Life Cycle Assessment of a Circularity Case Study Using Additive Manufacturing

Abstract

Currently, considering the rising concern in climate change, there is a clear necessity for technologies that can prolong the useful life of products through the ability to repair, re-manufacture and refurbish. As such, additive manufacturing has been a subject of research due to its design and resource consumption capabilities. However, there is a lack of more detailed information regarding environmental performances, especially in Directed Energy Deposition technology. The present paper presents a life-cycle assessment of the production and use of Directed Energy Deposition, making use of foreground data to build a life-cycle inventory and quantify the potential impacts. The equipment is analyzed for its refurbishment capabilities on an obsolete mold, and compared with the environmental impact of producing a new mold through conventional technology. The compiled inventory with detailed and primary information will enrich the current literature on this technology. The impact results show that the robot, deposition table and security cell are the most relevant subsystems for the system production impacts. In the refurbishment analysis, the refurbished mold part has lower impacts than the conventionally produced, thus showing that there is great potential in using additive manufacturing for circular economy loops.

1. Introduction

The industry is faced with a huge challenge: to reduce the consumption of raw materials and energy consumption, while at the same time increasing plant productivity to meet growing consumer demand.

Additive Manufacturing (AM) has been the subject of extensive research in recent years, and it is seen as a feasible alternative to traditional manufacturing procedures. This technology allows the production of constant improvements to the products created without having to change tools of great economic value; hence, it has undeniable usefulness in prototyping, which is why, early in the life of this technology, many referred to it as “rapid prototyping”.

Recognized as one of the pillars of Industry 4.0, AM is changing the logic of industrial processes by redefining the way things are performed, which results in economic, efficient and sustainable improvements. The capacity to easily make very complicated components, the near-complete automation, the range of materials that may be utilized, and the quantity of material saved, all translate into cost savings by removing the need for molds, punches, and skilled personnel.

Metal additive manufacturing, particularly Directed Energy Deposition (DED) technology, has been successfully used in the repair of high and low-mass metallic components [1,2,3,4,5,6,7,8,9,10,11,12,13]. During its lifetime, an extrusion mold may suffer from damage due friction and wear, or even become obsolete due to new requirements and specifications. The field of remanufacturing emerges to avoid the necessity of developing and producing, for example, a new mold [14], and to reduce the expenses related to these elements by machining the damaged region and filling it with deposited material. According to the literature [7], molds that are repaired using DED technologies can last as long as the respective original mold. In contrast, the use of traditional repairing processes, such as gas tungsten arc welding (GTAW), typically results in repaired parts that last between 12.5% and 29.2% of the original life without requiring additional repair [7]. This brings potential benefits in terms of environmental impact mitigation in production processes.

Indeed, the environmental performance of AM technology has been a current topic in research, usually analyzed through life cycle assessment (LCA) studies. LCA is a a standardize methodology [15,16] commonly applied to analyze and quantify the environmental impacts of products and systems in a wide range of sectors [17,18,19,20,21,22,23]. This methodology allows to gather the flows of energy and materials that occur in each life cycle stage, and quantify them into environmental impacts, to identify the main hotspots that contribute most to the total environmental impacts generated. LCA can be used as a tool for support to decision makers in establishing significant impact reduction action plans. However, only a few investigate the DED technology [7,13,14,24,25,26]. In this small bundle, all of them fail to provide a life cycle inventory (LCI) that details the composition of the machinery and auxiliary systems, to assess the system’s production in terms of environmental performance.

Three studies [24,25,26] conducted a comparative LCA with conventional production of steel parts and came to contrary conclusions in determining the production process with the least environmental impacts, highlighting the broad variability of the DED technology in terms of production capabilities. Nonetheless, it may not be fully understood how different part designs and materials affect the environmental performance of the DED system, since only one study provided more information in terms of energy consumption and material deposition efficiency. One of the most recent studies [26], did not take into consideration the impacts resulting from the equipment’s production in both scenarios. When comparing Computer Numeric Control (CNC) technology with AM, the results quantified through the ReCiPe method, showed that neither technology is better than the other. However, when analyzing by endpoints they find that CNC technology has a higher damage to the resources than DED. Additionally, as DED is a fairly new technology when compared to the traditional ones, the operation parameters have the potential to be even more optimized. In an energy-based approach to determine the potential environmental impacts of metal powders used by the DED, one paper [27] carried out a cradle-to-gate analysis based on the cumulative energy demand (CED). The DED system’s production was once again left out of the scope. The authors provided a model that determines the CED for parts made of Ti-6A1-4V by DED which was applied to a topologically optimized bracket. They find that for the paper’s case study, raw material production was the most CED-intensive process step.

