TY - JOUR
AU1 - Stolt,, Roland
AU2 - Elgh,, Fredrik
AB - Abstract The selective laser melting (SLM) process has created new possibilities for the manufacture of new lightweight jet engine components with lattice structures replacing solid sections. Hopes are to reduce the density of the component and thereby saving weight. To introduce the new manufacturing process, the components need to be redesigned and verified to comply with an array of requirements concerning, for example, strength, aerodynamics, and manufacturing. To find out how a capability of designing and evaluating components for the SLM process can be built into an organization, an interview investigation has been conducted at an aerospace company finding the state of practice in technology and product development. The impact of introducing SLM is thereafter estimated. The result is that introducing a novel manufacturing process will primarily influence the methods used to predict product lifecycle performance. An important finding is that it is currently difficult to include a topology optimization step in the multiobjective design evaluation environment used at the company due to which the complexity would increase significantly. Graphical Abstract Open in new tabDownload slide Graphical Abstract Open in new tabDownload slide SLM, selective laser melting, additive manufacturing, aerospace, DFAM, design for additive manufacturing, manufacturability, topology optimization Highlights A proposal is made to integrate topology optimization in multidisciplinary design space exploration. Technology and product development processes in an aerospace company have been described from interviews. Design for additive manufacturing of aerospace components has been proposed. 1. Introduction There is an ongoing endeavour to reduce the weight of aircraft since it gives an immediate payback in reduced fuel consumption and thereby lower emissions and improved economy. Aircraft manufacturers are therefore constantly exploring new designs, materials, and manufacturing concepts to reduce weight. Selective laser melting (SLM) is an additive manufacturing (AM) powder bed fusion process that currently is not commonly used in structural parts of jet engines. Other AM processes such as directed energy deposition are more common in aerospace applications. There are expectations that SLM possibly can complement or replace the commonly used sheet-metal and casting processes for manufacturing turbine parts and thereby gain advantages. For that, new design concepts will have to be developed to utilize the 'complexity for free' characteristic of the SLM process. Designing a component for manufacture with AM is referred to as design for additive manufacturing (DFAM). For example, see Diegel (2019). It is central in DFAM to bring down the amount of material in the part to keep the printing time low and to prevent thermal distortion of the shape. Low weight is also beneficial in many applications from a functional perspective. For these reasons, it has become a common practice to do topology optimization (TO) when designing parts for AM. This indicates where material can be removed or replaced with lightweight structures without compromising the structural integrity of the part. To successfully design for SLM, TO needs to be integrated in the component design process in companies so that all requirements can be taken into account. To detail how this can be done in aerospace, a manufacturer of jet engine components has been studied in this paper. The objective has been to find out how the design of parts for the SLM process can be incorporated in the design process of the company. A central part of the company's design process involves a system that semi-automatically makes multidisciplinary evaluations of conceptual designs to build knowledge around new product designs in early stages, finding how trade-offs can be made between the sometimes conflicting requirements. The interview studies have previously been published in the transdisciplinary engineering conference TE2018 (Stolt, Heikkinen, & Elgh, 2018). This paper is a reworked and extended version. It makes a novel contribution to the integration of TO in the process and exemplifies it with a component. 2. Literature There are general recommendations for designing for AM processes, such as Klahn, Leutenecker, and Meboldt (2015), Leutenecker-Twelsiek, Klahn, and Meboldt (2016), and Booth et al. (2017). These concern geometric recommendations, how to orient the part in the printer, and how to build support structures. Often, an AM process is more expensive than a traditional manufacturing process, especially in larger series competing with die-based processes. An additional advantage is needed to motivate the extra cost. AM is unique in the complexity of shape, material (property gradients), hierarchy (multiscale of features), and functionality (embedded parts) that it can produce. This introduces a number of opportunities such as manufacturing any shape without the need of investing in tools and dies, making highly individualized components without increasing the cost, embedding cellular structures to save material, and consolidating assemblies into single parts. A comprehensive review of applications is found in Rosen (2014). In Hällgren, Pejryd, and Ekengren (2016), a study of an aerospace application is performed showing that the cost of an aerospace part increased by a factor of 20 when changing from a forged to AM. Still, considering the high value of weight saving in the application, the change could be motivated. Another aerospace application is found in Thompson et al. (2016) involving a fuel injection nozzle. The number of components reduced, so the number of welds and bracings went from 25 to just 5. In contrast to traditional design for manufacturing where the focus is to eliminate manufacturing difficulties, the objective of DFAM has been defined as to 'Maximise product performance through the synthesis of shapes, sizes, hierarchical structures, and material compositions, subject to the capabilities of AM technologies' (Rosen 2014). To succeed with this objective, process- and designer-driven methods have been presented (Hällgren et al., 