A Direct Material Reuse Approach Based on Additive ... - ScienceDirect

Available online at www.sciencedirect.com

ScienceDirect Procedia CIRP 61 (2017) 229 – 234

The 24th CIRP Conference on Life Cycle Engineering

A direct material reuse approach based on additive and subtractive manufacturing technologies for manufacture of parts from existing components Van Thao Le*, Henri Paris, Guillaume Mandil, Daniel Brissaud Univ. Grenoble-Alpes, CNRS, G-SCOP, 38000 Grenoble, France * Corresponding author. E-mail address: [email protected]

Abstract This paper presents a direct material reuse approach. The approach consists in producing new parts directly from existing components avoiding the material recycling stage, using a combination of additive and subtractive technologies. The proposed approach seems able to reduce resources consumption and waste during the manufacturing process. The major steps of the approach are presented. The process planning for combined additive and subtractive processes is particularly discussed. The process planning is designed through the feature concept, based on the knowledge of individual processes, technological requirements, and available resources. Finally, the feasibility of proposed approach is validated through a case study. ©2017 2017The The Authors. Published by Elsevier B.V.is an open access article under the CC BY-NC-ND license © Authors. Published by Elsevier B.V. This Peer-review under responsibility of the scientific committee of the 24th CIRP Conference on Life Cycle Engineering. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the scientific committee of the 24th CIRP Conference on Life Cycle Engineering Keywords: Direct material reuse; Process planning; Additive manufacturing; Machining; End-of-life product.

1. Introduction Today environmental issues of manufactured products receive a significant attention from most countries in the world. Environmental policies and regulations [1] push manufacturing companies to produce their products in a cleaner and greener manner. The production process must be balanced from economic and social points of view by using resources efficiently and reducing environment impacts [2]. Moreover, the increase of end-of-life (EoL) products has also become an unavoidable social issue due to the increase of product demand. This causes an increase of environmental impacts, and contribute to reach the limits of landfill facilities. To solve this problem, the researchers and manufacturers are looking for efficient strategies to recover EoL products while taking into consideration the environmental benefits [3]. Until recently, there were two possibilities to recover EoL products. The first scenario consists in recycling the material of EoL products, and then using produced material in a new production cycle. However, the energy consumption of recycling systems remains important [4]. In addition, added

value, functionality and built-in energy of original products are generally lost during the recycling process [1]. On the other hand, in the second scenario EoL products are recovered via repairing or remanufacturing processes [4]. Remanufacturing is an industrial process, which allows the conversion of EoL products into products in a like-new condition one. This strategy can extend the life-time of products, reducing manufacturing cost, and waste, as well as environmental impacts [1]. Hence, the remanufacturing is today considered as an alternative solution for the recycling. In the past three decades, additive manufacturing (AM) has attracted an increasing attention of researchers both in the academic and industrial sectors. AM offers a special ability to build complex parts without using cutting tools, cooling fluid and fixture systems. This technique has been identified as having the potential to provide a number of sustainability advantages [5]. First, AM provides the capability of freeform production that removes the limitations of conventional manufacturing and opens new prospects for the design of innovative and lightweight parts [6]. This will reduce material and energy consumption in manufacturing. Secondly, the

2212-8271 © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the scientific committee of the 24th CIRP Conference on Life Cycle Engineering doi:10.1016/j.procir.2016.11.190