In one study [13], the authors analyzed the sustainability of producing a metallic mold by AM compared to conventional techniques utilized, namely die casting of zinc parts for the automotive industry. The authors found that AM has lower impacts justified by better cooling properties of AM molds. Other advantages of using AM found in this case study were less energy demand, faster processing and lower cost. Improvement opportunities were found to be related to process parameter optimization. Similarly, another study [28] performed a comparative LCA for the production of a mold core to be integrated in Carbon Fiber-Reinforced Plastics (CFRP) production. The study compared AM and conventional manufacturing, exploring four manufacturing processes: casting with low-melting alloy, milling from plaster-like material Aquapor, AM with High Impact Polystyrene (HIPS), and AM with powder materials like salt. The last one showed better environmental performance when compared to the others evaluated. In contrast, the conventional product with the low-melting alloy had the worst environmental results. The authors exalted the need for a better understanding of AM technologies to increase efficiency and reduce environmental impacts.

Cerdas et al. [29] analyzed another advantage of AM technologies: decentralized manufacturing. Through LCA, the authors compared the environmental performance between distributed and traditional centralized manufacturing system. The case study was a glass frame, assuming the same final quality of the product for both technologies. Although the authors did not establish which technology has better environmental performance, they identified three main factors influencing the impact results: energy efficiency, the composition of the electricity mix where the product is printed, and the material used. The scarcity of papers related to AM environmental analyses was exalted by Saade, Yahia and Amor [30], also mentioning the lack of transparency in some studies, which is also evidenced in this literature review. The author concluded by specifying some AM characteristics that appear to be immeasurable by usual environmental LCAs such as complex designs, individualization, decreased hazardous exposure to workers, and fewer man-hours needed.

It is clear that there is a pattern of overlooking the potential impacts of the DED system production and its influence on the overall life cycle performance. When analyzing LCA studies of other AM technologies to see if this approach is also present, only one [31] considered the AM system’s production.

Faludi et al. [31] presented an LCA where the authors measured the environmental impacts of Selective Laser Melting (SLM) technology, to determine the main hotspots, whether on the machine and supporting hardware, the material used (aluminum powder), or the energy consumed. Although in virtually all instances, the usage of power during printing had the greatest impact per part, the printer embodied effects, dominated these measures in several low-utilization situations. This emphasizes the importance of including machinery production in the scope of the LCA study, meaning that the current research may be overlooking relevant environmental impacts in their studies for different degrees of printer usage. In Výtisk et al. [32], the manufacturing of an air ejector and orifice plate was compared in terms of environmental performance, by conventional manufacturing and by SLM. Although the results showed that AM had higher impacts in more than half of the impact categories, the authors exalted the relevance in comparing AM and conventional manufacturing machines gate-to-gate, since the depreciation of machine tools for conventional production plays a major role in manufacturing processes.

In a broader analysis of the environmental performance of AM technologies, one research [33] elaborated an extensive review of eco-friendly AM processing of metal materials. The study conducted an analysis of energy efficiency and LCA of material production, product design and manufacturing, product distribution, usage and recycling stages. One of the main takeaways was that wire feedstock materials are more attractive than powder-based feedstock materials due to a faster printing rate. However, extrusion heads needed for wire feedstock are bigger than the ones needed for powder feedstock. Thus, it is crucial to include the impacts of the machinery when comparing technologies and even when comparing different feedstock materials.

More recently, Gouveia et al. [14] analyzed DED’s capability of conducting repairing activities instead of producing new parts. The authors compared the environmental and cost performance of repairing a damaged mold using DED, with discarding and conventional production of a new mold. Here, the authors provided energy and material parameters for the DED operations, specifically for the H13 steel material. Although the LCI of the DED system was not provided, the environmental impacts of the DED production were considered in the final assessment of the case study.The results showed that the repairing process generated less environmental impacts and less costs when compared with convention manufacturing.

This paper presents the continuing work on the environmental analysis of the DED technology [14]. Specifically, the main goal is to provide a robust LCI of DED and to environmentally characterize the subsystems that compose this technology, in the circular activities enabled by AM. The purpose is to provide data that can be used in the future when analyzing the impacts of parts produced by DED. A case study for the refurbishment of an obsolete mold is presented to determine the potential environmental impact reduction that this technology can bring, in the context of a future (more) circular manufacturing industry.

Section 2 presents the gathering of data to build the LCI and the main case study of a circular application of the DED. First, a description of the equipment and the work conducted to select the main components of each subsystem component is presented (Section 2.1). Then, the approach utilized to collect and organize the life cycle data of the DED equipment per subsystem (Section 2.2) is conducted. The matching of the identified flows of energy and mass with the ecoinvent database is referenced in Appendix A. In addition, a mass representation per subsystem was analysed to ascertain the main contributions to each subsystem, on a material level. Then, a detailed description of the LCA case study is presented, together with an explanation of the applied refurbishment process (Section 2.3). Finally, the results are presented in Section 3. The environmental impacts of the DED system production were analysed, as well as the main impacts generated by each subsystem (Section 3.1). Subsequently, the life cycle environmental impacts of the refurbishment and the conventional production process were presented and compared (Section 3.2). Lastly, a discussion on the results is presented in Section 4, including a comparison with other studies in the literature. The final conclusions are drawn in Section 5.

