have inputs that are
not coincident with other processes. Table 3-3 and table 3-4 present the specific process inputs
and general inputs, respectively.
The inputs for table 3-3 table 3-4 are explained below:
• The dedication input informs if the machine only produces the product to be
studied or if it also produces others. And the allocation represents the percentage
of time spent on that task.
• Then follows the number of workers assigned to the task and the percentage of
their working time that they spend on the operation, which corresponds to the
dedication.
• Acquisition cost is the cost of the machine.
• Set up and machine time are needed to prepare the machine to start operation
and the time the machine is running.
• Travel speed and standby time are the length of material deposited per hour
and the time between depositions, respectively.
• Maintenance and overheads are expressed as a percentage of the machine cost.
• At last, it is necessary to consider the quantity of consumables required for the
production and their costs and the number of tools and their costs.
The inputs for table 3-4 are explained below:
• Days per year corresponds to the days the factory operates per year.
• The wage Is the labour cost per hour.
• The unit energy cost Is the cost of the energy.
• The machine life is when the machine is available, and the production life is
when the part will be produced.
• Opportunity cost rate is the interest at which the investment is annualized.

Other inputs are essential to develop the cost model: part information, material infor-
mation, scrap input, and line utilization. The inputs used are the number of layers and the
layers' length within part information. Material information is the kind of material and cost.
In relation to scrap input, it is necessary to know the amount of material that is approved, the
quantity of the non-approved material and the amount of material reworked, and the cost at
which the waste material may be sold, as well as the material removal rate and the volume
32
that is removed from the part at the time of finishing. Regarding line utilization, it is necessary
to consider the scheduled and non-scheduled breaks of the operators, unplanned breakouts,
non-working times, and maintenance stoppages. Equation 6 shows how it is computed the
available time.
Table 3-3 Specific process inputs

Specific Process Inputs Unit

Dedication Yes/No
Allocation %
Nº of workers Number
Worker’s dedication %
Acquisition cost €
Set uptime h
Machine time h
Standby time h
Travel speed m/h
Maintenance %
Overheads %
Power consumption kW
Consumable required unit
Consumable cost €/unit
Tools number Number
Tools Cost €

Table 3-4 General inputs

General Inputs Unit

Days per year Days/year
Wage €/h
Unit Energy Cost €/kWh
Equipment Life years
Production life h
Opportunity cost rate %

𝐴𝑣𝑎𝑙𝑖𝑎𝑏𝑙𝑒 𝑡𝑖𝑚𝑒 = 24ℎ − (𝑠𝑐ℎ𝑒𝑑𝑢𝑙𝑒𝑑 𝑎𝑛𝑑 𝑛𝑜𝑛 𝑠𝑐ℎ𝑒𝑑𝑢𝑙𝑒𝑑 𝑏𝑟𝑒𝑎𝑘𝑠 + 𝑛𝑜𝑛 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑡𝑖𝑚𝑒 + (6)
𝑡𝑖𝑚𝑒 𝑓𝑜𝑟 𝑚𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒 𝑠𝑡𝑜𝑝𝑝𝑎𝑔𝑒𝑠)

To compute the product production cost, it is necessary to know how much material is
needed to produce the part. For this purpose, it is mandatory to consider the number of parts
that have quality and are passed to the next process, the number of defective parts, and the
number of parts that can be reworked. Equation 7 shows how to calculate the number of parts
to satisfy the requirements. This calculation is performed from the end of the production line
to the beginning. It is also important to mention that the only material that can be reworked
is the substrates since the printed part, if defective, is considered waste material.

33
 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑠𝑖+1
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑠𝑖 = (7)
%𝑝𝑎𝑠𝑠𝑒𝑠𝑖 + %𝑑𝑒𝑓𝑒𝑐𝑡𝑖𝑣𝑒𝑖 × %𝑟𝑒𝑤𝑜𝑟𝑘𝑖

To calculate fixed costs, it is necessary to calculate the uptime computed by equation 8,

but before it is needed to perform some intermediate calculations to compute the required
time through equation 9 and the cycle time. Cycle time depends on the process that is being
carried out. Equation 10 presents the calculation for cycle time for WAAM and equation 12
for calculating cycle time for finishing machining.

𝑇𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑
𝑈𝑝𝑡𝑖𝑚𝑒 = (8)
𝐴𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛

𝑇𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 = 𝐶𝑦𝑐𝑙𝑒 𝑡𝑖𝑚𝑒 × 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑠 (9)

𝑛
𝐿𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑙𝑎𝑦𝑒𝑟𝑠𝑖
𝐶𝑦𝑐𝑙𝑒 𝑡𝑖𝑚𝑒𝑊𝐴𝐴𝑀 = 𝑆𝑒𝑡𝑢𝑝 𝑡𝑖𝑚𝑒 + ∑ ( ) + 𝑆𝑡𝑎𝑛𝑑𝑏𝑦 𝑡𝑖𝑚𝑒 × (𝑁𝑙𝑎𝑦𝑒𝑟𝑠 − 1) (10
𝑇𝑟𝑎𝑣𝑒𝑙 𝑠𝑝𝑒𝑒𝑑
𝑖=1
)

If the length of the layer is the same for all the layers at the part to be analysed, then:

𝑙𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑙𝑎𝑦𝑒𝑟𝑠
𝐶𝑦𝑐𝑙𝑒 𝑡𝑖𝑚𝑒𝑊𝐴𝐴𝑀 = 𝑆𝑒𝑡𝑢𝑝 𝑡𝑖𝑚𝑒 + ( ) × 𝑁𝑙𝑎𝑦𝑒𝑟𝑠 + 𝑆𝑡𝑎𝑛𝑑𝑏𝑦 𝑡𝑖𝑚𝑒 × (𝑁𝑙𝑎𝑦𝑒𝑟𝑠 − 1) (11)
𝑇𝑟𝑎𝑣𝑒𝑙 𝑠𝑝𝑒𝑒𝑑

𝑉𝑜𝑙𝑢𝑚𝑒 𝑡𝑜 𝑏𝑒 𝑟𝑒𝑒𝑚𝑜𝑣𝑒𝑑 (12)

𝐶𝑦𝑐𝑙𝑒 𝑡𝑖𝑚𝑒𝐹𝑖𝑛𝑖𝑠ℎ 𝑚𝑎𝑐ℎ𝑖𝑛𝑖𝑛𝑔 =
𝑀𝑅𝑅

Knowing the uptime, it is possible to calculate downtime and idle time, computed by
equation 13 and equation 14, respectively.

24ℎ × 𝐷𝑎𝑦𝑠 𝑝𝑒𝑟 𝑦𝑒𝑎𝑟 − 𝑈𝑝𝑡𝑖𝑚𝑒

𝑇𝑜𝑡𝑎𝑙 𝑑𝑜𝑤𝑛𝑡𝑖𝑚𝑒𝑖 = (13)
𝐷𝑎𝑦𝑠 𝑝𝑒𝑟 𝑦𝑒𝑎𝑟

𝐼𝑑𝑙𝑒 𝑡𝑖𝑚𝑒𝑖 = 𝐴𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒 𝑡𝑖𝑚𝑒 − 𝑢𝑝𝑡𝑖𝑚𝑒 (14)

Henceforth, equations for determining the costs associated with the production of
WAAM products will be presented, starting with variable costs. The first cost to be considered
is the materials costs. For wire, it is necessary to consider the wire cost, the weight of the raw
material, and consider that the wasted material can be sold. Its value is taken from the total
cost of the raw material cost. The raw material cost is possible to compute with equation 15.
Equation 17 computes the shielding gas cost by multiplying the volume of shielding gas re-
quired by the cost per litre of shielding gas. Equation 18 calculates the substrate cost, and, as
in the wire cost equation, the waste material Is considered to calculate the total substrate cost.
It is important to note that shielding gas and substrate are consumables, not raw materials.