230

Van Thao Le et al. / Procedia CIRP 61 (2017) 229 – 234

adoption of AM offers shorter and simpler supply chains, and more localized production [7]. Finally, AM also provides a huge potential to reduce environmental impacts and offers a significant sustainability benefits [8]. Taking into account these benefits and performances of AM, this article presents a direct material reuse approach based on combined additive and subtractive technologies. The approach aims to reuse EoL components or existing components directly to produce new parts, avoiding the recycling phase. The obtained parts are then intended for another product. This is totally different from other strategies, such as repairing and remanufacturing of components. 2. Related works A number of studies have reported the possibility of AM for remanufacturing applications. Cottam et al. [9] used the laser cladding to build Ti-6Al-4V entities on Ti-6Al-4V substrate. By observing microstructures and microhardness of built samples in the function of cladding parameters, the authors could determine what parameters that can be used for repairing components. Terrazas et al. [10] presented a method, which allows the manufacture of multi-material parts using discrete runs of electron beam melting (EBM) system. The authors successfully built a copper entity on the top of an existing titanium part. Their results demonstrated the feasibility of EBM for remanufacturing. Dutta et al. [11] stated that directed material deposition (DMD) technology offers the particular benefit of a minimum heat affected zone that enables a high-quality repaired part to be achieved. In fact, these works focused on investigating microstructures of the samples to observe the phase transformation traversing the heat affected zone. This could predict the mechanical bonding between the AM-built zone and the substrate zone. However, this prediction has not yet validated by the tests on mechanical properties, such as the tensile testing. To fill this gap, Mandil et al. [12] carried out a study to confirm the feasibility of EBM for the build of new entities on existing parts. In their work, both the observation of microstructures and the tensile testing go to the conclusion of the existence of strong bond between the EBM-built entities and the existing part. Literature shows that the use of directed energy deposition (DED) techniques (e.g., DMD) have a significant efficiency in the remanufacturing of worn-out or damaged components, particularly for high value components, such as turbine blades, molds and dies. Wilson et al. [13] stated that laser direct deposition was efficient for remanufacturing turbine blades. Rickli et al. [14] also presented a framework for a remanufacturing system using DMD technique, which was able to restore high value EoL cores to original specifications and ensure their quality. However, these works only aim to return EoL components in a like-new condition, and extending their lifetime. In comparison with DED processes, the powder bed fusion (PBF) processes have some limitations for remanufacturing applications due to the limited build envelope and the deposition of material that is only conducted on horizontal flat surfaces. However, there are numerous components, which can be remanufactured by these processes.

For example, Navrotsky et al. [15] used selective metal melting (SLM) to repair gas turbine burner tips. This research also demonstrated the potential of SLM technique for building new features on existing components. Recently, Newman et al. [16] proposed a framework, entitled ‘iAtractive’, which can generate different strategies for manufacture of parts based on existing components. The development of such strategies is presented in [17]. These strategies use CNC machining, AM process (i.e., fused filament fabrication, FFF), and inspection process interchangeably. However, the strategy is only efficient for producing prismatic plastic parts. In some cases, the strategy is not time-effective and reduces the tensile strength of the produced parts. 3. Direct material reuse approach 3.1. Approach overview In this work, taking into account the performance of AM techniques a direct material reuse approach is proposed. This approach can give a new life to existing components by transforming them into new parts intended for another product [18]. The existing components could be extracted from EoL products. In comparison with the approach proposed in [16] and [17], the proposed approach is extended to metallic components. For this purpose, the current study focuses on metal-based AM techniques, namely DED processes (e.g., DMD and CLAD - construction laser additive deposition) and PBF processes (e.g., SLM and EBM). The approach consists in reusing the existing parts directly to produce the final parts, using a combined subtractive and additive manufacturing, and inspection processes. This combination takes the advantages and performance of AM and CNC machining techniques while minimizing their disadvantages. The general manufacturing process of parts based on existing components is described in Fig. 1. This process consists of three major stages: the pre-processing stage, the processing stage and the post-processing stage. Existing part Subtractive manufacturing Cleaning and initial inspection

Pre-processing of existing part

Inspection

Heat treatment

Additive manufacturing

Final inspection & additional operations

Final part

Processing: Subtractive and Post-processing additive manufacturing sequence

Fig. 1. Major stages of manufacturing process of parts from EoL components.

x Pre-processing stage. First, the existing part or component extracted from EoL products is cleaned and evaluated. Its actual geometry and shape are then achieved by a system of measurement and scanning to generate the CAD model. The information and the CAD models of the existing part and the final part are used for the processing stage. x Processing stage. This stage refers to design a manufacturing sequence containing subtractive, additive, and

Van Thao Le et al. / Procedia CIRP 61 (2017) 229 – 234

inspection operations, and even heat treatment. Such a process sequence will be designed through the concept of additive manufacturing and machining features. x Post-processing stage. The post-processing consists of final inspection operations, and additional operations, such as labeling, and so on. 3.2. Workflow for subtractive and AM process sequencing In the current study, the design of manufacturing process sequence is based on the concept of features (i.e., AM features and machining features). The definition of machining features (MFs) proposed in [19] is adopted. For AM features (AMFs), an AMF is defined as a geometrical form and associated technological attributes for which there exists at least an AM process known to build it. This AM process is independent from the processes of other features. The design of process sequence is done in two major steps (Fig. 2). In the first step, from the available information and the CAD models of the existing part and the final part, AMFs and MFs are extracted. The process planning for AMFs and MFs is then designed in the second step. Info and the CAD model of existing part

Info and the CAD model of final part

Extraction of AM and machining features List of AM and Machining features & Model of relations between features

Design of process planning for AM and machining features Process design for each manufacturing feature

Setup design

Setup sequencing

AM and machining process sequencing

Fig. 2. Workflow for design of process sequence using the feature concept.