2. Materials and Methods

2.1. DED Component Selection

The manufacturing process via DED is accomplished through the melting and projection of metallic powders, with the energy that triggers the process coming from a high-power laser beam. The following group of equipment refers to all the subsystems necessary for the functioning of this technology. Figure 1, Figure 2 and Figure 3 present an overview of the subsystems that compose the DED.

2.1.1. Robot

In this system, a robotic arm (Figure 1a) with 6 degrees of freedom is used to perform the metallic powder deposition trajectories. Given the importance of the path followed during deposition, a high-precision KUKA industrial robot was selected, with the model number “KUKA KR 30 HA”. It is characterized by a payload of up to 30 kg, a maximum reach radius of 2033 mm and repeatability of ±0.100 mm. The robot’s control system is the KUKA KR C4.

2.1.2. Powder Feeder

The powder feeder (Figure 1b) has the specific function of transporting the metallic powder(s) to the printer head (located in the robot’s terminal axis as seen in Figure 1f) with the desired mass flow. It is a vibrating platform-type device with its own controller. To protect the metallic powders from oxidation and to conduct their transport, the powder feeder uses an inert gas line. In this subsystem, Argon (Ar) is used as the inert gas. The powder feeder is from MEDICOAT, having two powder reservoirs and, consequently, two output lines.

2.1.3. Laser

The energy for the DED process is delivered by a high-energy laser source, depicted in Figure 2a. This laser acts as the source of energy for the fusion of metallic powders. The laser light is created by a ROFIN FL030, with a maximum power of 3 kW, which is guided to the process head by a 100 µm optical fiber cable. The laser source uses the manufacturer’s controller to create the laser beam with a controllable duty cycle and power between 300 W and 3 kW.

2.1.4. Deposition Table

The system is designed to have a heated work surface (Figure 1c), to allow better control of the cooling of the deposited material and thus the quality of the final parts. The work surface is placed on a welding table with nominal dimensions of 2000 × 1000 mm. The table sits bellow a layer of insulating refractory fiber board. This insulation material acts as a thermal barrier between the table and the heaters. This is a set of 8 metal plates of equal size, evenly distributed on top of the insulation. In each of the heating plates, 12 electrical cartridge heaters are placed, each one with a maximum output power of 900 W.

It is intended that the heating system allows for a maximum and uniform 600 °C temperature over the entire work surface. In addition, the table system includes a metal sheet on top of the heaters to protect and isolate them from the process metallic powders. This metal shield also serves the purpose of enabling the collection of the waste powder left by the process.

2.1.5. Chiller

To ensure ideal operating temperatures of several components, the system has a water chiller. The chiller is pictured in Figure 2b. The water chiller is a LAUDA Ultracool UC-0180 with a total cooling capacity of 32.9 kW. Through a network of water hoses, the chilled water is pumped from the chiller unit to the internals of the laser source, to the optical fiber and tool head optics, the nozzle of the printer head and to the solid-state relays that control the output power of the heating table.

2.1.6. Security Cell

Taking into account the nature of the process and its equipment, the DED system design requires a prominent safety component. As such, the system is installed inside a metal cabin (Figure 2c), properly equipped with security elements. From a control point of view, the safety logic is implemented in a Beckhoff Programmable Logic Controller (PLC) with TwinSAFE safety terminals and programmed in TwinCAT 3 in the TwinSAFE environment. The PLC is pictured in Figure 3a.

Inside the metallic cabin, there is an electrical panel (Figure 1e) with safety functions. It has safety PLC terminals that communicate with the safety central PLC, where the respective logic is programmed. This electrical panel is equipped with an emergency button that can be activated inside the cell in an emergency, as well as a button to open the cabin door in case an operator is locked inside. The system is also equipped with a mobile safety console that has an emergency button, as well as buttons to control the locking/unlocking of the cabin door from the operator work station.

The system is equipped with a set of warning lights at strategic points so that the status of the system is intuitively detected by the operator and visible in the industrial area of the building. There are 3 sets of warning lights, two of which are located outside the cabin and one inside.

Each of the equipment that makes up the system has a safety circuit. In particular, the robot controller, the laser, and the powder feeder have internal and external safety circuits that are interconnected via the safety central PLC. Its installation was performed in such a way that if any of the equipment in the system is in an emergency, this state is propagated to the other equipment. This safety configuration makes it possible to use the robot and laser emergency stop button as an external emergency button for the entire system.

2.1.7. Structure

The structure of the cell is a metallic cabin, located in the industrial area of the building and corresponds to a large structure with a working volume of 6.5 × 5 × 3.4 m. The existence of this structure establishes a physical barrier between the place where the process occurs and the operator. Due to the high energy of the laser beam, it is not possible to carry out a print job with people inside the cabin, which thus acts as a fundamental safety element to protect humans from the laser light used. The metallic cabin has an access door to its interior. It is equipped with two door closing sensors, and an actuator that guarantees its locking. The control of the locking/unlocking of the door is performed from the safety PLC, properly programmed according to the state of the system.