34
 𝑊𝑖𝑟𝑒 𝑐𝑜𝑠𝑡 = 𝐶𝑜𝑠𝑡𝑊𝑖𝑟𝑒 × 𝑊𝑖𝑟𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 − 𝐶𝑜𝑠𝑡𝑀𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑤𝑎𝑠𝑡𝑒 × 𝑀𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑤𝑎𝑠𝑡𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 (15)
− 𝐶𝑜𝑠𝑡𝑆𝑐𝑟𝑎𝑝 × 𝑆𝑐𝑟𝑎𝑝 𝑤𝑒𝑖𝑔ℎ𝑡

Scrap weight is computed by equation 16. This equation can be used for WAAM scrap
and substrate scrap.

𝑆𝑐𝑟𝑎𝑝 𝑤𝑒𝑖𝑔ℎ𝑡 = 𝑀𝑎𝑡𝑒𝑟𝑖𝑎𝑙 𝑢𝑠𝑒𝑑 𝑤𝑒𝑖𝑔ℎ𝑡 × (1 − %𝑝𝑎𝑠𝑠𝑒𝑠 − %𝑑𝑒𝑓𝑒𝑐𝑡𝑖𝑣𝑒𝑖 × %𝑟𝑒𝑤𝑜𝑟𝑘) (16)

𝑆ℎ𝑖𝑙𝑑𝑖𝑛𝑔 𝑔𝑎𝑠 𝑐𝑜𝑠𝑡 = 𝑉𝑜𝑙𝑢𝑚𝑒 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑𝑆ℎ𝑖𝑒𝑙𝑑𝑖𝑛𝑔 𝑔𝑎𝑠 × 𝐶𝑜𝑠𝑡𝑆ℎ𝑖𝑒𝑙𝑑𝑖𝑛𝑔 𝑔𝑎𝑠 (17)

𝑆𝑢𝑏𝑠𝑡𝑟𝑎𝑡𝑒 𝑐𝑜𝑠𝑡 = 𝐶𝑜𝑠𝑡𝑆𝑢𝑏𝑠𝑡𝑟𝑎𝑡𝑒 × 𝑆𝑢𝑏𝑠𝑡𝑟𝑎𝑡𝑒 𝑎𝑟𝑒𝑎 (18)

− 𝐶𝑜𝑠𝑡𝑆𝑢𝑏𝑠𝑡𝑟𝑎𝑡𝑒 𝑢𝑠𝑒𝑑 × 𝑆𝑢𝑏𝑠𝑡𝑟𝑎𝑡𝑒 𝑢𝑠𝑒𝑑 𝑤𝑒𝑖𝑔ℎ𝑡 − 𝐶𝑜𝑠𝑡𝑆𝑐𝑟𝑎𝑝 × 𝑆𝑐𝑟𝑎𝑝 𝑤𝑒𝑖𝑔ℎ𝑡

As with shielding gas, the cost of other consumables is also computed by the number of
tools used per hour of the process. The equations for calculating the cost of shielding gas and
substrate are not presented as the same equation because the way to achieve the substrate cost
is different from the shielding gas cost. The tools costs are given by equation 19, where the
number of tools needed and their cost are considered.

𝑇𝑜𝑜𝑙 𝑐𝑜𝑠𝑡 = 𝐶𝑜𝑠𝑡𝑇𝑜𝑜𝑙 × 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑜𝑜𝑙𝑠 (19)

The labour cost is presented in equation 20. Only the directly involved workers in the
process are considered for the labour cost, as the others are considered in the overheads. On
the other hand, since workers do not need to constantly work on a given process, the opera-
tor's dedication is also considered. Labour costs also include the number of workers, the
hourly wage and the time required.

𝐿𝑎𝑏𝑜𝑢𝑟 𝑐𝑜𝑠𝑡 = 𝐷𝑒𝑑𝑖𝑐𝑎𝑡𝑖𝑜𝑛 × 𝑊𝑎𝑔𝑒 × 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑤𝑜𝑟𝑘𝑒𝑟𝑠 × 𝑇𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 (20)

(18)
)
For the energy cost, two types of formulas can be used, one for all processes and the
other for the WAAM process, although for calculation purposes, the general equation for all
processes can be used for the WAAM process since the difference between the costs obtained
is negligible compared to other costs associated with the process studied. In the case of the
general equation for calculating the energy cost, it is necessary to consider the time required,
the process energy consumption and the unit energy cost. For the more accurate calculation
of the WAAM machine consumption, the methodology is similar, the only difference is that
the energy consumption during deposition and during standby time is detached, which are
then multiplied by the respective times and then summed. Finally, they are multiplied by the
number of pieces and the unit energy cost. The equations for the energy cost for the different
processes and the WAAM machine are the equations 21 and 22, respectively.

𝐸𝑛𝑒𝑟𝑔𝑦 𝑐𝑜𝑠𝑡 = 𝑇𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 × 𝑃𝑜𝑤𝑒𝑟 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛 × 𝑈𝑛𝑖𝑡 𝑐𝑜𝑠𝑡𝐸𝑛𝑒𝑟𝑔𝑦 (21)

35 (18)
)
 Or
𝐸𝑛𝑒𝑟𝑔𝑦 𝑐𝑜𝑠𝑡𝑊𝐴𝐴𝑀
𝑛
(22)
𝐿𝑒𝑛𝑔𝑡ℎ 𝑜𝑓 𝑙𝑎𝑦𝑒𝑟𝑠𝑖
= (∑ ( ) × 𝑃𝑜𝑤𝑒𝑟 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛𝐷𝑒𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛 (18)
𝑇𝑟𝑎𝑣𝑒𝑙 𝑠𝑝𝑒𝑒𝑑 )
𝑖=1

+ 𝑆𝑡𝑎𝑛𝑑𝑏𝑦 𝑡𝑖𝑚𝑒 × (𝑁𝑙𝑎𝑦𝑒𝑟𝑠 − 1) × 𝑃𝑜𝑤𝑒𝑟 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛𝑆𝑡𝑎𝑛𝑑𝑏𝑦 )

× 𝑈𝑛𝑖𝑡 𝑐𝑜𝑠𝑡𝐸𝑛𝑒𝑟𝑔𝑦 × 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑠

Next, the equations are presented to obtain the value of fixed costs. For this purpose, a
logic is adopted in which fixed costs are annualised so that the investment becomes a set of
payments over the useful life of the machine (n) with an opportunity cost of capital (r). Equa-
tion 23 represents the way to calculate annualised fixed costs.
(1 + 𝑟)𝑛𝑗 × 𝑟
𝐹𝑖𝑥𝑒𝑑 𝑐𝑜𝑠𝑡𝑗𝑖 = 𝐼𝑖 × (23)
(1 + 𝑟)𝑛𝑗 − 1
(18)
)
The machine investment is calculated considering the cost of the machine, the quantity,
and allocation. This is because the machine can be allocated to more than one workpiece,
which means that the machine investment must be divided by more than one product. Equa-
tions 24 calculate machine investment.

𝑀𝑎𝑐ℎ𝑖𝑛𝑒 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 = 𝐶𝑜𝑠𝑡𝑀𝑎𝑐ℎ𝑖𝑛𝑒 × 𝑎𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛 × 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑚𝑎𝑐ℎ𝑖𝑛𝑒𝑠 (24)

(18)
)
Maintenance investment and overhead investment are calculated as a percentage of
fixed costs. In the case of maintenance investment, it is calculated as a percentage of machine
investment since it is the only kind of equipment that requires maintenance. The overhead
investment is related to all existing costs not directly associated with production and is calcu-
lated as a percentage of all fixed costs. Equations 25 and 26 compute maintenance investment
and overheads investment, respectively.