2016) where either the focus is on leveraging computational synthesis to optimize performance or human knowledge to reduce costs. According to Rosen (2016), computational synthesis methods can be classified into: (i) size optimization – given dimensions are optimized, (ii) shape optimization – altering control vertices for surfaces and curves (generalized size optimization), (iii) TO – finding the optimal distribution of material, changing the overall shape, arrangement of shapes, as well as the connectivity, and (iv) evolutionary optimization – utilizing evolutionary algorithms to explore a variety of shapes. TO can be categorized into ground truss and volume based (Rosen 2016). In ground truss methods, the domain is comprised of a mesh of struts along nodes that are then optimized with respect to size dimensions (e.g. diameter of strut) and/or position of nodes with respect to compliance. The topological differences are made when struts are removed due to the size dimension getting near zero. Volume-based methods instead focus on optimizing densities or boundaries of volumetric elements (in 3D called voxels). There are several different types of volume-based TO methods; homogenization, bi-directional evolutionary structural optimization, solid isotropic material with penalization (SIMP), and level set. In homogenization-based TO, the material properties of varying porosity (amount of microscale voids) (Suzuki 1991) or density (Cheng et al., 2017) are obtained by utilizing calculated scaling laws and used in the optimization of deflection or volume, removing volumetric elements where the density or non-void constituents are close to zero. Bi-directional evolutionary structural optimization originates from evolutionary structural optimization that worked by hard-killing FE elements where low stress levels could be identified and was first reversed, adding material to high-level stress areas called additive evolutionary structural optimization, and finally bi-directionality was added to allow elements to be removed and added (Xia, Xia, Huang, & Xie, 2018). SIMP is the most commonly implemented method in commercial software (Rosen 2016) and optimizes the density of each voxel, where close to zero density voxels are removed and intermediates are penalized. Finally, the level set optimizes a level-set function that implicitly defines the part boundary (Wang, Wang, & Guo, 2003). See Verbart, Langelaar, van Dijk, and Van Keulen (2012) for an example of level-set TO approach used to optimize an L-bracket and 2D aerofoils. TO often leads to complex structures that are difficult or impossible to manufacture. For this reason, the software for TO has been developed to account for tool access and tool removal for machining and casting or mould manufacturing, respectively. AM, however, allows much more geometric complexity and does not necessarily affect manufacturing costs for different levels of complexity. In a way, the challenges with TO efficiency to explore the complete design space have led to a shift from manufacturing limiting the realization of optimal design to the design stage (Brackett, Ashcroft, & Hague, 2011). The trouble is no longer to manufacture the geometries, but to computationally generate them. TO and AM have shown a good match for orthopaedic implants such as bone implants where it is very important that the implants have similar characteristics as the host (Wang et al., 2016). AM can be used to produce porous, geometrically complex, and anisotropic material properties and TO can eliminate or complement a trial-and-error approach. To address the TO efficiency challenges, a number of approaches have been presented: hard-kill elements (remove from FE model), iterative remeshing and coarsening where possible and only keeping fine meshes where needed, and boundary-based methods (Brackett et al., 2011). Further, even though AM is less constrained, there are still design aspects that require consideration. In Liu et al. (2018), an extensive summary of the state of the art within TO with respect to DFAM can be found. A number of DFAM topics are listed including material properties, lattice structure infill, and post treatment. Lattice structures can be defined as a type of cellular structure and are synonymous with lightweight structures and foams. In Panesar, Abdi, Hickman, and Ashcroft (2018), a strategy for realizing lattice structures that are derived using TO is presented. The strategy takes into consideration both mechanical performance (in terms of defined as well as uncertain loading looking at strain energy) and manufacturing performance (in terms of support structure, processing effort, and design to manufacture discrepancy). The strategy is separated into three main parts: Realizing lattice design – Starts by utilizing an unpenalized SIMP TO method to minimize structural compliance (inverse stiffness) by varying densities. The density map can then be used to define the overall topology in three ways: (a) 'Intersected lattice' – Create a discrete solid–void solution and intersect it with a uniform lattice structure. (b) 'Graded lattice' – Vary lattice structure according to the density map and remove the ones under a density value and create solid from the ones over a specific density value. (c) 'Scaled lattice' – Vary lattice structure according to the density map without removing elements or creating complete solids. These three are in addition to the extremes (no lattice and entire design space as lattice) assessed with respect to solution optimality, design effort, and support requirements. Which lattice structure to choose from is not clear but different types are presented, either strut based or surface based. Numerically evaluate mechanical performance – Made with respect to defined/intended loading as well as loading with uncertainty. Assess with respect to manufacture – Done by looking at three criteria: (a) Support structure requirements – Especially important for AM where support structure can be required where there are downward facing faces as well as where there are risks for residual stresses causing distortion. (b) Processing effort – Considering number of triangles or size of STL file, which will require