The metallic cabin is equipped with a specific safety sensor for the laser. The Laser Spy (Figure 1d) corresponds to a set of sensors installed on the cabin walls that are linked together and form a redundant safety electrical circuit. Its function is to detect a possible breach of the cabin sheet metal by the laser beam and trigger the safety alert on the PLC that will immediately switch off the laser source.

Figure 1. Inside view of the Security Cell. (a) Robot; (b) Powder Feeder; (c) Deposition Table; (d) Photosensitive Sensors; (e) Video Camera and electrical panel; (f) Printer Head.

Figure 1. Inside view of the Security Cell. (a) Robot; (b) Powder Feeder; (c) Deposition Table; (d) Photosensitive Sensors; (e) Video Camera and electrical panel; (f) Printer Head.

Figure 2. Outside back view of the Security Cell. (a) Laser; (b) Chiller; (c) Security Cell.

Figure 2. Outside back view of the Security Cell. (a) Laser; (b) Chiller; (c) Security Cell.

Figure 3. Inside view of the main control system electrical board. (a) Control PLC; (b) Electrical power sensors.

Figure 3. Inside view of the main control system electrical board. (a) Control PLC; (b) Electrical power sensors.

2.1.8. Control System

The central control point of the entire system is a Beckhoff PLC, pictured in Figure 3a. It establishes the communication between all the equipment, namely, between the process, safety, and auxiliary ones. The central PLC communicates via the EtherCAT protocol with the PLC inside the cabin, the robot controller and the powder feeder. Communication with the other components is performed by using the input and output terminals. The central PLC is installed in a central electrical panel, which also contains the solid-state relays that control the heating of the deposition surface. The information diagram of all equipment can be translated into a simplified schematic according to Figure 4.

The system has a video camera (Figure 1e) for real-time monitoring of the process from the outside of the security cell, as well as video recording for data analysis. It is a high-resolution camera with optical zoom from the manufacturer Abus. The camera is located inside the cabin and it gives the viewing point of the interior while deposition work is taking place. For this reason, the camera is exposed to laser radiation and must be equipped with a laser radiation filter that protects its lens and sensor from reflections of the laser beam. It is mounted on the electrical panel inside the cabin. Its data is sent via RTSP (Real Time Streaming Protocol) to the DED cell’s internal network, from where it can be accessed to be transmitted by a video player.

Figure 4. Schematic of the control interconnections for the DED system.

Figure 4. Schematic of the control interconnections for the DED system.

2.1.9. HMI

The system also includes an interface with the controller, which is depicted in Figure 5. It is an HMI (Human Machine Interface) where it is possible to establish communication with the various components of the system, being able to read their status and send commands. The HMI makes it possible to monitor the process variables of the different equipment in real time. This interface is implemented in a Beckhoff PC with a 23 inches monitor and is located next to the metallic cabin, on the outside.

2.1.10. Printer Head

The printer head (Figure 1f) consists of a nozzle that conveys metal powder and inert gas to the deposition surface. The nozzle is model COAX12-V6 and is produced and marketed by Fraunhofer IWS. The laser, which contains optical fibers, enter the printer head in the collimator and passes through a set of optics that allow the laser beam to be focused, coaxially with the metallic powder, on the deposition surface. The ideal deposition point is located at a distance of 13 mm from the bottom of the printer head and corresponds to a laser beam focus of Ø1 mm.

The thermal camera installed on the printer head is the EMAQs camera system, developed by Fraunhofer IWS. This camera is tasked with measuring the temperature of the deposited material throughout the process. The camera system is controlled by a Beckhoff PLC. Inside the camera, there is a compressed air circuit to control a pneumatic actuator that enables an internal protection filter. The camera is mounted on the focusing optics assembly and can be integrated into an advanced temperature control system developed by the same manufacturer. LompocPro 8 software has a control algorithm that allows modulating the laser power as a function of the temperature of the deposited metallic powder, measured by the EMAQs camera. LompocPro 8 produces data files with measured temperature and/or laser power as a function of time, as well as camera images, for studying the process.

2.2. DED Life Cycle Inventory

The first step for the development of the study is related to the survey of the composition of the equipment referring to the DED system. Thus, it was necessary to collect information on the constitution of the machine at the component level, taking into account the total mass of each component and constituent materials, as well as manufacturing processes inherent to the production of the components themselves. It should be noted that the energy and mass flows that occur during the assembly of the DED machine were not considered. The DED machine was assembled and placed in the laboratory, where the study was conducted. As such, it was possible to organize field sessions to collect all the data necessary for the characterization of the equipment. Due to its complexity, the DED system was divided into several subsystems according to their function, and characterized at the component level within each of these subsystems, as shown in Figure 6.

The main data source came directly from the DED system development team. Through an exhaustive data collection, either in the collection of the references of the manufacturers of the components in situ or from the measurement of their dimensions, among other approaches, as well as through bibliographical research to define the material composition and respective mass quantities, it was possible to collect most of the information needed to build the DED system LCI. However, whenever there was a need to address the lack of information, the ecoinvent V3.7.1 database [34] and the existing literature were also consulted. For some components, it was not possible to collect enough information from the manufacturer, so it was necessary to search for similar components from other manufacturers that had the same function and apparent composition, to fill the lack of data. Thus, it was possible to build a DED machine inventory with high reliability and a similar level of detail for most subsystems.