𝑀𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 = 𝑀𝑎𝑐ℎ𝑖𝑛𝑒 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 × 𝑀𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒 (25)

(18)
)
𝑂𝑣𝑒𝑟ℎ𝑒𝑎𝑑𝑠 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡
= (𝑀𝑎𝑐ℎ𝑖𝑛𝑒 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 + 𝑇𝑜𝑜𝑙 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 + 𝑀𝑎𝑐ℎ𝑖𝑛𝑒 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡) (26)
× 𝑀𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒 (18)

)
3.5. Life Cycle Inventory

Multiple sources of data were used to collect and compile data regarding each input
required to produce WAAM products. The data collection process included the compilation
of secondary data available in public websites such as EDP (2022), Alibaba.com (2022) and

36
Haas CNC Machines (2022). Also, primary data was collected through unstructured inter-
views with researchers who developed and worked with the WAAM equipment, since they
have access to experimental data and also have a deep knowledge about the equipment func-
tion. The data was collected in the period of time between September 2021 and February 2022.
For the application of the cost model, the production of 500 hollow cubes with 7x7x7 cm
was considered. The wall thickness was defined as 5.5 mm for the side and top faces, and for
the bottom face, a thickness of 1.3 mm was considered due to deposition parameters. There-
fore, for the part information, the number of layers and their dimensions for the lower face,
the four lateral faces and the upper face were obtained. For the base, 12 layers with 7 cm in
length each, 49 layers with 257 mm in length for the side faces and 54 layers with 7 cm each
for the top surface. These inputs were calculated considering that the material is deposited
initially to create the bottom surface, then the side surfaces are deposited, wherein each layer
the material is deposited over the edges of the base and finally, the top surface is deposited
with the mouth in a lateral position, thus making the thickness of the wall similar to the thick-
ness of the side walls.
Regarding the information on materials, three materials required for production were
identified: stainless steel plates, shielding gas composed of Argon, and the wire of stainless
steel AISI316LSI. The last one is raw material, and the other two consumables are used during
the process. The costs of the materials were provided by experts working in NOVA SST and
can be seen in table 3-5.
Table 3-5 Material information

Material Wire Shielding gas Substrate

material required 1.01 kg 717.84 l 0.015m2
Cost 5.33 €/kg 0.03 €/l 257.07 €/m2
Note: Data obtained by interviews with experts working in NOVA SST.

Although scrap was considered for the cost model development, it was not considered
in the estimate made since the value to be received for it is quite neglectable compared with
the other production costs. Furthermore, it was considered that all the material passed the
inspections, so there is no rework. This is due to a lack of information on passes and reworks.
However, the volume value removed from the part and the material removal rate was
calculated concerning scrap information. Once again, the cost of waste material was not con-
sidered, and these values were used to calculate the machining time of the finish machining
process and substrate removal. The material removal rate is 262.5 cm3 /min, considering the
depth of cut 0.5 mm, the width of cut 35 mm and the feed rate 250 mm/s, these values were
obtained in the laboratory of NOVA SST. For the calculation of the removed volume, it was
considered that in each exterior wall of the cube, excluding the inferior surface, 0.5 mm was
removed and that in the inferior face, only the substrate was removed. Given the dimensions
of the cube, the volume removed from the walls was 12.25 cm3 . For the substrate part, it was
computed that 15 cm3 were removed, considering that the substrate measures are 15 cm in
length, 10 cm in width and 1 cm in height. The buy to fly ratio of the deposited part under
these conditions is 1.1, which agrees with the literature (Priarone et al., 2019). Researchers

37
from the area provided the data mentioned above through unstructured interviews. Regard-
ing the line utilization, the values presented in table 3-6 were considered for the duration of
each break.
Table 3-7 shows the values corresponding to the general inputs. To obtain the days per
year were considered the working days and stoppages due to periods in which production
may be stopped, as is the case of holidays. The wage was calculated based on a gross wage of
1500 €, a unique social tax of 23.75% and a holiday and Christmas bonus. For the other data,
the typical values of the industry were considered. The unit energy cost was obtained from
EDP (2022).
Table 3-6 Line utilization inputs

Non-working Time 14.3 h/day

Scheduled Breaks 1.0 h/day
Unscheduled Breaks 0.5 h/day
On Shift Maintenance 0.1 h/day
Unplanned Downtime 0.1 h/day
Note: Data obtained by interviews with experts working in NOVA SST.

Table 3-7 General inputs data

Days per Year 230 days/year

Wage 14.12 €/hour
Unit Energy Cost 0.169 €/kWh
Interest 10 %
Equipment Life 10 years
Production Life 10 years
Note: Data obtained in public websites.

In table 3-8 are the specific process inputs. For the cost model, it was considered that
none of the machines are dedicated, although later, we will study the scenarios in which this
happens. The acquisition costs of the heat treatment and CNC milling were obtained through
Alibaba.com (2022) and HAAS CNC MACHINES (2022). The setup times were obtained from
unstructured interviews with researchers, and the machining times were obtained in the same
way, except for the WAAM and finish machining and substrate removal processes since the
latter was estimated as previously mentioned. Since, for the WAAM machine, the machine
time is computed, the travel speed was obtained through unstructured interviews with re-
searchers. In the case of finish machining and substrate removal, the setup time is 30 minutes,
but it was considered that there are two setups since the workpiece must be machined from
above and below. In the case of power consumption, a conservative approach was used, so
the values used correspond to the maximum power of the WAAM machine. In the case of the
WAAM machine, the power consumption considered was the highest of the range of values
obtained in the laboratory since only the deposition time was measured, which may represent
an almost insignificant impact on the overall study. The standby time was considered as 1
minute for the bottom and top faces and 30 seconds for the construction of the side faces. The
maintenance and overhead percentages were obtained through unstructured interviews with
38
researchers. The tools considered are the forging nozzles for the WAAM process and cutters
for finishing machining and substrate removal. The cost of the tools was collected through
Table 3-8 Specific process input data unstructured interviews with researchers.
Finish
Substrate
machining
cutting Heat
Reception WAAM and Inspection
and Oxide treatment
Substrate
removal
removal
Dedication - No No No No No
Allocation - 5 90 40 40 20 %
units/
1 1 1 1 1 1
Number of workers shift
Worker’s dedication - 50 20 25 5% 30% %
Acquisition cost - 2500 300000 60000 5 500 2 350 €
Power consumption - 1.2 1.89 11.2 15 - kW
Setup time 0.25 0.05 1 1 0.008 0.08 h
Machine time - 0.106 - 0.15 1 0.5 h
Standby time - - 0.017 - - - h
Travel speed - - 21.6 - - - m/h
Maintenance - 10 10 10 10% 10% %
Overheads - 40 40 40 40% 40% %
1.007 kg
Consumable - - 717.840 - - l
- 0.015 2
m
5.333 €/kg
-
Consumable cost - - 0.025 - - €/l
257.066 - €/m2
Tools number - - 4 301 - - unit
Tools cost - - 8 100 - - €/unit
Note: Data obtained by interviews with experts working in NOVA SST.

3.6. Results

3.6.1. Model Validation and Sensitivity Analysis

Once all the cost model inputs have been clarified, it is necessary to perform a model
validation. Therefore, the model's inputs were varied to check whether the total cost varied
as expected. This validation, carried out using a sensitivity analysis, allows to identify each
input's variation in the final cost and find out if it is significant.
The sensitivity analysis was performed by varying the inputs of each process, keeping
the other inputs the same, to determine the variation of the total cost per piece. To do so, the
inputs were varied between -30% and 30% and in the case of percentage variables, a variation

39
was made between 10% and 100%. It should be noted that inputs directly impact costs line-
arly, making the graphical representations show a linear trend.
The following charts will show the variation of the total cost as a function of the varia-
tion of the inputs of each process, as well as the part information and material information.
The variation of the tools variables was made separately from the process.
For the analysis of the variation of inputs of part information, material information, and
the processes except for the WAAM process and finish machining and substrate removal, a
maximum of 290 € and a minimum of 250 € were considered since the largest variation that
occurs is 2.1 %. This way, it is possible to make a better comparison between inputs. In the
case of the WAAM process, finish machining and tools, it was considered a maximum of 320
€ and a minimum of 200 € since the largest variation that occurs in this process is about 15.6
%. In the case of percentage inputs, a similar approach was considered, but in this case, con-
sidering a different range for the WAAM process. Therefore, a maximum of 290 € and a min-
imum of 250 € has been considered for all the processes and a maximum of 400 € and a mini-
mum of 100 € for the WAAM process.
Figure 3-5 shows the part information sensitivity analysis. For this, a cost of 259.95 €
per part produced was calculated. For the sensitivity analysis, it was considered that at each
variation, the layers with different lengths vary proportionally, the same approach was taken
for the number of layers since depending on the place where the material is being deposited,
the standby time also varies. In figure 3-5, it is possible to observe that the number of layers
has more effect on the total cost than the length of the layers. Since the number of layers, when
varied to the maximum, varies the total cost by 1.17%, while the variation caused by varying
the layer's length is 0.8 %. This happens because the amount of raw material and the deposi-
tion time vary significantly when varying the number of layers than layer length.
Figure 3-6 represents the sensitivity analysis of the material information for the three
materials. The amount of material and its cost vary almost equally the part's total cost. How-
ever, there is a significant difference between the materials since both the wire and the sub-
strate cause a variation of less than 1% each, while the shielding gas causes a variation of 2.04
% while the required material varies by 30%.