3.3. Extraction of AM and machining features The extraction process of AMFs and MFs is based on the available information of the existing part and the final part, the knowledge of AM and machining processes, the tool accessibility constraint, and the available resources. In this study, this process was manually performed using the Boolean functions of a CAD software. The description of this process has been deeply presented in [20]. Note that the AMFs and MFs are extracted independently. This allows the design of process for a given feature independently from the process of all other features. In addition, the relationships between these features are also built during the feature extraction. These relations can be categorized into geometrical relations, topological relations and precedence relations.

x The geometrical relations. The geometrical relations refer to either dimensional tolerances or tolerances on the position or the orientation of the geometrical bodies of the features [19]. x The topological relations. These represent the associativity of a feature with the neighbor features. For MFs, they can be classified into three types: “opens-into”, “opens-onto” and “intersects-with” [19]. For AMFs, this relation is defined in term of “starts-onto”, namely, an AMF is built on a surface of another feature. These relations are also considered as the attributes of features. x The precedence relations. These relations are built considering the tool accessibility constraints in AM and machining processes. In PBF processes, such as EBM and SLM, the build of AMFs must start from a flat surface of existing part. Hence, the flat surface must be achieved by machining before the build of the corresponding AMF. In machining, the tool accessibility constraint is also used to avoid the collision between cutting tools and the part. For instance, if there exists at least a MF between two AM features, AMF1 and AMF2, and the build of AMF2 on AMF1 causes an inaccessibility of cutting tool for machining the MF, the precedence relations are then assigned to MF, AMF1 and AMF2, as follows: AMF1 o MF o AMF2. The list of extracted features and their relationships achieved from this step are used for the process planning. 3.4. Design of process planning for AM and machining features From the list of extracted features and the relations between the features, the process planning is designed based on the knowledge of AM and machining processes, the technological requirements and the available resources (e.g., AM machines, machine-tools, and so on). The design of process sequencing consists of three tasks (Fig. 2). a. Process design for each feature. In this task, a process (a sequence of operations) is selected for each MF considering the required specifications. Depending on the geometrical specifications (e.g., dimensional tolerances and surface roughness) a sequence of operations is decided. For example, if the geometrical specifications of a MF are tight (e.g., IT ≤ 0.05 or Ra ≤ 0.8), then this MF is performed by three operations: roughing (R), semi-finishing (SF) and finishing (F). In addition, the tools and a machine-tool, and a fixture system are also selected to perform each MF. In this study, we consider that each AMF is built in an independent setup. Thus, the process design for an AMF refers to select the most compatible AM technique while considering available AM machines. The selection of the fixture system to assure the stability of the part in AM process is also considered. b. Design of the setups. The design of machining setups contains number of tasks, such as selection of machine-tools and fixture systems, grouping and sequencing of operations. However, in current work we only focus on the grouping and sequencing of operations. Other tasks, such as selection of