In the case of the chiller subsystem, the information available in one paper for a chiller in an SLM system [31] was used (model SMC HRGC002-A). Through the analysis of this model, it was found that the underlying components were similar to the DED machine model LAUDA Ultracool UC -0060/0240. Thus, for the information in this inventory to be suitable for the DED system chiller, the quantities of the inventory components were adapted through a mass proportion, so that the total mass quantity of the original chiller inventory was equal to the total chiller mass of the DED system:

$$amountofcomponentintheS5Chiller=\frac{amountofcomponentintheliteratureChiller\times weightofS5Chiller\left[\mathrm{kg}\right]}{weightoftheliteratureChiller\left[\mathrm{kg}\right]}$$

(1)

It should be noted that the inventory built for the DED system also includes most of the transformation processes in the production of each component, based on the approach present in Faludi et al. [31].

Taking into account the data collected for each subsystem, the life cycle data was uploaded into the SimaPro V.9.2.0.1 software [35]. Thus, for each material and manufacturing process, it was necessary to carry out the matching between the inventoried materials and manufacturing processes and those present in the ecoinvent V3.7.1 database. To avoid allocation within the transformation and treatment systems [15,16], the allocation library at the point of substitution (APOS) was selected. The efficiency of the system in terms of material deposition was determined by computing the relation between the amount of material that was deposited in the piece and the total amount that was provided by the powder feeder:

$$materialdepositionefficiency[\%]=\frac{materialdeposited\left[\mathrm{g}\right]}{materialdeposited\left[\mathrm{g}\right]+wastedpowder\left[\mathrm{g}\right]}$$

(2)

To this end, both the powder feeder and the mold were weighed before and after the deposition. For the current case study, the material deposition efficiency of the system was 37.5%. With the imputation of the processes in Simapro and the corresponding matching, the initial acquisition of data was completed to characterize in detail the mass and energy flows underlying the production of the components that make up the systems of each of these technologies. Regarding the mass flows, it was possible to obtain a preliminary mass distribution by the DED equipment subsystems through the collected inventory, shown in Figure 7.

As expected, the security cell (S6) is the subsystem with the highest mass contribution compared to the others, as it includes the steel of the surrounding structure of the machine that protects the employee from the effects of the laser when it is in operation. Next is the steel table system (S4), where the material deposition takes place, and the Robot (S1) which includes the cast iron robotic arm and the stainless-steel plinth. The printer head (S9) and Media (S10) do not make a significant contribution to the total mass of the DED system.

The analysis of the energy consumption (kWh) of this system was made possible with the installation of energy sensors connected to the robot, the laser head and the heating table. In parallel with the web service system in the robot controller, the monitoring system allowed a holistic approach to the production parameters, in terms of its visualization and monitoring in real-time, as well as its historical evaluation.

2.3. Case Study

The present work will follow the standardized methodology of LCA [15,16]. The main goal of this study is to compare the environmental performance of the refurbishment of an obsolete thermoplastic mold piece for new applications with the business-as-usual (BAU) scenario of disposal of the obsolete mold and conventional manufacturing of the new one. The scope will focus on a European context, and consider the main energy and mass flows that occur from extraction of raw materials, to transport and production of the mold, including the AM and conventional machinery production. Figure 8 presents the system boundaries of this study with the quantified energy and mass flows obtained at the laboratory, through the use of sensors, data acquisition equipment and measurement techniques made on-site. The functional unit is one fully functional mold piece with a life expectancy of 1 million cycles for the production of thermoplastic components. It is assumed that the refurbishment of the obsolete mold will produce a new useful mold, with the same durability as its counterpart from the conventional production process.

The BAU scenario starts with the disposal of the obsolete mold for landfill. With this, it is necessary to manufacture a new mold part with the desired features. For this, the conventional process starts with the machining of an 18.11 kg block of steel to the approximate dimensions of the final part, generating steel scrap. Then, the block goes through two finishing processes of Electrical Discharge Machining (EDM), first through wire EDM, and after with die sinking EDM to obtain the mold piece in the final form with the required characteristics. The conventionally produced mold piece weighs 10.32 kg and is considered ready to go to the use stage.

The new circular scenario to handle the obsolete mold piece allows to take advantage of the existing piece as a basis and work on providing its new features through the AM. In the DED process, the laser used for the heat input is guided to the printer head by a optical fiber cable, where a series of optical elements focus it on the surface to be deposited. The feedstock material to be added is H13 powder, which is fed to the deposition head through a series of hoses. The printer head then focuses the powder in the area to be deposited. Alongside the H13 powder, Ar gas is introduced to shield the metal against oxidation.

The first step for the refurbishment via DED is to optimize the printing parameters according to the powder and the material where it will be deposited. A Design of Experiment (DOE) was performed on substrates using different ranges of laser power, speed, powder feed rate, carrier gas flow and shield gas flow as presented in Figure 9. This combination of parameters generates a specific layer height and width, and optimal hatch, which are important inputs for the deposition path. The optimal printing parameters used for the deposition can be found in

Table 1. Printing parameters used on the mold depositions.