Figure 3-5 Sensitive analysis of the part information

40
 Figure 3- 6 Material information sensitive analysis

In appendix 2 are presented the figures A2-1 and A2-2, which represent the sensitivity
analysis of the absolute and percentage inputs of the substrate cutting and oxide removal pro-
cess, respectively. Due to the short cycle time, low allocation, and low machine cost, it is pos-
sible to observe that the effect inputs have on total part cost is very low, less than 0.2% for
absolute inputs.
Figure 3-7 shows the sensitivity analysis of the inputs to the WAAM process. As de-
scribed previously, this process has a different cost range as the variable acquisition cost has
a huge effect on the part's total cost, varying it by 15.62 %. This is due to the high cost of the
WAAM machine and the allocation being very high for this process, 90%. The other input that
most influences the total cost is travel speed, with a variation of 1.82 %. This variation is due
to changing the deposition time that, besides changing the machine's cycle time, also causes
the change in the amount of gas consumed, which, as seen previously, is the raw material
with the greatest effect on the final cost. As expected, the slope of the travel speed is negative
since increasing it decreases gas consumption and deposition time.

Travel speed

Figure 3-7 WAAM inputs sensitive analysis

41
 In its turn, figure 3-8 represents the sensitivity analysis of the WAAM percentage inputs.
As can be seen from the previous analysis and as will be verified later, the WAAM process is
the one that most influences the total cost since it represents the largest part of it. Therefore, it
is observable that the percentage inputs also vary the product's final cost significantly. Alt-
hough the value of inputs is 10%, they are in quite different positions. It is notable that
throughout the variation between 10% and 100%, the input that has the highest impact on cost
is allocation due to machine cost.
On the other hand, the second input that makes the final cost vary the most is mainte-
nance, which again happens because it is a cost that depends on the cost of the machine. Fi-
nally, the other input influenced by the machine's cost is overheads, which have a lower var-
iation than maintenance because they depend on all fixed costs. The final proof that machine
cost is very relevant for the variation of the final cost is that dedication is the percentual input
that makes the part cost vary the least.
Figure 3-9 shows the sensitivity analysis for finish machining and substrate removal
inputs. Although the variation that the inputs cause in the final cost is smaller than in the
WAAM process, they have a superior influence than the substrate cutting and oxides removal
process. In this case, the input that causes a bigger alteration is the machining time, with a
variation of around 7.01 %. On the other hand, the other inputs cause a variation of less than
1% each unless the machine cost, which causes a variation of 1.39 %. These variations are

Figure 3-8 WAAM percentage inputs sensitive analysis

according to the forecasted since the machine time directly influences the number of cutters
that are used, and the machine cost is the second highest with a big difference, 7.01 % to 1.39
%, due to the low allocation value, around 40 %.
Figure 3-10 shows the sensitivity analysis of the percentage inputs. Again, this process
causes a more significant variation in final cost than substrate cutting and oxides removal and
less than the WAAM process. In this case, the input that makes more variation in the final cost
is the allocation that directly influences the cost of the machine allocated to the process, fol-
lowed by the overheads since these depend on all fixed costs, including the machine. Finally
comes maintenance, again due to the cost of the machine and the fact that the allocation is
40% and in last, the dedication since the labour cost is much lower than the others.

42
 Figure 3-9 Finish machining and substrate removal inputs sensitive analysis

Figure 3-10 Finish machining and substrate removal percentage inputs sensitive analysis

As shown in figures A2-3 an A2-4, due to the low acquisition cost and low allocation,
the heat treatment and inspection processes practically do not vary the final cost, with less
than 0.4 % variations. In the case of figure 3-11 and figure 3-12, it is possible to notice that due
to the low acquisition cost and allocation values, the percentage inputs hardly vary the total
cost, except for the dedication that in both cases influences the total cost. However, the heat
treatment dedication is slightly inferior to the dedication trend of the finish machining process
and superior to the inspection process due to the cycle time being slightly inferior to the ma-
chining cycle time and almost twice the inspection cycle time.
Figure 3-13 shows the sensitivity analysis for the number of nozzles and the number of
cutters and their cost. As with the material information sensitivity analysis, the quantity of a
given input and its cost change the total cost equally. One effect previously observed is that
the input of the cutters has a more significant expression in the total cost of the part, of about
7 %, while the number of nozzles does not change the total cost. This is due to the number of
cutters needed and their cost being much higher than the cost of the nozzles.

43
Figure 3-11 Heat treatment percentages inputs sensitive analysis

Figure 3-12 Inspection percentage inputs sensitive analysis

Figure 3-13 Tools sensitive analysis

44
 3.6.2. Final Results

Obtaining the production costs of a part through the PBCM allows cost drivers to be
identified, analysed, and compared, thus helping the decision-making process when studying
the use of a new technology, such as the WAAM technology.
As already mentioned in the model validation phase, a cost per piece of 259.95 € was
estimated to produce 500 cubes. Even though this is a scenario, using the PBCM allows esti-
mates to be made whose final values are close to a real scenario. Since the main objective of
this thesis is to develop a cost model that allows estimating the costs inherent to the produc-
tion of parts using WAAM technology, the production costs mentioned above will be pre-
sented, separating the weight that variable and fixed costs have in the final cost of the part, to
identify the cost drivers better.
In table 3-9 and table 3-10, it is possible to observe the variable and fixed costs and the
respective percentages. As previously mentioned, the values presented are related to the cost
per piece in the production of 500 hollow cubes. Observing table 3-9, it is possible to see that
the total variable costs are 111.41€, representing approximately 43 % of the total costs per
piece. In the case of table 3-10, it is possible to see that the value of fixed costs is 148.54 € which
is approximately 57 % of the total costs per piece. On the other hand, table 3-9 shows that
tooling cost is the highest, representing approximately 54.1 % of variable costs. This happens
due to the cost of the milling cutters and the number needed. The second highest variable cost
is the material cost because the cost of the substrate, the cost of the gas and the cost of the wire
are considered. In the same way, observing table 3-10, the machine cost is the highest, as it
was already observed in the sensitivity analysis. Note that the overheads also have a high
value because it is considered that they represent about 40% of the sum of the other fixed
costs.
Table 3-9 Variable costs and percentages

VARIABLE COSTS Part Cost Percentage of Variable Costs

Material Cost 27.17 € 24.4 %
Labour Cost 17.97 € 16.1 %
Tooling Cost 60.26€ 54.1 %
Energy Cost 6.00 € 5.4 %
Total Variable Cost 111.41 € 100%

Table 3-10 Fixed costs and percentages

FIXED COSTS Part Cost Percentage of Fixed Costs

Main Machine Cost 96.37 € 65.9 %
Fixed Overhead Cost 42.51 € 28.6 %
Maintenance Cost 9.66 € 6.5 %
Total Fixed Cost 148.54 € 100%