231

232

Van Thao Le et al. / Procedia CIRP 61 (2017) 229 – 234

machine-tools and fixture systems, have been presented in other works [19, 21]. As mentioned above, each AM feature is built in an independent setup. Hence, the grouping and sequencing of operations are only performed for machining operations. Grouping operations of MFs is based on several criteria, such as tool approach direction (TAD), tolerance requirements, and precedence relations among the features. Normally, the MFs having the same TAD could be grouped in the same setup. However, depending on their tolerances, they can be assigned in the same setup with a suited machine-tool, or machined in different setups. The features with tight tolerances must be assigned to the same setup to eliminate fixture system errors. In order to reduce the number of tool changes in a setup, the machining operation using the same cutting tool should be grouped in the same cluster. Sequencing of operations within a setup is generally based on the machining precedence constraints. They are determined from the MFs’ geometrical information, and tolerance requirements, as well as the tool interactions and fixture interactions. If MF1 is the precedence of MF2, then MF1 is sequenced before MF2. For example, if a through hole is divided by a slot, then the through hole should be machined before the slot. Moreover, it is also necessary to consider the burr problem in machining, especially when there is a geometrical intersection between the features. The burrs must be rejected to ensure the surface quality. For instance, if there is an intersection of two holes. Thus, the hole with smaller diameter is referred to be machined before the last one. c. Setup sequencing. The setup sequencing is also based on the precedence constraints between the features. These precedence constraints are determined from the geometrical and topological relations between the features, and the tool accessibility constraints. Following the topological relations, if a MF provides the starting surface for an AMF, the setup of MF must be sequenced before the setup of AMF. Similarly, if the rough state of MF is the actual state of AMF, then the setup of MF must be scheduled after the setup that builds AMF. Taking into account the tool accessibility, if the build of AMF causes an inaccessibility of tool for machining MF, then the machining setup of MF should be performed before the setup of AMF, and so on. In addition, the locating and clamping stability constraints during the manufacturing process are also taken into account. One must first machine the MFs in the setups that enables the most stable positioning and a good clamping in the next setups. If the machining of MF1 causes an instability of the part in the next setup that performs another feature (e.g., MF2 or AMF1), then MF1 should be machined after the setups of MF2 or AMF1. These ensure the best stable positioning and a correct clamping of the part during manufacturing process. In the case of sequencing two AMFs, particularly when they are built by PBF techniques, the stability of the part and the required powder quantity are also two important factors to be taken into account. Normally, if there are not any precedence constraints between AMF1 and AMF2, we prefer to build in first time the feature with smaller dimension along the build direction.

d. Identifying intermediate states of features in manufacturing processing. Identifying intermediate states of MFs is a very challenging task in process planning. In this study, a MF can pass though different machining and AM phases. The state of the MF could vary from raw state (e.g., the state of the existing part) to final state with some intermediate states. For example, the states after R and SF operations in machining, the state after AM process of another AMF, and the final state after F operation. In fact, the thermal during AM process, particularly in EBM and SLM, has a significant influence on the surface finish. Consequently, if the MF is finished before the process of AMFs, the final state of machined surfaces may not satisfy the required surface quality. So the following question is proposed: “is a MF performed before or after the AM phases?” To solve this question, the following constraints are taken into account. (1) The required quality of the final part. The first purpose of the process planning aims to achieve the quality of the final part. Thus, this constraint is first considered and respected. (2) The productivity. Taking into account this constraint allows us to reduce the manufacturing time by eliminating the intermediate setups and the tool change as much as possible. (3) The material use efficiency. This means reusing material of existing part as big as possible to make a final feature. As a result, the powder to be used to build the feature in AM process and the chip removed in machining will be reduced. These constraints and their combinations enable us to determine the intermediate states of a feature. Note that the constraint (1) is considered as the mandatory constraint; and other ones are considered as optional constraints. An illustrating example is presented in Fig. 3. Taking into account the constraints (1), (2) and (3), the intermediate states of the cylinder of final part are determined through the scenario 1 or 2. The scenario 1 is used when the required roughness of the surface A is tight (e.g., Ra = 1.6 µm) and/or the material volume (orange, with the height H) to be reused is important. Following the scenario 1, the required quality of surface A is respected while using material of the existing part and the powder to create the cylinder efficiently. Otherwise the scenario 2 is applied when the volume (orange, with the height H) is not important and the required quality of surface A is compatible with AM processes. In this case, MF1 is machined before the build of AMF1 by using a flat end mill with big diameter that could reduce the time to machine MF1 when compared to that in the scenario 1. MF1

AMF1

H

Scenario 1

Final part

Existing part

A

MF1 AMF1

A

Scenario 2 Phase 10

Phase 20

Fig. 3. Example illustrating the combination of the constraints (1), (2) and (3).