Printing Parameters	Values
Laser Power	400 W
Speed	3 mm/s
Powder feeder rate	4 g/min
Carrier gas flow	3 L/min
Shield gas flow	35 L/min
Layer height	1.5 mm
Layer width	1.5 mm
Hatch spacing	0.85 mm

.

In general, the refurbishment process of a typical part has three main phases: (1) material removal, (2) material deposition, and (3) high-precision machining. In this case, between the material removal and the deposition, the cavities of the mold that had to be partially removed were filled with solder. Therefore, the material deposited on the top of the cavities would then stand and create a surface to cover them. An intermediate machining process was performed before the deposition to remove excess solder, as can be seen in Figure 10. After depositing the material, the mold is ready for the final machining process, which is normally performed using EDM due to tight tolerance requirements. Details of the main operations are given as follows.

The 3D models of the original mold and the final aimed geometry were provided by the project partners. First, a comparison between these two models was performed in SolidWorks® using the command “Compare Volume”. From this command, it was possible to determine what material regions needed to be either removed or added.

Figure 11 represents the different models developed to define the manufacturing processes. Focusing on models (a) and (f), the original and final geometry of the mold respectively, it is clear that several cavities and slots were eliminated. On the other hand, the operation resulted in surface inclinations and grooves.

The definition of the first machining volume was based on the Z height of the slots to be eliminated and on making all the adding-material regions reachable by the deposition tool. By comparing the original model with the model after first machining (see Figure 11b), it was possible to obtain the volume that needed to be machined. This model was exported on a .step format, uploaded into the MasterCAM® software, and then the machining procedure was written into a .H file using a post processor.

To ensure the correct deposition of the powdered metal on the cavities, filler metal was soldered manually (highlighted in red in Figure 11c). Then, to machine the resulting soldered surface, the second machining 3D model (Figure 11d) was adopted and the surfaces were machined to the final height before deposition. As a result, complete and planar surfaces were obtained, which are very suitable for depositing the powdered metal.

Figure 10. Different steps adopted throughout the mold refurbishment.

Figure 10. Different steps adopted throughout the mold refurbishment.

The deposition volume was defined based on the comparison between the model of intermediate machined mold and the model of the final aimed geometry (highlighted in blue in Figure 11e). After the deposition a final machining was still required, and as such, extra material was added to the deposition volume, i.e., the height was increased by 1 mm. The resulting volume was exported to a .stl file, uploaded into an open-source slicer software (Slic3r®), and then exported to a .gcode file. The .gcode file includes information about the robot’s position and movements, but the actual process parameters are defined by a different post-processor.

Lastly, by comparing the deposited (Figure 11e) and the final models (Figure 11f), it was possible to obtain the new volume to be machined by EDM (using the .step file of the final geometry). Although the tool program had several complex paths with challenging perimeters, the deposition occurred without any problems. The deposited material had no visible cracks or major imperfections.

Figure 11. 3D models used on the different phases of the mold refurbishment: (a) original model; (b) mold after first machining process; (c) mold after filling cavities with solder; (d) mold after second machining process; (e) mold after deposition; (f) final mold geometry.

Figure 11. 3D models used on the different phases of the mold refurbishment: (a) original model; (b) mold after first machining process; (c) mold after filling cavities with solder; (d) mold after second machining process; (e) mold after deposition; (f) final mold geometry.

3. Results

The results regarding the potential environmental impacts of the production of the DED system, as well as the life cycle impact assessment comparison of the mold production by refurbishment and the BAU scenario are presented in this section. The quantification of the environmental impacts was conducted through the ReCiPe Endpoint (H) method [36] in the Simapro software. The results include impacts from infrastructure and do not consider the transformation processes for each component.

3.1. Environmental Impacts of the DED System Production

The main environmental impacts from the DED system are presented in Figure 12, with the total ecopoints (pt) for the production of one DED system, aggregated by impact category (Figure 12a), as well as the percentage in contribution from each category to the total DED impacts (Figure 12b).

The results of the environmental damage generated by the production of the DED system show that the main impact categories that contribute to the total impact damage are Human carcinogenic toxicity, HCTox (45%), Fine particulate matter formation, PM (23%), Human non-carcinogenic toxicity, HNCTox (19%) and Global Warming, GW (12%). The remaining categories (defined as “Others” in Figure 12) do not reach more than 1% of the total environmental impacts, whereas the highest only generates 0.3% of the total impacts (Fossil resource scarcity). A more detailed analysis was conducted to assess the source of the environmental hotspots for the most relevant categories for the environmental performance of the DED production. In Figure 13, the contributions from each subsystem of the DED equipment for the relevant categories are shown.

Figure 12. Total environmental damage from the DED system production per impact category, using the ReCiPe Endpoint (H) method: (a) in absolute values and (b) in percentage.