45
 Figure 3-14 shows the percentage of each variable and fixed cost as a function of the
total cost of each part produced. The chart shows that the acquisition cost continues to be a
relevant driver cost due to the high allocation value and the high cost of the WAAM machine.
The second driver cost is the tooling cost, as was expected since this input was the higher of
the variable costs due to the cost and number of the milling cutters. In third comes the over-
heads since they are a percentage of fixed costs, including the acquisition cost. Finally, the
fourth most important driver cost is material cost, representing about 10% of the part's total
cost. This is justified by the low raw material consumption of the WAAM machine. However,
it should be noted that the machine and tools costs represent about 60.25 % of the production
cost.
After understanding the structure of cost types, it is necessary to analyse the costs ac-
cording to each process. Figure 3-15 shows in blue the costs of each process. It is possible to
observe that the WAAM process and the process of finish machining and substrate removal
are the processes that represent the largest portion of the total cost. This is justified because
the machines used here are the most expensive in the production cycle because most of the
raw materials are used in the WAAM process and because the most expensive tools are used
in the finish machining and substrate removal process. It is observed that the other processes
represent only 5% of the total production costs. In order to analyse the weight that raw mate-
rials and tools have in the process costs, figure 3-15 shows as orange the process costs without
material and without tools. In this way, it is possible to observe that the reception, heat treat-
ment, and inspection processes remain the same, whereas the substrate cutting and oxides
removal process reduces significantly, although their cost is reduced. This is due to the

Figure 3-14 Cost allocation

46
 cost of the substrate. On the other hand, when the cost of the wire, gas and tools used in the
WAAM process is removed, its cost reduces from 168.92 € to 145.54 €, and as expected, the
cost remains high due to the high WAAM machine cost and allocation. Finally, once again,
the weight that the cutters have in the cost of the piece is clear since the machining process is
greatly reduced when the cost of the cutters is removed, representing 76.7% of the total cost
of the process.

Figure 3-15 Cost per process with and without material and tools costs

In table 3-11, the cycle times of each process are represented. This analysis is important
because it allows observing which processes occupy more time during the 6.49 h needed to
produce a part. One of the conclusions that can be obtained is that the most time-consuming
process is WAAM, which occupies 50 % of the production time. On the other hand, although
the finish machining and substrate removal process is the second most expensive process and
the second most time consuming with 18 % of the production time, it is very close to the time
that heat treatment takes, which corresponds to 17 % of the production time. Finally, the other
processes represent only 15 % of the production time, these being 4 % for reception, 2 % for
substrate cutting and oxide removal and 9% for inspection. Figure 3-16 shows the cost of each
process per hour. Through the analysis, it is possible to observe that the most expensive pro-
cess per hour is the finish machining and substrate removal, which agrees with the previous
analysis since this process is the second most expensive but has about one-third of the dura-
tion of the WAAM process that passes to the second place because its cost is diluted by the
process time. The heat treatment process remained with a slight decrease due to the setup
time. However, the reception and inspection processes, although of low value, increased their
value due to their low cycle time, and the reception increased a lot, going from 0.35 € to 1.41
€. Finally, the substrate cutting and oxide removal process, similarly to the reception process,
due to its low cycle time value, increased from 5.05 € to 32.46 €.

47
 Table 3-11 Processes cycle times and percentages

Process Time Percentage

Reception check and storage 0,25h 4%
Substrate cutting and oxide removal 0,16h 2%
WAAM 3,26h 50%
Finish machining and substrate removal 1,15h 18%
Heat treatment 1,08h 17%
Inspection 0,58h 9%
Total 6,49h 100%

Finally, one of the advantages of using a PBCM is analysing costs according to produc-
tion volume. This allows more information to be available for informed decision-making. In
figure 3-16, it is represented the total cost curve per piece according to the produced volume
for a scenario where the machines are all dedicated and for a scenario where the uptime and
available time are equal. For a better visibility of the chart, it was established that the total cost
would be 20000 € and that the maximum number of pieces would be 200. Observing figure 3-
17, it is possible to observe that in the scenario where the uptime is equal to the available time,
the cost value is almost constant since there is no idle time. The cost is only allocated to the
part when it is being used. In the scenario where the machines are all dedicated, it is possible
to observe that at the beginning, the cost of the piece is very high since the idle time is very
high, and therefore the cost of the machine is allocated to the piece even if the production line
is stopped. However, as the number of parts increases, the graph shows that the cost curve
with all dedicated machines approaches the cost where the uptime equals the available time.
So, it can be seen the effect that fixed costs have on product cost depending on the variation
of the machine's allocation.
Another interesting analysis to understand the effect of fixed costs on part cost is to
analyse the effect of the WAAM process on total part cost when machine allocation varies
since the WAAM process is the process with the highest cost. Figure 3-18 shows the variation

Figure 3-16 Process cost per hour

48
in product cost when the WAAM machine is dedicated and when the uptime is equal to the
available time depending on the production volume. It should be noted that the effect in both
curves is very similar to what was seen previously, with the only difference being when the
uptime and available time are equal, and the production volume is reduced, the curve shows
a high value of about 7300 € due to the other machines have constant allocation values, and
therefore the fixed costs allocated to a part are higher.

Figure 3-17 Model cost boundaries depending on processes allocation

Figure 3-18 Model cost boundaries depending on WAAM allocation

49
50
 4.
Conclusions and Future Work

Considering the smart WAAM project, it was established as an objective to develop a

LCC model that would allow to compute and estimate the cost of producing WAAM prod-
ucts. This way, a PBCM that estimates the costs of implementing a hybrid WAAM system
with additive and subtractive technologies was developed. The developed model allowed to
estimate the costs associated with each phase of the product life cycle using a cradle-to-gate
approach. Furthermore, the model was applied to estimate the cost of producing 500 hollow
cubes of stainless steel AISI316LSI with 7x7x7 cm. Finally, it was possible to identify the main
cost drivers of the production system studied. Therefore, this paper provides a PBCM that
allows to analyse the economic viability of a WAAM production system before its implemen-
tation or obtaining the cost of an existing system. In this way, the PBCM developed is an es-
sential tool for decision-making since, besides estimating the costs inherent to the 500 hollow
cubes scenario of WAAM technology implementation, it also allows an economic analysis of
an existing system to make it more efficient.
By applying the proposed PBCM, it was possible to estimate that the production of 500
hollow cubes would cost 259.95 € per piece. This value includes all variable and fixed costs
associated with production. It was also possible to identify that the fixed costs are higher than
variable costs, the former being 57% and the last 43% of the total cost. On the other hand, it
can be observed that, in general, the highest costs are the machine cost, which represents about
37% of total costs, the tooling cost with about 23% and the overheads cost with about 16% of
total costs.
In a subsequent analysis, it was identified that the processes that most influence the total
costs, approximately 84%, are the WAAM process, due to the high cost of the machine, and
the finishing machining and substrate removal process, due to the cost of the machine and the
cost of the milling cutters required, since they cost 168.92 € and 78.47 € per piece respectively.
Another interesting point was the analysis of the weight of the cutters' cost in the finish ma-
chining and substrate removal, which was identified as being around 77 % of the total cost pf
the process.
When the process is analysed in terms of cycle time, the longest process is WAAM, with
about 50 % of the total time. For this reason, when costs are calculated per hour, the WAAM
process becomes the second highest, with the cost of finish machining and substrate removal
rising to the highest.
51
 Finally, the variation curves of the total cost per piece for a scenario where the processes
are dedicated and another where the available time was equal to the uptime was analysed.
This analysis made it possible to identify the cost boundaries when the production volume
and idle time vary. And, as expected, when the production cost increases, the total cost per
piece decreases significantly. This analysis was also made for a scenario where the only pro-
cess considered a dedicated one was WAAM, and the other processes were in normal condi-
tions. The conclusions were identical, being the only difference the first values of the curve,
where the available time is equal to uptime, since the cost per piece was higher due to the
other processes being in normal conditions.
Throughout the development of the model, some study limitations were identified.
Firstly, the developed PBCM does not consider all WAAM products' life cycle phases, which
are raw materials, design, transport, use, maintenance and end of life. On the other hand, it
also does not consider the costs associated with the space occupied by the machine, as it was
considered that the space occupied would be identical to traditional manufacturing produc-
tion.
Concerning the model application, some limitations were also identified, such as the
fact that a scenario was considered without the possibility of verifying the results. Another
limitation of this study was that a low complexity piece was studied. This could hinder pos-
sible comparison with other production methods.
Since the production volume was small, making it impossible to assess the effect of the
price at which the wasted material could have been sold. Data regarding the quality of the
processes were also not considered. Although it was considered during the model develop-
ment, the amount of material that must be reworked and the costs associated with these ac-
tivities were not considered for the model application. At last, the sales price of the scrap from
the parts that do not pass and cannot be reworked was also not considered.
Considering that WAAM technology is understudied compared to other technologies
and has a lot of interest in the industrial context, some suggestions for future works will now
be proposed. Firstly, the developed PBCM must be used for other parts with different com-
plexities and materials to ascertain the model's ability to study WAAM parts and improve it.
Since WAAM technology is emerging, the model should be updated to consider other devel-
opments and improvements that could change the PBCM accuracy in future works. On the
other hand, it may be helpful to complement this model with the costs inherent to the other
life cycle stages to compare better this technology to the traditional ones. Finally, it is also
important to develop further studies on the economic viability of the WAAM technology and
compare it with other processes in order to better understand in which scenarios the WAAM
technology represents an added value for the production process in economic terms.