more computer memory and processing effort to generate sliced file. (c) Design-to-manufacture discrepancy – Evaluated using a function depending on surface area, volume, and build as well as designed weights. In Cheng et al. (2017), a similar approach that also takes advantage of the intermediate densities and utilizes lattice structures that are self-supported is presented. It is called a homogenization-based topology optimization methodology and keeps the original shape (like the Scaled method introduced above). In Li et al. (2018) a 'generative design optimization method' is presented, which takes advantage of both lattice design (gyroid-based) and TO, where the variation of the lattice (gyroid) structure is given by the optimum density variations from the TO that also takes advantage of homogenization. In the example shown, robustness is also discussed and tested by simulating the effects of varying load directions. These two examples might not in the end be considered topologically optimized structures since the topology is known in hindsight but still take advantage of TO methods. Lattice-based TO methods implemented in commercial software have also been identified (Arisoy, Musuvathy, Mirabella, & Slavin, 2015) SLM has been shown to be an excellent AM technique to produce strong and stiff lattice structures (Challis et al., 2014). However, electron beam melting (EBM) is better than SLM with respect to lower residual stress, fabrication speed, and energy efficiency but requires holes about twice as large to remove powder (due to elevated build temperatures). Powder removal can be difficult in lattice structures. A generally rougher surface is obtained with EBM (Takezawa, Yonekura, Koizumi, Zhang, & Kitamura, 2018). Material produced with SLM is not completely dense, which makes it less resistant to tensile loads than compressive. Therefore, the structures that carry tensile strength are where failure occurs. This can be improved with smoother surfaces (Sercombe et al., 2015). To make the topologically optimized SLM component useful in an actual aerospace application, all aspects such as aerodynamics, fatigue, thermal loads and assembly, inspection, and manufacturability must be taken into account when designing. This is a very complex multidisciplinary problem that currently cannot be represented in a single optimization. Multidisciplinary optimization frameworks specifically for AM have been presented (Yao, Moon, & Bi, 2017), which is promising but still lacks the topological shape optimization capabilities that are an important aspect in many products to motivate a transformation to AM. One strategy used is to generate a multitude of variants and analyse them from all aspects in a multidisciplinary manner and thereafter selecting a set of variants for further elaboration. A case where this strategy has been applied is reported in Heikkinen, Stolt, Elgh, and Andersson (2016) and Stolt, André, Elgh, and Andersson (2017). The requirements as expressed by customers can vary throughout the design process. To handle this fluctuation, the design process needs to be agile so that it becomes possible to quickly respond to these changing requirements. An example from the automotive industry is found in Almefelt, Berglund, Nilsson, and Malmqvist (2006). An organization that can quickly respond to the fluctuations will have a competitive edge (Stolt, Johansson, André, Heikkinen, & Elgh, 2016). Therefore, increasing the company's ability to respond to fluctuating requirements should be addressed when the design process of a company is being formalized. Some of the referenced papers concern how computational tools are used to drive DFAM towards a design with less material and better adaptation to printing. Here, the generation and use of lattice structures is central to why it is elaborated in some of the referenced papers. Two papers concern AM processes and the properties that result from them. In five of the papers, multidisciplinary design frameworks and how to systematically include the design knowledge from several disciplines are presented. It should be noted that designing for the AM process will often include complex lightweight structures that are computationally expensive to generate. This highlights the problem addressed in this paper, namely creating a multidisciplinary design environment that is responsive to changing requirements in the design process and still including the geometric shape optimization for the AM process. 3. Finding the State of Practice To find out how the current development process is conducted, a semi-structured qualitative interview was carried out. The studied organization employs about 2000 people and is part of a large global corporation. The company manufactures components for jet engines such as turbine frames, axles, and fans. The interviews involved four different interviewees at senior positions in the company. The interviews were conducted in the spring of 2017. The results of the interviews were reported in Stolt et al. (2018). The interviews were semi-structured and qualitative, encompassing 20 questions. The four interviewed professionals had the following functions: A technical lead engineer, responsible for the functions of technical systems, and requirements in manufacturing processes An engineer in thermal calculations A manager responsible for technical sales and preparation of quotations An engineer responsible for the product development for specific products after agreement with customers. The interviews were conducted separately, and the duration of each interview was around 1 hour. The answers were voice recorded and then summarized in writing. There were four categories of questions that were applicable in this study: Technology development Product development Requirements Tools and systems The interviews showed that technology development is directed towards identifying future needs from customers and improving the current working procedures and manufacturing processes. One part of technology development is developing and qualifying new conceptual designs resulting in a proof of concept as shown in the top part of Fig. 1. Figure 1: Open in new tabDownload slide Technology and