Figure 12. Total environmental damage from the DED system production per impact category, using the ReCiPe Endpoint (H) method: (a) in absolute values and (b) in percentage.

Figure 13. Contributions from each subsystem to the total environmental damages for the relevant impact categories, using the ReCiPe Endpoint (H) method.

Figure 13. Contributions from each subsystem to the total environmental damages for the relevant impact categories, using the ReCiPe Endpoint (H) method.

In all categories, it is evident that the Robot (S1), the Deposition Table (S4) and the Security Cell (S6) are the most contributing subsystems to the total impacts generated during production. S1 generates 30% and 26% of the total impacts from the GW and the PM categories, respectively. This is mainly due to the electronic equipment within the controller (60% in GW and 54% for PM), namely the printed wiring board and integrated circuits that are mostly manufactured in countries highly dependent on fossil fuels for energy production. The manufacturing of the steel for the plinth and the cast iron for the robot arm also have significance in these categories, with 38% and 42% of the system’s total impact in the GW and PM categories, respectively. These metals are the main drivers of the HCTox category, with almost 94% of S1’s impact, because of the generated slag from the electric arc furnace during steel production, which is related to the presence of hazardous toxic metals which can leach into the ecosystem [37,38,39,40]. In the HNCTox category, the main impacts generated by S1 come from the copper and the gold present in the cables and in the integrated circuits existing in the robot arm and the controller, respectively, with a total of 32% of the DED system HNCTox impacts.

Regarding the other impact-relevant systems, the results show that the impacts generated by the S4 system in all categories are mostly related to the steel structure that composes the deposition table, specifically its production and associated energy consumption. This is especially relevant in the PM and HCTox categories, where the system contributes 33% and 42% to the total impacts, respectively. The S6 system contributes 24% to the total GW impacts. In this category, the production of the aluminum from the safety cell structure represents 58% of S6 contributions, followed by the production of the tin plate steel rails with 37%. In the PM category these materials are responsible for 95% of the total impacts. In both of these categories, the production of these materials has higher contributions due to the high level of energy consumption, which is mostly derived from fossil fuel resources. Regarding the HCTox category, the tin plate steel rails are responsible for 68% of the system’s contributions (because of the toxic metal sludge generated during production), whereas the aluminum structure generates 30%. However, this material is responsible for 89% of the S6 total impacts in the HNCTox category due to the sulfidic tailings (that contain sulfide minerals, a combination of sulfide with metals such as iron, copper, nickel, lead and zinc) [41] generated in the extraction of raw materials through mining operations, which can leach acid to the soil and groundwater if not properly handled. Other systems, such as S3 and S7, have also some relevant contributions in the HNCTox category, with 11.5% and 10.8%, respectively. This is mostly due to the production and refining of the copper present in the wiring.

Overall, the results indicate that the main drivers of the DED machine production impacts are the extraction and processing of steel and aluminum, as well as the production of electronic components, mostly due to the generation of toxic waste and the combustion of fossil fuels for energy purposes. The total impacts for the production of one DED machine are 3149.65 pt, from which 3107.64 pt are derived from these four categories. The total damage points will be integrated into the assessment of the case study, considering the amount of usage during the lifetime of the machine. For this, a 10-year lifetime scenario was considered, with a daily operation of 4 h, during which one refurbishment takes 45 min to be completed and one minute for it to be exchanged to start a new process. With this, a total of 12,650 refurbishment processes were estimated. This number was used to include the environmental impacts of the DED system for one refurbishment process in the case study.

3.2. Environmental Impacts—Conventional Production vs Refurbishment Using DED Technology

The environmental impact comparison between the conventional production of a new mold piece and the refurbishment of the obsolete piece are presented in Figure 14. The results are presented in ecopoints (pts), aggregating all of the environmental effects from the analyzed activities that are within the previously established system boundaries. European conditions were assumed and the electricity generation was adjusted for the Portuguese 2020 electricity mix, using data from national entities [42,43].

The results show that there is a significant reduction (98%) in environmental damage when the mold piece is refurbished with the DED technology, instead of being discarded and a new mold piece being produced. In both scenarios, there is a clear influence of the four major categories previously mentioned: GW, PM, HCTox and HNCTox. The remaining categories (marked as “Others”) do not contribute more than 2% in the BAU scenario and 3% in the refurbishment. In the conventional production, the HCTox contributes with more than 50% of the total damage, whereas in the refurbishment this category contributes with 47%. This is mostly due to the electric arc furnace slag formed during steel production of the new mold (around 74% of HCTox impacts in the conventional production are due to the production of the initial block) and, in the refurbishment scenario, these impacts are due to powder waste treatment (in this case landfill was selected).

Figure 14. Total environmental damage comparison between conventional production of a new mold piece and the refurbishment of the obsolete version via DED, using ReCiPe Endpoint (H): (a) in absolute values and (b) in percentage.

Figure 14. Total environmental damage comparison between conventional production of a new mold piece and the refurbishment of the obsolete version via DED, using ReCiPe Endpoint (H): (a) in absolute values and (b) in percentage.