52
53
 Bibliography

ASTM International - Standards Worldwide. (n.d.). Retrieved March 31, 2022, from
http://www.astm.org/

Baumers, M. (2012). Economic Aspects of Additive Manufacturing: Benefits, Costs and Energy
Consumption. Doctorial Thesis, Loughborough University, 1-256. Retrieved from
http://www.americanbanker.com/issues/179_124/which-city-is-the-next-big-
fintech-hub-new-york-stakes-its-claim-1068345-1.html

Baumers, M., Tuck, C., Wildman, R., Ashcroft, I., Rosamond, E., & Hague, R. (2012). Combined
build-time, energy consumption and cost estimation for direct metal laser sintering.
23rd Annual International Solid Freeform Fabrication Symposium - An Additive Man-
ufacturing Conference, SFF 2012, 932-944. Retrieved from https://www.re-
searchgate.net/publication/287719627_Combined_build-time_energy_consump-
tion_and_cost_estimation_for_direct_metal_laser_sintering

Bekker, A. C. M., & Verlinden, J. C. (2018). Life cycle assessment of wire + arc additive manu-
facturing compared to green sand casting and CNC milling in stainless steel. Journal
of Cleaner Production, 177, 438–447. https://doi.org/10.1016/j.jclepro.2017.12.148

Bjørn, A., Laurent, A., Owsianiak, M., & Olsen, S. I. (2018). Goal Definition. Em M. Z.
Hauschild, R. K. Rosenbaum, & S. I. Olsen (Eds.), Life Cycle Assessment (pp. 67–74).
Springer International Publishing. https://doi.org/10.1007/978-3-319-56475-3_7

Bjørn, A., Moltesen, A., Laurent, A., Owsianiak, M., Corona, A., Birkved, M., & Hauschild, M.
Z. (2018). Life Cycle Inventory Analysis. Em M. Z. Hauschild, R. K. Rosenbaum, & S.
I. Olsen (Eds.), Life Cycle Assessment (pp. 117–165). Springer International Publishing.
https://doi.org/10.1007/978-3-319-56475-3_9

Bjørn, A., Owsianiak, M., Laurent, A., Olsen, S. I., Corona, A., & Hauschild, M. Z. (2018). Scope
Definition. Em M. Z. Hauschild, R. K. Rosenbaum, & S. I. Olsen (Eds.), Life Cycle As-
sessment (pp. 75–116). Springer International Publishing.
https://doi.org/10.1007/978-3-319-56475-3_8

Campatelli, G., Montevecchi, F., Venturini, G., Ingarao, G., & Priarone, P. C. (2020). Integrated
WAAM-Subtractive Versus Pure Subtractive Manufacturing Approaches: An Energy
Efficiency Comparison. International Journal of Precision Engineering and Manufac-
turing-Green Technology, 7(1), 1–11. https://doi.org/10.1007/s40684-019-00071-y

54
Colegrove, P. A., Coules, H. E., Fairman, J., Martina, F., Kashoob, T., Mamash, H., & Cozzo-
lino, L. D. (2013). Microstructure and residual stress improvement in wire and arc ad-
ditively manufactured parts through high-pressure rolling. Journal of Materials Pro-
cessing Technology, 213(10), 1782–1791. https://doi.org/10.1016/j.jmatpro-
tec.2013.04.012

Colegrove, P. A., Martina, F., Roy, M. J., Szost, B. A., Terzi, S., Williams, S. W., Withers, P. J.,
& Jarvis, D. (2014). High Pressure Interpass Rolling of Wire + Arc Additively Manu-
factured Titanium Components. Advanced Materials Research, 996, 694–700.
https://doi.org/10.4028/www.scientific.net/AMR.996.694

Cunningham, C. R., Wikshåland, S., Xu, F., Kemakolam, N., Shokrani, A., Dhokia, V., & New-
man, S. T. (2017). Cost Modelling and Sensitivity Analysis of Wire and Arc Additive
Manufacturing. Procedia Manufacturing, 11, 650–657.
https://doi.org/10.1016/j.promfg.2017.07.163

Ding, D., Pan, Z., Cuiuri, D., & Li, H. (2015). Wire-feed additive manufacturing of metal com-
ponents: Technologies, developments and future interests. The International Journal
of Advanced Manufacturing Technology, 81(1–4), 465–481.
https://doi.org/10.1007/s00170-015-7077-3

Duarte, V. R., Rodrigues, T. A., Schell, N., Miranda, R. M., Oliveira, J. P., & Santos, T. G. (2020).
Hot forging wire and arc additive manufacturing (HF-WAAM). Additive Manufac-
turing, 35, 101193. https://doi.org/10.1016/j.addma.2020.101193

EDP- (2022). Retrieved 9 June 2022, from https://www.edp.pt/empresas/energia/tarifarios/

Hai-ou, Z., Wang, R., Liang, L., & Gui-lan, W. (2016). HDMR technology for the aircraft metal
part. Rapid Prototyping Journal, 22(6), 857–863. https://doi.org/10.1108/RPJ-05-
2015-0047

Hauschild, M. Z., Bonou, A., & Olsen, S. I. (2018). Life Cycle Interpretation. Em M. Z.
Hauschild, R. K. Rosenbaum, & S. I. Olsen (Eds.), Life Cycle Assessment (pp. 323–334).
Springer International Publishing. https://doi.org/10.1007/978-3-319-56475-3_12

Huang, S. H., Liu, P., Mokasdar, A., & Hou, L. (2013). Additive manufacturing and its societal
impact: A literature review. International Journal of Advanced Manufacturing Tech-
nology, 67(5–8), 1191–1203. Scopus. https://doi.org/10.1007/s00170-012-4558-5

Kara, S., & Li, W. (2011). Unit process energy consumption models for material removal pro-
cesses. CIRP Annals, 60(1), 37–40. https://doi.org/10.1016/j.cirp.2011.03.018

Industrial 1200 Degree 5 Heating Sides Molds Heat Treatment Electric Furnace - Buy Molds
Heat Treatment Furnace,5 Heating Sides Heat Treatment Furnace, Industrial 1200 De-
gree Electric Furnace Product on Alibaba.com. (2022). Retrieved 9 June 2022, from
https://www.alibaba.com/product-detail/Heating-Furnace-Industrial-Heating-Fur-
nace-Industrial_1600115441417.html?spm=a2700.galleryofferlist.normal_offer.d_ti-
tle.6ae28a31nN8J7t&s=p

Kokare, S., Godina, R., & Oliveira, J. P. (2022). Digital Platform for Environmental and Eco-
nomic Analysis of Wire Arc Additive Manufacturing. Em L. M. Camarinha-Matos
(Ed.), Technological Innovation for Digitalization and Virtualization (Vol. 649, pp.