product development. Adapted from Stolt et al. (2018). Figure 1: Open in new tabDownload slide Technology and product development. Adapted from Stolt et al. (2018). Technology development begins with the creation of a conceptual design. This is done in a creative manner involving professionals from several disciplines. When a conceptual design has been created, a parametric computer aided design (CAD) model of it is made. This is used as a base for a multidisciplinary design exploration to find out the response on the performance of the conceptual design when varying the design parameters. By doing so, knowledge about the conceptual design is gained. This makes it possible to elaborate the concept to the level that it be qualified for presentation to potential customers (proof of concept). The technology development goes on continually and is focused on preparing suggestions for new products and quotations for prospective customers. When there is a customer and a signed contract, product development commences. It serves to develop the proven concept into an actual product. This is done in cooperation with the customer and involves detailing the design and qualifying the new product. In Sections 3.1–3.4 below, the answers from all four interviewees have been summarized. Some clarifications have been added by the authors. 3.1. Technology development The R&T (research and technology) department works primarily with technology development. This is often carried out in cooperation with universities and research institutes. The drivers are mainly: (i) addressing the need of new and improved products and (ii) the necessity of bringing down the production cost. Addressing the first driver results in technology demonstrators that are run in test rigs so that improvements in performance such as reduced weight, noise, and fuel consumption can be demonstrated. The second driver leads to new manufacturing concepts. In the qualification phase of the technology development, they are passed through a validation program, assessing their performance so that they can be approved for use in actual manufacturing. The company needs to have a close contact with current and prospective customers to perceive the future needs. One important channel is the product launches that are made by the aircraft manufacturers at aviation exhibitions. Unlike product launches in other businesses, there are in general no actual products ready at the time of the launch. Only preliminary designs and specifications of the envisioned aircraft are shown. This is done to indicate what type of aircraft the manufacturer wants to have in its future product portfolio. After the launch, suppliers offer systems that they believe can fulfil the specifications. However, in practice, unofficial contacts with potential sub-suppliers are taken beforehand. This is a way for the aircraft manufacturers to get indications if it is feasible for the suppliers to design and manufacture the products that the manufacturers plan for the launch. The customers are not directly involved in the details of the manufacturing at the supplier. They want a low price, but they seldom specify how the component should be manufactured if it fulfils their specifications. In the company, it is the responsibility of the department's 'chief engineer's office' to first try to obtain beforehand information on the needs of the customer prior to the product launch and then instruct the R&T department to develop conceptual products accordingly. It is also the chief engineer's office that prepares the quotation and the contract with the aircraft manufacturers. The chief engineer's office also conducts the product development of the agreed product. When making the quotation, the specific fuel consumption is valuable information as well as the estimated cost and weight. In the quotation work, there are efforts made to minimize the cost. There is competition among suppliers for the contracts. Therefore, some risk is taken when making an offer. There is perhaps not a complete proof of concept before it is offered. In such a case, indications that it is possible to get the technology ready must be presented internally at the company before making the quotation. The expectation is to get the new technology validated as a part of the product development after the contract has been signed. When problems with products are discovered, root cause analyses are always performed. The results are documented in concept books, and records and lessons learned at the end of each project. There are also reports from the design verifications available. This documentation makes it possible to understand how future products can be improved. 3.2. Product development There are several types of product development projects. It may involve developing a completely new product or reusing a successful product by scaling it for a smaller or bigger aircraft. These projects are extensive, with a duration of several years and require certification of the final product. There are also projects involving updates of existing products. They are carried out in a shorter time. Still, multidisciplinary calculations may be needed. If it can be proved that no form, fit, or functions are affected then minor changes can be implemented within a few weeks to an existing product in production. To some extent, existing products are offered to several aircraft manufacturers. However, there are always some minor differences between the variants. At least, the attachments and the software will be unique for each product. The first priority in product development is to get a functioning engine. Simultaneously, efforts are made to reduce the weight and manufacturing cost. Function, weight, and cost are addressed in an iterative manner, where the refinement steps get smaller and smaller changes until the final design is reached. 