233

Van Thao Le et al. / Procedia CIRP 61 (2017) 229 – 234

4. Case study The proposed approach is illustrated using the test parts, as presented in Fig. 4. Here, the pocket (P), the hole (H) and the surfaces (fS1 to fS7) of the final part require a high surface precision. The roughness of the surfaces (eS1, eS2, and eS3) of existing part satisfies the quality of the final surfaces (fS1, fS2 and fS3) and they are compatible with AM processes.

surface quality (e.g., Ra = 1.6 µm). If MF2 is finished before the setup of AMF3, the thermal in AM process building AMF3 may cause an influence on its surface roughness. AMF1 is considered as the precedence feature of AMF2 because the build of AMF2 before AMF1 may cause an instability of the part when AMF2 has been built on the part. MF9 AMF1

The Fig. 5 presents extracted features. The intermediate part refers to the material volume of existing part to be reused to produce final part. Here, we have nine MFs, MF1 to MF9 (e.g., MF1 and MF2 corresponding to the top surface and the “irregular” step of intermediate part and other MFs corresponding final surfaces have to be machined), and three AM features, AMF1, 2 and 3. Note that AMF2 and AMF3 are extracted independently by considering the tool accessibility constrains even though they can be built in the same AM phase. Existing part

MF8

MF1

AMF2

MF3

AMF3

MF4

MF2

MF5

MF6

Final part P

MF7

fS7

eS1 fS1

fS6

Fig. 6. Relations between the features. eS2

H eS3

fS2

fS3

fS4 fS5

Fig. 4. Test parts of the case study.

Extracted features

AMF3 MF3

AMF2

MF4

MF9

MF1 MF6 MF5

MF7

From the information of extracted features (i.e., the attributes of features), the process of each feature is designed. Considering the relations between the features, as presented in Fig. 6, the setups and the setup sequencing are designed (Fig. 7). MF1 is machined in the first setup to ensure the stability of the part in the setup 2. The features MF3, 4, 6 and 7 having the same TAD are grouped in the setup 4, in which MF6 is sequenced before MF7. Similarly, the features MF2, 5, 8 and 9 having the same TAD are grouped in the setup 6 and performed after the build of AMF3 (in the setup 5) to ensure the required quality.

MF8 AMF1

Setup 1

Setup 2

Setup 3

Setup 4

Setup 5

Setup 6

MF1

AMF1

AMF2

MF3

AMF_3

MF2

MF2

Existing part

Intermediate part

MF4

Final part

MF6

Fig. 5. Extracted features: MFs, AMFs.

The relations between the features are presented in Fig. 6. The blue links present the geometrical and topological relations between the features. For example, AMF3 starts onto the surface obtained by machining the MF3, and so on. The red dash links present the precedence relations. MF1 is precedence feature of AMF2 because AMF2 must be built from the top surface of the intermediate part achieved by machining the MF1 (Fig. 5). Taking into account tool accessibility, MF4 corresponding to pocket (P) and MF7 corresponding to hole (H) are also the precedence features of AMF3. The green dash links show the precedence constraints, which are made by considering the quality constraints and the stability constraints. For example, MF2 must be performed after the final AM phase (AMF3) to assure the required

MF7

MF5

MF8

MF9

Fig. 7. Setups and setup sequencing.

5. Discussion and conclusions The major challenge in the development of the proposed approach is how to perform the process planning for additive and subtractive processes while considering the technological constrains, the constrains of individual processes, the tool accessibility constrains, and the available resources. The proposed method for process planning design was based on

234

Van Thao Le et al. / Procedia CIRP 61 (2017) 229 – 234

the concept of AMFs and MFs. To generate the process planning the relationships between the extracted features and different precedence constrains were determined and respected (Section 3.4). In addition, the method proposed in [22] was also adopted to solve conflicts or trade-offs between some constraints during the design of process planning. This method allows the decision of compromises among conflicting constraints, and the generation of an optimal solution. Compared with the traditional remanufacturing of components (e.g., the works presented in [13] and [14]), the proposed approach allows achieving new parts from existing parts intended for another product; meaning that the final part has new functionalities that might be totally different from those of the existing part. Compared with the approach developed in [16] and [17], the proposed approach is applicable for the manufacturing of metallic parts. Moreover, the produced parts have mechanical characteristics comparable to those of the parts produced from workpiece by traditional processes (e.g., machining), as demonstrated in our recent work [12]. The generated process planning not only respects the quality and manufacturing constraints, but also minimizes the production time by eliminating intermediate setups and a number of tool changes. The feasibility of the proposed approach has been demonstrated by the case study (Section 4). However, it remains necessary to evaluate the environmental impacts and compare the proposed approach with the traditional manufacturing process, and the remanufacturing in terms of environmental benefits. 6. Future work Future work will consist in the development of models for the evaluation of the economic efficiency and environmental impacts of the proposed approach. Another interesting issue to investigate is to determine which EoL product types would give the best components for the manufacturing of the desired parts. Acknowledgements The authors would like to thank Auvergne-Rhône-Alpes Region of France for its support in this project. References [1] Gehin A, Zwolinski P, Brissaud D. A tool to implement sustainable endof-life strategies in the product development phase. Journal of Cleaner Production 2008;16:566–76. [2] Bashkite V, Karaulova T, Starodubtseva O. Framework for innovationoriented product end-of-life strategies development. Procedia Engineering 2014;69:526–35.