55
 233–243). Springer International Publishing. https://doi.org/10.1007/978-3-031-
07520-9_22

S. Kokare, J.P. Oliveira, R. Godina, “A Cost Model for Wire Arc Additive Manufacturing”,
11th International Conference on Industrial Technology and Management - ICITM
2022, 18-20 February 2022, Oxford, United Kingdom, accepted, presented

Li, J. Z., Alkahari, M. R., Rosli, N. A. B., Hasan, R., Sudin, M. N., Ramli, F. R. (2019). Review
of Wire Arc Additive Manufacturing for 3D Metal Printing. International Journal of
Automation Technology, 13(3), 346–353. https://doi.org/10.20965/ijat.2019.p0346

Priarone, P. C., Campatelli, G., Montevecchi, F., Venturini, G., & Settineri, L. (2019). A model-
ling framework for comparing the environmental and economic performance of
WAAM-based integrated manufacturing and machining. CIRP Annals, 68(1), 37–40.
https://doi.org/10.1016/j.cirp.2019.04.005

Priarone, P. C., Pagone, E., Martina, F., Catalano, A. R., & Settineri, L. (2020). Multi-criteria
environmental and economic impact assessment of wire arc additive manufacturing.
CIRP Annals, 69(1), 37–40. https://doi.org/10.1016/j.cirp.2020.04.010

Ribeiro, Matos, Jacinto, Salman, Cardeal, Carvalho, Godina, & Peças. (2020). Framework for
Life Cycle Sustainability Assessment of Additive Manufacturing. Sustainability, 12(3),
929. https://doi.org/10.3390/su12030929

Rosenbaum, R. K., Hauschild, M. Z., Boulay, A.-M., Fantke, P., Laurent, A., Núñez, M., &
Vieira, M. (2018). Life Cycle Impact Assessment. Em M. Z. Hauschild, R. K. Rosen-
baum, & S. I. Olsen (Eds.), Life Cycle Assessment (pp. 167–270). Springer International
Publishing. https://doi.org/10.1007/978-3-319-56475-3_10

Seow, C. E., Zhang, J., Coules, H. E., Wu, G., Jones, C., Ding, J., & Williams, S. (2020). Effect of
crack-like defects on the fracture behaviour of Wire + Arc Additively Manufactured
nickel-base Alloy 718. Additive Manufacturing, 36, 101578.
https://doi.org/10.1016/j.addma.2020.101578

Singh, S., Sharma, S. kumar, & Rathod, D. W. (2021). A review on process planning strategies
and challenges of WAAM. Materials Today: Proceedings, 47, 6564-6575.
https://doi.org/10.1016/j.matpr.2021.02.632

Super Mini Mill 2 | 40-Taper Mill | Vertical Mills – Haas CNC Machines. (2022). Retrieved 9
June 2022, from https://www.haascnc.com/machines/vertical-mills/mini-
mills/models/sminimill2.html

Wang, F., Williams, S., Colegrove, P., & Antonysamy, A. A. (2013). Microstructure and Me-
chanical Properties of Wire and Arc Additive Manufactured Ti-6Al-4V. Metallurgical
and Materials Transactions A, 44(2), 968–977. https://doi.org/10.1007/s11661-012-
1444-6

Wu, B., Pan, Z., Ding, D., Cuiuri, D., Li, H., Xu, J., & Norrish, J. (2018). A review of the wire
arc additive manufacturing of metals: Properties, defects and quality improvement.
Journal of Manufacturing Processes, 35, 127–139. https://doi.org/10.1016/j.jma-
pro.2018.08.001

56
Xie, Y., Zhang, H., & Zhou, F. (2016). Improvement in Geometrical Accuracy and Mechanical
Property for Arc-Based Additive Manufacturing Using Metamorphic Rolling Mecha-
nism. Journal of Manufacturing Science and Engineering, 138(11), 111002.
https://doi.org/10.1115/1.4032079

57
 APPENDIX

Appendix 1

Goal Definition
The first step of a LCA is defining the goal, where the study's objective is well defined
and described. This is a vital step due to this phase's influence on all the LCA stages. Since the
decisions made in other steps should be consistent with the goal. However, other phases' con-
straints may require a revision of the goal definition. (Bjørn, Laurent, et al., 2018)
According to Bjørn, Laurent, et al. (2018), the definition of the goal involves six aspects
accordingly to ISO requirements, which are:
1. Intended applications of the results
2. Limitations due to methodological choices
3. Decision context and reasons for carrying out the study
4. Target audience
5. Comparative studies to be disclosed to the public
6. Commissioner of the study and other influential actors.

Aspects 1 and 3 are vital due to their importance in the decisions made in other stages
of the LCA. The other aspects are related to communicating the results of an LCA.
The determination of the intended applications of the results is an important aspect due
to the influence that it will have in later LCA phases, like drawing system boundaries, sourc-
ing inventory data and interpretation of the results (Bjørn, Laurent, et al., 2018).
The second aspect is related to the limitations due to methodological choices, this will
determine what the LCA results can and cannot be, for example, if a comparative study does
not consider one or more life phases, it is essential to highlight how that choice limits the
interpretation of results. The limitations stated here must be related to the goal and scope
phases of an LCA, other choices made due to constraints must be documented at a later point
(Bjørn, Laurent, et al., 2018).
Like the first aspect, the decision context and the reasons for carrying out the study are
vital due to its influence in later stages. The reasons for carrying out the study must be under-
stood and clearly connected to the intended application of the study (Bjørn, Laurent, et al.,
2018).
The goal definition must state the target audience of the LCA study and clearly state if
the LCA is a comparative study, and if it is to be disclosed to the public, it must have to fulfil
the ISO requirements. At last, the commissioner and other influential actors must be stated to
prevent conflict of interest (Bjørn, Laurent, et al., 2018).

58
 Scope Definition
The second step of a LCA study is the scope definition. This phase clarifies the product
system and how it should be assessed. For this purpose, the scope must have nine essential
items, which, according to Bjørn, Owsianiak, et al. (2018), are:

• Deliverables: the deliverables must be in accordance with the proposed application of

results defined in the goal definition.
• Object of assessment: the scope must have a precise and quantitative description of
the function of the product system. The function unit defines the qualitative and quan-
titative aspects of the function. This usually includes a function and answers typically
the following questions: what? How much? For how long? Where? How well? After
the definition of the function unit, it is necessary to determine the reference flow. The
quantity of product is needed to realise the function unit.
• LCI modelling framework and handling of multifunctional processes: in this part, the
choice is made for the appropriate LCI modelling framework and the way that the
multifunctional processes will be handled. The LCI framework can be defined in two
different approaches, the attributional and consequential modelling frameworks. The
attributional modelling framework answers the question: What environmental impact
can be attributed to the product? The consequential modelling framework answers the
question: What are the consequences of consuming the product?
• System boundaries and completeness requirements: the system boundaries demar-
cate the product system to be studied from the surrounding environment. Usually, the
system boundaries are not ideal due to three reasons. The first reason is that most of
the studies do not focus on all the life cycle stages (e.g., cradle-to-gate approach), the
second reason is that, if the LCA has a comparative nature, some of the unit processes
and its quantities may be similar. The last reason is that ideal boundaries are almost
impossible to achieve since a unit process typically needs around 5-10 material or en-
ergy inputs that require another 5-10 inputs to be produced. The completeness require-
ments share the actual environmental impact of a product that the LCA aims to cap-
ture.
• Representativeness of LCI data: the representativeness of LCI data has three interre-
lated dimensions: geographical, time-related, and technological. The scope definition
must express the guidance and requirements for the inventory analysis regarding to
representativeness of the LCI data.