3.3. Requirements management The process concerning the requirements specification is central and is therefore carefully managed. The requirements specification that is applicable in product development is obtained from the customer. The requirement specifications are extensive documents specifying requirements for engine performance, service intervals, and so on. It also specifies how to verify the compliance of the product with these requirements. From the requirement specifications obtained from the customer, an in-house requirement specification is made that includes internal manufacturing process requirements. The requirement specification is also adapted to different disciplines such as market and engineering. Some of the requirements are absolute, for example the certification requirements concerning strength and safety. In most cases, the expected lifetime versus weight and cost are negotiated with the customer. It is possible to make changes to the initial requirement specification provided that convincing arguments are presented. There are no special tools for requirements handling (except word and excel). The customer relations management is used for keeping track of some of the requirements such as when a customer requires a certain analysis to be performed in a certain way. There is a variation in the requirements during the development process, especially those that are related to loads and interfaces. Lately, there have been some changes in requirements related to the environmental legislations such as the European Union legislation for chemicals (REACH), stipulating the phasing out of certain materials and manufacturing processes. 3.4. Tools and systems The company uses many different types of software for simulation. These are used as standalone applications for interactive use, or as part of automated routines to make design studies involving many variants. Both the chief engineer's office and the R&T department use the same IT systems and tools. This creates a vast amount of information that needs to be stored in an indexed way allowing rapid information retrieval. Work is going on to use the same product lifecycle management system throughout. SharePoint is used for information sharing in the non-formalized part of the development work. The company is striving towards a uniform catalogue structure in all projects to facilitate information retrieval and sharing between projects. All information sharing between projects is subject to prior assessment to avoid violation of confidentiality agreements. The aforementioned environment used for making explorative design studies incorporate many of the IT systems and software of the company such as CAD, finite element analysis (FEA), and computational fluid dynamics (CFD). These have been connected using scripting into an automated environment. After the creation of a concept design CAD model, many geometrically different variants of it are created automatically according to a design of experiment, varying 5–20 geometrical design parameters. This results in 100 or more different variants that are sent to software that analyse them from several aspects such as aerodynamic, structural, and thermal. These are called design cases (DC). Figure 2 below shows this procedure. Figure 2: Open in new tabDownload slide Automated design exploration. Adapted from Hjertberg, Stolt, and Elgh (2018). Figure 2: Open in new tabDownload slide Automated design exploration. Adapted from Hjertberg, Stolt, and Elgh (2018). The results are then visualized and reviewed. This works as a decision support to find regions of the design space that are interesting for further elaboration. The design's response to variation is explored allowing the design parameters to be set so that the customer expectations are met. Manufacturability is part of these analyses as seen in Fig. 2. This is elaborated in Stolt et al. (2017) and Stolt, Elgh, and Andersson (2017a). The results of the analyses are visualized graphically. This leads to a knowledge build-up on the behaviour of the conceptual design. The whole process of analysing the concept is automated using in-house developed scripts and software that are integrated via the application programmable interfaces of the different software. Automation is necessary since there are often 100 or more different variants to be analysed from many different aspects. Efforts have been made to keep the time required to analyse each variant low, so that the whole process can be completed within a week or so. When developing the product for the customer in the product development stage, the same environment is used. This time the variation is smaller since it was found roughly how to set the design parameters in the technology development phase. In the product development phase, more fine tuning is done, increasing the precision of the analyses. The environment is also useful when requests for changes are made by customers. What-if analyses can be made quickly, increasing the ability of the company to react to fluctuating requirements. The company uses design practices that are textual documents to convey a preferred way of conducting the development work. They specify, for example, when and what types of analysis should be carried out and how the results should be interpreted. The design practices make it easy to get an overview, but they lack detail, so it can be difficult to understand exactly how to perform the work from the design practices. The sources of information in the product development process are web pages, both internal and external along with the product data management system and databases for patents and materials. Some of the systems are not integrated. It would be useful with a connection between the materials database and the FEA system allowing material properties to be directly used when making FEA. After each project, everything is documented in concept books and records and lessons are learned. In addition, there are reports from the design verifications. This information is found in the product lifecycle management system. 