The second most relevant category in these activities is the PM, with 23% of total damage in the conventional production and 21% in the refurbishment. In the conventional production, the main drivers of these impacts are the atmospheric emissions from the steel production of the new mold, whereas, in the refurbishment, this category has mainly contributions from the argon and Portuguese electricity consumption during the DED process. In these results, GW is the third most relevant category, with 12% of total impacts in the conventional production, due to steel production and its additives, together with the resins present in the EDM wiring processes. In the refurbishment scenario, this category represents 17% of total environmental damage, mostly generated by the argon and electricity consumption in the DED, similarly to the PM category. Finally, in the HNCTox category, most impacts are due to consumables with copper and electronic components, in the wiring EDM in the conventional production (10% of total damages). In the refurbishment (12% of the total damages), the production and end-of-life treatment of the H13 powder is the main contributor to this category, followed by argon consumption during deposition.

4. Discussion

The results of the LCA study for the refurbishment application clearly show a high level reduction in environmental impacts when compared with the conventional process. This indicates that AM can have a significant role in the future circular manufacturing industry. In all analyzed categories, the environmental damage reduction was over 90%. When analyzing current literature, in one study [7], the comparison between conventional repairing and DED repairing (using also H13 powder) resulted in similar results as this paper, but had a high-level of uncertainty due to lack of information, a proper LCI of the DED system and the use of estimates from the industry. In this study, the reliability of the results provide from a robust LCI data acquisition process, mainly the on-site measurements for this specific process. The resulted LCI can be useful to researchers to validate their previous studies and conduct a more solid investigation on the life cycle of this technology. Other repairing studies [14], reach the same conclusion, even considering that the repaired mold only reaches one-third of the conventional lifetime of cycles. In other studies [13,26], the authors compare conventional manufacturing with DED processes. Although the scope is different, the authors also come to the same conclusion that the 3D printed product has less environmental impacts, although this relation is not always clear when comparing production processes due to the different levels of technology development.

The decision to opt for the traditional route or DED technology depends on the degree of confidence and reliability that the final finish of the part corresponds to the necessary specifications, or that guarantees beneficial sustainability for the end user. Moreover, additional studies must be conducted regarding the technical performance of the refurbished mold piece to ensure that it is up to par with the conventionally produced mold. Up until the conclusion of this study, the characterization of the refurbished mold piece had not yet been conducted. With the number of cycles to determine the comparability for the use stage, the results for the refurbishment would be adapted to match the number of cycles of the BAU scenario (as conducted in [14]) in a future study, together with other relevant mechanical performance features that may differ from the conventionally produced mold piece in finishing or treatment processes, as well as in the use stage. Nonetheless, a recent study [7] of a similar process, using the same technology and the same material, concluded that the durability of the mold is equivalent to the mold produced by a conventional process, and as such, the comparison made in this study is very plausible.

5. Conclusions

The main goal of this study was to create life cycle models for the production and use of the DED system, from a circular economy perspective, to measure the environmental performance of the additive manufacturing system in a real-life application. To do so, a mapping of materials and components was conducted to create a reliable LCI for the machine system and its subsystems.

The inventory was built using foreground data from laboratory measurements, invoices, and sensors with a high level of reliability. The first step consisted of describing the systems into which the machine can be divided to facilitate the data collection. The second step involved the description of all components for each defined system. The third step consisted of allocating the components to the life cycle impact assessment software processes. This was a complex process that involved large amounts of data. The data collection was possible in a detailed manner since the DED system was built in the laboratory, through the purchase and assembly of various parts and components. Thus, it was possible to collect information about the description of most of the components. Overall, the study successfully measured and obtained detailed data regarding the main components and materials that compose each subsystem of the DED. It is expected that the built LCI can provide LCA researchers and the scientific community with important life cycle information for this AM technology.

The built LCI provided a clear picture of the material demand and resource consumption for the production of these systems, where the Security Cell (S6), the Deposition Table (S4) and the Robot (S1) had the highest quantity. When analyzing the environmental impacts of the production of the DED, most of the environmental damage derives from the emissions of toxic compounds, particulate matter and GHG during the production of the constituting metal and electronic components.

The case study referred to the refurbishment of an obsolete thermoplastic injection mold into a new geometry. Yet, for repairing processes, for example, a mold with cracking problems, the process may involve fewer resources than the traditional processes, and translate into lower environmental impacts. The refurbishment process of the mold included machining processes (CNC and EDM), manual solder application and DED. There were not any major problems during the processes, and the material was machined effectively and deposited with no visible cracks. The complex tool path with challenging borders was performed successfully and the deposition process was a success in terms of providing the obsolete mold piece with the desired features for new applications. Future studies will conduct a characterization of the refurbished mold to confirm its equivalency to the conventionally produced pieces, together with an economic and social assessment to assess other relevant parameters, such as operation time and production costs, and the overall sustainability of the process. Concluding, the quantified environmental performance of this refurbishment has great potential in reducing the environmental footprint of the tooling industry. As such, this study showed that there is a high potential for repairing processes using AM technology in the circularity sphere.