[3] Lindahl M, Sundin E, Östlin J. Environmental issues within the remanufacturing industry. Proceedings of CIRP Life Cycle Engineering 2006:447–52. [4] King AM, Burgess SC, Ijomah W, McMahon CA. Reducing waste: repair, recondition, remanufacture or recycle? Sustainable Development 2006;14:257–67. [5] Huang R, Riddle M, Graziano D, Warren J, et al. Energy and emissions saving potential of additive manufacturing: the case of lightweight aircraft components. Journal of Cleaner Production 2015. [6] Thompson MK, Moroni G, Vaneker T, Fadel G, et al. Design for Additive Manufacturing: Trends, opportunities, considerations, and constraints. CIRP Annals - Manufacturing Technology 2016;65:737–60. [7] Huang S, Liu P, Mokasdar A, Hou L. Additive manufacturing and its societal impact: a literature review. The International Journal of Advanced Manufacturing Technology 2013;67:1191–203. [8] Gebler M, Schoot Uiterkamp AJM, Visser C. A global sustainability perspective on 3D printing technologies. Energy Policy 2014;74:158–67. [9] Cottam R, Brandt M. Laser Cladding of Ti-6Al-4V Powder on Ti-6Al-4V Substrate: Effect of Laser Cladding Parameters on Microstructure. Physics Procedia 2011;12:323–9. [10] Terrazas CA, Gaytan SM, Rodriguez E, Espalin D, et al. Multi-material metallic structure fabrication using electron beam melting. The International Journal of Advanced Manufacturing Technology 2014;71:33–45. [11] Dutta B, (Sam) Froes FH. The additive manufacturing (AM) of titanium alloys. Elsevier Inc.; 2015. [12] Mandil G, Le VT, Paris H, Suard M. Building new entities from existing titanium part by electron beam melting: microstructures and mechanical properties. The International Journal of Advanced Manufacturing Technology 2016;85:1835–46. [13] Wilson JM, Piya C, Shin YC, Zhao F, Ramani K. Remanufacturing of turbine blades by laser direct deposition with its energy and environmental impact analysis. Journal of Cleaner Production 2014;80:170–8. [14] Rickli JL, Dasgupta AK, Dinda GP. A descriptive framework for additive remanufacturing systems. International Journal of Rapid Manufacturing 2014;4:199–218. [15] Navrotsky V, Graichen A, Brodin H. Industrialisation of 3D printing (additive manufacturing) for gas turbine components repair and manufacturing. VGB PowerTech 12 2015:48–52. [16] Newman ST, Zhu Z, Dhokia V, Shokrani A. Process planning for additive and subtractive manufacturing technologies. CIRP Annals Manufacturing Technology 2015;64:467–70. [17] Zhu Z, Dhokia V, Newman ST. A novel decision-making logic for hybrid manufacture of prismatic components based on existing parts. Journal of Intelligent Manufacturing 2014:1–18. [18] Le VT, Paris H, Mandil G. Using additive and subtractive manufacturing technologies in a new remanufacturing strategy to produce new parts from End-of-Life parts. 22nd Conference of French Mechanics (CFM2015), 24th - 28th August, Lyon, France 2015. [19] Paris H, Brissaud D. Modelling for process planning: The links between process planning entities. Robotics and Computer-Integrated Manufacturing 2000;16:259–66. [20] Le VT, Paris H, Mandil G. Extraction of features for combined additive manufacturing and machining processes in a remanufacturing context. Advances on Mechanics, Design Engineering and Manufacturing, Springer International Publishing; 2017, p. 179–89. [21] Paris H, Brissaud D. Process planning strategy based on fixturing indicator evaluation. The International Journal of Advanced Manufacturing Technology 2004;25:913–22. [22] Descotte Y, Latombe JC. Making compromises among antagonist constraints in a planner. Artificial Intelligence 1985; 27:183–217.