4. Integrating SLM in Technology and Product Development The interviews provided an insight in the development process of the company. The designers working both in technology and product development need to learn more about the SLM process and its capabilities and constraints to come up with designs that can be fully or partly manufactured by SLM. When doing modifications of existing engine components, it will not be possible to simply replace them with SLM parts since their geometries are not suited for SLM. Redesigning is required. This must be followed by a certification of the redesigned engine, requiring methods and tools that are not yet available. SLM parts will therefore only emerge in future engine concepts. Currently, it is in the technology development phase that the efforts are being made for exploring the potential of SLM. New concepts could include the consolidation of several parts into one since SLM has few geometric restrictions. Another expectation is that the preparation for production will be shorter for SLM parts (no moulds or fixtures). One can assume that this would allow the company to quickly respond to fluctuating requirements from customers, because the impact on the manufacturing process will be easier to assess. There are also sustainability issues that possibly can be addressed when manufacturing with SLM. However a thorough analysis of the lifecycle of the alternatives will have to be made to make any definitive predictions. 4.1. Design space exploration As elaborated in the literature section, TO is an integral part of DFAM and must therefore be integrated into the environment. The main steps in the current procedure are shown in Fig. 3. From a conceptual design, a variety of design cases are generated, e.g. DC1–DCn in Fig. 3. These are represented as CAD models and a variety of analyses are performed on them including FEA, CFD, and other types of calculations so that the different design cases can be compared according to their performance from different aspects. Here, interesting regions of the design space are identified for further refinement. When SLM and TO come into the picture, it is not possible to let all the various disciplines control the TO. It will be too complex. Instead, a single objective function such as minimizing the strain energy should be used. Figure 3: Open in new tabDownload slide Current automated design space exploration. Figure 3: Open in new tabDownload slide Current automated design space exploration. Having a single objective function means that it will be necessary to have a separate step for TO in the environment. The TO can as a suggestion involve minimizing the strain energy with the constraint such that the volume fraction of part should be kept at a predetermined level. This is standard in TO and readily available in current commercial FEA software. The added step is seen in Fig. 4. Compare Figs 3 and 4. The TO and lattice step will alone determine the location and density of the lattice structure. It is also possible in current software to include some SLM constraints such as the exclusion of downward facing surfaces to increase the SLM printability. Hence, there are 'constraints' in the TO and lattice step of Fig. 4. In addition to the manufacturing constraints, it should also be a possibility to exclude regions in the geometry that for various reasons should not be affected by the TO. These are, for instance, surfaces in contact with the airflow that should not have exposed lattice structures for aerodynamical reasons. The result of the TO is a density distribution of the elements, that in turn should determine the parameters in the CAD models representing each of the design cases. These have been named DC/TO as seen in Fig. 4. Figure 4: Open in new tabDownload slide Proposed integration of TO in the explorative design evaluation. Figure 4: Open in new tabDownload slide Proposed integration of TO in the explorative design evaluation. These CAD models are subject to the following multidisciplinary analyses at the company, as seen in the multidisciplinary design case evaluation step of Fig. 4. This part is already established at the company for casting and fabrication processes. It is currently not known how to use the FEA meshes of each design case to govern the CAD models. A design engineer can of course make an interpretation of the FEA mesh and adjust the CAD model accordingly. However, considering the number of designs cases that are evaluated, this would take very long time. An automated method is needed. Figure 5 shows a single design case. From the design case CAD model, an FEA mesh is generated. This is seen to the left in Fig. 5. A TO (SIMP) has been performed on it as shown in the middle picture. It indicates areas that have strain energy levels allowing the density to be decreased. The missing elements have low density and have therefore been penalized and excluded by the SIMP algorithm. In these regions, a lattice is inserted as seen in the rightmost third picture. Here, the densities from the calculations have been assigned to the lattice structures of the parametric CAD model. This has been done manually in the Fig. 5 example. Figure 5: Open in new tabDownload slide Mapping the TO density distribution to lattice features in the CAD model. Figure 5: Open in new tabDownload slide Mapping the TO density distribution to lattice features in the CAD model. 5. Discussion This paper has investigated the expected impact of the introduction of SLM on the technology and product development processes in an aerospace company. The purpose has been 2-fold: first, finding the state of practice on product and technology development and then using this knowledge to try to predict the consequences on technology and product development of introducing SLM. It must be understood that SLM was not a part of the interviews. They were conducted to understand the technology and product development processes in the company. Additional insights, especially concerning the environment for automated design case evaluation, have been made by the authors working at the company with the environment and with predicting component producibility using other production processes such as welding (Pabolu, Stolt, & Johansson, 2017; Stolt, André et al., 2017) and inspection (Stolt, Elgh, & Andersson, 2017b). There have been frequent interactions with company staff in several research projects. Through these interactions, it has been found that designs for the SLM process are currently being considered in the technology development phase. However, important pieces are still missing such as methods for verifying SLM components against aviation regulations. Although