59
 • Preparing the basis for the impact assessment: This item has two purposes. The first
is to ensure that it is done according to the goal definition. The second is to prepare
the inventory analysis since the elementary flows depend on the impact categories of
the LCA.
• Special requirements for system comparisons: To compare two product systems that
fulfil the same function, the ISO 14044 standard proposes some special requirements
for the scope definition to guarantee that the product systems may be compared. To
ensure the trustworthiness of the LCA, the international life cycle data guidelines pro-
pose that the uncertainties must be evaluated and communicated and that if a single
indicator makes the comparison, this information must be highlighted.
• Critical review needs: A critical review is not always required, but a review made by
experts not involved with the study is useful to increase the study's quality and cred-
ibility.
• Planning reporting of results: It is essential that the report must be clear and transpar-
ent to prevent erroneous and misleading use of the LCA.

Life Cycle Inventory Analysis

The Life Cycle Inventory Analysis aims to collect and compile data on elementary flows
from all processes. The LCI output is a list of quantified elementary flows that cross the system
boundary, and it will be the input in the LCIA phase. Sometimes, this phase involves the ad-
justment of the scope requirements. The LCIA, typically, is the step that needs more resources
and effort. It is difficult to collect the highest quality data due to the cost that this would in-
volve, but, fortunately, it is not required in most cases. Thus, the LCIA needs a structured
approach to ensure that efforts are spent on collecting data related to the most important
phases of the product’s life cycle (Bjørn, Moltesen, et al., 2018).
Bjørn, Moltesen, et al. (2018) refer to six steps to do the LCIA, which are:

• Identifying processes for the LCI model: This step identifies the processes and their
links and draws the initial system diagrams. The initial system diagram is made under
the scope item, system boundaries. First, to detail the physical value chain, it is neces-
sary to start with the reference flow and then construct the foreground. This process
is done by levels, where: level 0 is the unit process; level 1 are the processes needed to
deliver physical flows; level 2 is the processes required to provide flows that perform
a supporting function; level 3 is the processes necessary to deliver services to the level
0 processes; level 4 are the processes required to produce and maintain the infrastruc-

60
 ture. After identifying all the processes, this method is applied to the other level pro-
cesses, commonly, the processes belonging to levels 3 and 4 are omitted because their
contribution is insignificant, and it is difficult to find data.
• Planning and collecting data: The main objective of the planning step is to balance the
data collection efforts based on the relevance of the data information. The data collec-
tion depends on the specificity level, which can be very high and high, medium, low,
and very low. The very high specificity data are the inputs and outputs measured on-
site, while high-quality data, usually, can be modelled from other site-specific data.
Medium specificity data are collected from LCI databases or from the literature related
to the specific process. The low specificity data are collected from the generic LCI da-
tabase process or data from the literature. At last, very low specificity data is collected
based on an expert or LCA practitioner's judgement.
• Constructing and quality checking unit processes: It is advisable to study several pro-
duction cycles to guarantee that all the data covering a product is analysed. Sometimes
the data collection covers one year of production to gather enough data that represents
the full operation cycle of the process studied. To avoid the risk of constructing incom-
plete unit processes or existing errors in the data flow, it is advisable to check them
before using them in the LCI model.
• Constructing LCI model and calculating LCI results: The inventory modelling is com-
monly done using dedicated software that supports the construction of the product
system model, connects the relevant unit processes, links to available databases, and
links the elementary flows in the inventory results. The LCI results are all the elemen-
tary flows in the process to be studied.
• Preparing the basis for uncertainty management and sensitivity analysis: It is im-
portant to analyse the study's uncertainty and sensitivity, understand how robust the
LCI results are, and advise where future studies should focus. Uncertainty analysis
allows to quantify the uncertainties of the result. Sensitivity analysis allows identify-
ing the parameters that most influence the LCIA results. It is important to consider the
parameters data to be analysed since it is unnecessary to collect all the process data for
the sensitivity and uncertainty analysis.
• Reporting: The reporting of the inventory analysis must have the documentation of
the LCI model at the system level, of each unit process, metadata, LCI results, data
collected for uncertain and sensitivity analysis, and it must contain the assumptions
of each life cycle stages.

Life Cycle Impact Assessment

Life cycle impact assessment is the fourth step of the LCA, and it aims to assess the
magnitude of the contribution of each elementary flow to an impact on the environment. It is
done by examining the product system using impact categories and category indicators and
combining those with the inventory analysis results. This phase is very important since it
61
transforms elementary flows into potential environmental impacts. The LCIA, in most cases,
is done by using a LCA software that enables the automatization of this phase, letting the LCA
practitioner to choose the LCIA model and the necessary settings. It is important to note that
impacts resulting from this phase must be interpreted as impact potentials and not as actual
impacts, risks, or safety margins (Rosenbaum et al., 2018).
The ISO 140040/14044 distinguish the LCIA steps between mandatory and optional
steps. So, according to ISO 14040/14044, the mandatory steps of a life cycle impact assessment
are:

• Selection of impact categories, category indicators and characterisation models: This

phase must be in accordance with the goal and must be done during the scope defini-
tion to guarantee that the inventory data collection is targeted towards what is to be
assessed in the end. The impact categories selection must be in accordance with the
ISO guidelines that define that the impact categories cannot be redundant nor lead to
double counting nor disguise significant imparts and must be complete and allow
traceability. According to ISO, the selection for impact categories, category indicators,
and characterisation models shall be consistent with the goal and scope definition, jus-
tified in the study report, comprehensive regarding environmental issues and well
documented with all the information and sources being referenced.
• Classification: Normally, this step is handled automatically due to the considerable
understanding that is needed.
• Characterisation: In this section, all elementary flows are assessed according to their
contribution to an impact.
According to ISO 14040/14044, the optional steps of a LCIA are:

• Normalisation
• Weighting
• Grouping

Life Cycle Interpretation

The last phase of a LCA study is the interpretation, where the other phases results are
analysed considering the limitations, uncertainties and assumptions made during the study.
Hauschild et al. (2018) point out three main steps for the interpretation phase, which are: the
identification of significant issues, the evaluation of the issues and the presentation of conclu-
sions, limitations, and recommendations.
Significant issues to be identified can be methodological choices and assumptions, in-
ventory data, weighting factors, characterization, and standardization. To assess the relevance
of each issue, it can be used a sensitive analysis or a dominance analysis. To assess discrete
choices done during the other stages of an LCA study, new scenarios with different possibil-
ities can be created to determine the influence on the results (Hauschild et al., 2018).
62
 The second element of the life cycle interpretation is the evaluation stage, where the
stability and reliability of the results from the identification step are determined. This phase
guarantees the strength of the conclusions and recommendations. The evaluation involves
three elements: completeness check, sensitivity analysis and uncertainty analysis and con-
sistency check. The first element serves to see if the significative data is complete. The second
element is the sensitive check, identifying the most important elementary flows and key pro-
cesses. The last element is the consistency check, verifying if the assumptions, methods, and
data are consistent with the goal and scope. When the consistence check is being applied to a
comparative study, it is necessary to evaluate the consistency of the allocation rules, bounda-
ries settings and impact assessment (Hauschild et al., 2018)
The last step of the life cycle interpretation is drawing conclusions, limitations, and rec-
ommendations. The conclusions must be in accordance with the requirements defined on the
scope and should be based on the significant issues and the evaluation of sensitivity, com-
pleteness, and consistency. The recommendations should be based on significant results and
relate to the intended application of the study (Hauschild et al., 2018).

Appendix 2

The following appendix presents some of the sensitive analysis charts.

Figure A2-1 Substrate cutting and oxides removal inputs sensitive analysis

63
Figure A2-2 Substrate cutting and oxides removal percentage inputs sensitive analysis

Figure A2-3 Heat treatment inputs sensitive analysis

64
Figure A2-4 Inspection inputs sensitive analysis

65
66
67
 TOMÁS MEALHA MARTINS A PROCESS-BASED COST MODEL FOR WIRE

2022
RODRIGUES CUNHA

⟨ano⟩
AND ARC ADDITIVE MANUFACTURING

68