the environment for design evaluation has a role in product and technology development, it is not completely central. A lot of the development work is conducted outside this environment. It should be regarded as an assistive tool for exploring the design space. One of the most important findings in the paper is that introducing TO in the environment will require a separate step. This is because bringing in all the aspects – structural, aerodynamic, and so on − into the TO will not be possible. Only finding an objective function to weigh in all these aspects will probably be insurmountable for the time being. The consequence of this division is that there will be more iterations. When the TO geometry is evaluated against all other criteria, the geometry will possibly have to be revised and re-evaluated. If all aspects could be incorporated in one step, a more ready design could be found with less iterations. The example shown in Fig. 5 is a simplified version of a product intended for sheet metal and welding. This is most likely not what future components for SLM manufacture will look like. However, lattice will probably be part of the future SLM components considering the high importance of bringing down the amount of material used to a minimum from both manufacturing and functional points of view. Another development will perhaps be conformal cooling channels inside the components to achieve a better temperature distribution and thereby reducing thermal stresses. An important observation made in the paper is that it will be a challenge to make an automated interpretation of the TO results into parametric CAD models. In the example, an interpretation was made from the TO mesh for where to place the lattice structures in the CAD model and what their density should be. This was a lengthy and subjective process. In the actual environment, considering the high number of design cases that are evaluated, it remains to find an automated rapid way of making parametric CAD models from the TO meshes. As for DFAM, there are more manufacturing issues than just introducing lattice. The design of the parts needs to consider the orientation in the printer and the amount and removal of the support material. There are also extensive dependencies between the manufacturing and the structural performance since printed parts are anisotropic, meaning the structures can be optimized with respect to the expected outcome of the SLM process in terms of material properties. This adds to the motivation for including DFAM simultaneously in a single optimization loop together with all other aspects. 6. Conclusions and Future Work This paper has contributed to the understanding of the technology and product development processes within an aerospace company. The development of SLM components is currently addressed in the early stages of technology development and the qualification of actual SLM products in actual products is still years into the future. There is a high degree of dependency between the SLM manufacturing process and the resulting properties of the part. That is why, ideally all requirements should be evaluated simultaneously. In future work, the criteria allowing the automated evaluation of SLM manufacturability need to be developed. Footnotes ☆ Selected paper from the 25th ISTE International Conference on Transdisciplinary Engineering, 3–6 July 2018, Modena, Italy. References Almefelt L. , Berglund F., Nilsson P., Malmqvist J. ( 2006 ). Requirements management in practice: Findings from an empirical study in the automotive industry . Research in Engineering Design , 17 ( 3 ), 113 – 134 . OpenURL Placeholder Text WorldCat Arisoy E. B. , Mirabella L., Slavin E. ( 2015 ). Design and topology optimization of lattice structures using deformable implicit surfaces for additive manufacturing , In Proceedings of the ASME 2015 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2015, 2–5 August, 2015 Boston, United States. 4 , (pp. 1 – 11) . OpenURL Placeholder Text WorldCat Booth J. W. , Alperovich J., Chawla P., Ma J., Reid T. N., Ramani K. ( 2017 ). The design for additive manufacturing worksheet . Journal of Mechanical Design , 139 ( 10 ), 1 – 9 . OpenURL Placeholder Text WorldCat Brackett D. , Ashcroft I., Hague R. ( 2011 ). Topology optimization for additive manufacturing , In Proceedings of the 22nd Annual International Solid Freeform Fabrication Symposium - An Additive Manufacturing Conference, SFF 2011, 8–10 August, 2011, Austin, TX; United States . (pp. 348 – 362) . OpenURL Placeholder Text WorldCat Challis V. J. , Xu X., Zhang L. C., Roberts A. P., Grotowski J. F., Sercombe T. B. ( 2014 ). High specific strength and stiffness structures produced using selective laser melting . Materials and Design , 63 , 783 – 788 . OpenURL Placeholder Text WorldCat Cheng L. , Zhang P., Biyikli E., Bai J., Robbins J., To A. ( 2017 ). Efficient design optimization of variable-density cellular structures for additive manufacturing: Theory and experimental validation . Rapid Prototyping Journal , 23 ( 4 ), 660 – 677 . OpenURL Placeholder Text WorldCat Diegel O. ( 2019 ). A Practical Guide to Design for Additive Manufacturing . Singapore : Springer . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hällgren S. , Pejryd L., Ekengren J. ( 2016 ). (Re)Design for additive manufacturing . Procedia CIRP , 50 , 246 – 251 . OpenURL Placeholder Text WorldCat Heikkinen T. , Stolt R., Elgh F., Andersson P. ( 2016 ). Automated Producibility Assessment Enabling Set-Based Concurrent Engineering . Transdisciplinary Engineering: Crossing Boundaries TE2016 , Curitiba, Brazil : IOS Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hjertberg T. , Stolt R., Elgh F. ( 2018 ). A tool for obtaining transparency and traceability in heterogeneous design automation environments . Computer-Aided Design and Applications , 15 ( 4 ), 1 – 13 . OpenURL Placeholder Text WorldCat Klahn C. , Leutenecker B., Meboldt M. ( 2015 ). Design strategies for the process of additive manufacturing . Procedia CIRP , 36 , 230 – 235 . OpenURL Placeholder Text WorldCat Leutenecker-Twelsiek B. , Klahn C., Meboldt M. ( 2016 ). Considering part orientation in design for additive manufacturing . Procedia CIRP , 50 , 408 – 413 . OpenURL Placeholder Text WorldCat Li D. , Liao W., Dai N., Dong G., Tang Y., Xie Y. M. ( 2018 ). Optimal design and modeling of gyroid-based functionally graded cellular structures for additive manufacturing . CAD Computer Aided Design , 104 , 87 – 99 . OpenURL Placeholder Text WorldCat Liu J. , Gaynor A. T., Chen S., Kang Z., Suresh K., Takezawa A., Li L., Kato J., Tang J., Wang C. C. L., Cheng L., Liang X., To A. C. ( 2018 ). Current and future trends in topology optimization for additive manufacturing . Structural and Multidisciplinary Optimization , 57 ( 6 ), 2457 – 2483 . OpenURL Placeholder Text WorldCat Pabolu V. K. R. , Stolt R., Johansson J. ( 2017 ). Flexible manufacturability analysis applied to the welding process to increase sustainability in engineer to order businesses . International Journal of Agile Systems and Management , 10 ( 3/4 ), 271 – 294 . OpenURL Placeholder Text WorldCat Panesar A. , Abdi M., Hickman D., Ashcroft I. ( 2018 ). Strategies for functionally graded lattice structures derived using topology optimisation for additive manufacturing . Additive Manufacturing , 19 , 81 – 94 . OpenURL Placeholder Text WorldCat Rosen D. W. ( 2014 ). Research supporting principles for design for additive manufacturing: This paper provides a comprehensive review on current design principles and strategies for AM . Virtual and Physical Prototyping , 9 ( 4 ), 225 – 232 . OpenURL Placeholder Text WorldCat Rosen D. W. ( 2016 ). A review of synthesis methods for additive manufacturing . Virtual and Physical Prototyping , 11 ( 4 ), 305 – 317 . OpenURL Placeholder Text WorldCat Sercombe T. B. , Xu X., Challis V. J., Green R., Yue S., Zhang Z., Lee P. D. ( 2015 ). Failure modes in high strength and stiffness to weight scaffolds produced by selective laser melting . Materials and Design , 67 , 501 – 508 . OpenURL Placeholder Text WorldCat Stolt R. , André S., Elgh F., Andersson P. ( 2017 ). Introducing welding manufacturability in a multidisciplinary platform for the evaluation of conceptual aircraft engine components . International Journal of Product Lifecycle Management , 10 ( 2 ), 107 – 123 . OpenURL Placeholder Text WorldCat Stolt R. , Elgh F., Andersson P. ( 2017a ). Design for inspection – evaluating the inspectability of aerospace components in the early stages of design . Procedia Manufacturing , 11 , 1193 – 1199 . OpenURL Placeholder Text WorldCat Stolt R. , Elgh F., Andersson P. ( 2017b ). Design for inspection: Evaluating the inspectability of aerospace components in the early stages of design . In Proceedings of the 27th International Conference on Flexible Automation and Intelligent Manufacturing, FAIM2017, 27–30 June, 2017, Modena, Italy . 11 , ( pp. 1193 – 1199 .). Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Stolt R. , Heikkinen T., Elgh F. ( 2018 ). Integrating additive manufacturing in the design of aerospace components . In Proceedings of the 25th ISPE International Conference on Transdisciplinary EngineeringIntegrating (TE2018), 3–6 July, Modena, Italy . Margherita Peruzzini, IOS Press , (pp. 145 – 154 .). Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Stolt R. , Johansson J., André S., Heikkinen T., Elgh F. ( 2016 ). How to challenge fluctuating requirements: Results from three companies . In Proceedings of the 23rd ISPE Inc. International Conference on Transdisciplinary Engineering, Parana, Curitiba, October 3–7, 2016 . Transdisciplinary Engineering: Crossing Boundaries, IOS Press , ( pp. 1061–1070 ). Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Suzuki K. ( 1991 ). A homogenization method for shape and topology optimization . Computer Methods in Applied Mechanics , 93 ( 3 ), 291 – 318 . OpenURL Placeholder Text WorldCat Takezawa A. , Yonekura K., Koizumi Y., Zhang X., Kitamura M. ( 2018 ). Isotropic Ti–6Al–4V lattice via topology optimization and electron-beam melting . Additive Manufacturing , 22 , 634 – 642 . OpenURL Placeholder Text WorldCat Thompson M. K. , Moroni G., Vaneker T., Fadel G., Campbell R. I., Gibson I., Bernard A., Schulz J., Graf P., Ahuja B., Martina F. ( 2016 ). Design for additive manufacturing: Trends, opportunities, considerations, and constraints . CIRP Annals - Manufacturing Technology , 65 ( 2 ), 737 – 760 . OpenURL Placeholder Text WorldCat Verbart A. , Langelaar M., van Dijk N., Van Keulen F. ( 2012 ). Level set based topology optimization with stress constraints and consistent sensitivity analysis . In Proceedings of the 53rd AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference 20th AIAA/ASME/AHS Adaptive Structures Conference 14th AIAA(April) . (pp. 1 – 15 .). Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Wang M. Y. , Wang X., Guo D. ( 2003 ). A level set method for structural topology optimization . Computer Methods in Applied Mechanics and Engineering , 192 ( 1–2 ), 227 – 246 . OpenURL Placeholder Text WorldCat Wang X. , Xu S., Zhou S., Xu W., Leary M., Choong P., Qian M., Brandt M., Xie Y. M. ( 2016 ). Topological design and additive manufacturing of porous metals for bone scaffolds and orthopaedic implants: A review . Biomaterials , 83 , 127 – 141 . OpenURL Placeholder Text WorldCat Xia L. , Xia Q., Huang X., Xie Y. M. ( 2018 ). Bi-directional evolutionary structural optimization on advanced structures and materials: A comprehensive review . Archives of Computational Methods in Engineering , 25 ( 2 ), 437 – 478 . OpenURL Placeholder Text WorldCat Yao X. , Moon S. K., Bi G. ( 2017 ). Multidisciplinary design optimization to identify additive manufacturing resources in customized product development . Journal of Computational Design and Engineering , 4 ( 2 ), 131 – 142 . OpenURL Placeholder Text WorldCat The Author(s) 2020. Published by Oxford University Press on behalf of the Society for Computational Design and Engineering. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
TI - Introducing design for selective laser melting in aerospace industry
JF - Journal of Computational Design and Engineering
DO - 10.1093/jcde/qwaa042
DA - 2020-08-01
UR - https://www.deepdyve.com/lp/oxford-university-press/introducing-design-for-selective-laser-melting-in-aerospace-industry-ozGPW4HARS
SP - 489
EP - 497
VL - 7
IS - 4
DP - DeepDyve
ER -