Design and manufacturing are intimately coupled in the conceptualization and realization of products. Engineering design is regularly a key component of innovation and successful product development, yet promising designs that are very costly or challenging to manufacture may never be realized in practice. Advancement of manufacturing technologies tends to lower barriers to realization and expands design freedom but may also introduce unique restrictions on the design space. A clear example of this connection is the maturation of additive manufacturing technologies, which has created immense design freedom for both researchers and practitioners to explore. New capabilities, such as the ability to fabricate complex shapes, tailor materials and properties, and provide complex functionalities, present a whole new paradigm and a wider range of opportunities for intricately design, highly functional, cost-effective, high-value products. However, additive manufacturing technologies also bring new types of constraints to be carefully considered in design. New design methods and tools are therefore needed to help engineers navigate both the novel design space created by advanced manufacturing and the new manufacturing constraints. This special issue aims to consolidate some of the most recent design research in support of advanced manufacturing.
During the conception of this special issue, we examined the connection of design research with the advancement of manufacturing technology. We identified several general design research topics that are often influenced by the capabilities of manufacturing technologies, including design methods, geometric modeling techniques, uncertainty analysis, and design education. Each of these topics can be further divided into more specific research domains. For example, design methods for advanced manufacturing include design for additive manufacturing, design optimization to support advanced manufacturing, design principles, and guidelines for advanced manufacturing.
We had hoped to attract manuscripts that encompass all types of advanced manufacturing technologies such as hybrid manufacturing processes, high-precision CNC processes, improved injection molding, and advanced sheet metal forming processes. Nevertheless, the diverse body of articles that comprise this special issue largely focuses on additive manufacturing technology. This may reflect the reality that design for additive manufacturing continues to have many open research challenges, attracting significant research interest. This special issue comprises 12 papers that can be divided into five categories: (1) lattice structure design considering manufacturability by additive manufacturing, (2) design considering material behavior associated with manufacturing processes, (3) design theory and methodology considering manufacturability and fixation, (4) design of manufacturing processes, and (5) uncertainty in design and manufacturing. These categories represent the breadth of emerging techniques employed in design for advanced manufacturing. For example, AI techniques, such as generative models, have been used to cover high-dimensional design representations, such as lattice geometries or processing sequences. Since the use of physics-based processes or product models is critical while the computational demand is high, a few accepted papers utilize multi-fidelity modeling to achieve better process and material behavior predictions. Design heuristics and design ideation are important research aspects to understand manufacturing constraints on design consideration and fixation. Finally, the advancement of additive manufacturing continuously promotes research in shape and topology optimization and other structural optimization techniques. We summarize the submissions to this issue under the aforementioned categories.
Lattice Structure Design Considering Manufacturability of Additive Manufacturing
In the paper titled “Neural Network-Assisted Design: A Study of Multiscale Topology Optimization With Smoothly Graded Cellular Structures,” Rastegarzadeh, Wang, and Huang use a neural network as an optimizer in the topology optimization problem. This approach reparametrizes the density field into a function-representation-based microstructure. The volume and shape of the microstructure are controlled by a level surface which serves as design representation. The multiscale topology optimization problem is reformulated with this level surface and a neural network is used to map the input spatial coordinates onto the level surface. The neural network parameters are optimized through backpropagation that leads to optimized designs. The design approach is adapted to be able to address the issue of sharp transitions between adjacent cellular structures in multi-scale design. Experimental results demonstrate that this approach outperforms previously reported works. Future research discussed in this work includes integrating the finite element analysis solver with the neural network to make the design process more streamlined.
Letov and Zhao in the paper titled “Beam-Based Lattice Topology Transition With Function Representation” present a function representation approach for the geometric modeling of beam-based heterogeneous lattice structures. This research is a further development of its previously reported function representation of general heterogeneous lattice structures by defining a set of geometric parameters to achieve lattice topology transition (i.e., the continuous connection of unit cells with different topologies). One of the key geometric parameters is the transition plane, which is defined as a function. Once the design engineer identifies where the topology transition should happen, the parameters of this transition plane are set to support the connection of chosen different topologies. The function representations of some beam-based topologies have an optional truncation parameter that is needed to fully define the skeletal graph. This paper further explores using truncation as a controllable parameter to support topology transitions. These topology transition techniques are implemented in a lattice geometric modeling software. The authors believe this approach has the potential to be integrated with manufacturability analysis to improve the manufacturability of the designed heterogeneous beam-based lattice structures.
In the paper “G-Lattices: A Novel Lattice Structure and Its Generative Synthesis Under Additive Manufacturing Constraints,” Armanfar and Gunpinar present a novel strut-based lattice structure called G-Lattices and a generative synthesis method incorporating manufacturing constraints. G-Lattices consist of straight and curved struts. This type of lattice can be generated by a particle-tracing algorithm. Two synthesis algorithms are presented to generate straight-strut and curved-strut G-Lattices, respectively. These algorithms are used with different parameter settings to generate multiple G-Lattices. Further development of the algorithms is achieved by allowing customization of the lattice structure to improve mechanical performance under vertical loading. The designed G-Lattices are fabricated using the fused-deposition modeling technique. Results demonstrate that this type of lattice often is self-standing and self-supporting, which eliminates some manufacturing constraints.
Design Considering Material Behavior Associated With Manufacturing Process
In the paper titled “Unit-Based Design of Cross-Flow Heat Exchangers for LPBF Additive Manufacturing,” Liang et al. developed a unit-based design framework to optimize the channel configuration for cross-flow heat exchangers to improve the heat dissipation performance while controlling the pressure drop between the fluid inlet and outlet. A gradient-based optimization method is used to drive the design process, which results in the change of shape of the fluid channel. In the design framework, 2D designs are extruded and manufacturability is considered with respect to the laser powder bed fusion process. A rounding technique in the re-design process is presented to improve the manufacturability of the final design. Numerical simulation is subsequently carried out after the design to predict the residual stress and deformation. The effect of build orientation is studied, and a 45 deg orientation is identified to be the preferred build orientation for fabricating the presented cross-flow heat exchanger designs.
In the paper titled “Shape Optimization for the Temperature Regulation in Extrusion Dies Using Microstructures,” Zwar et al. present a novel approach to design a lattice infill for an extrusion-die design with the goal of optimizing the temperature profile of plastic melt along the flow-channel wall and hence prevent excessive deformations and residual stresses in the extruded component. The infill consists of a morphed lattice with a simple-cubic unit cell (which the authors call a microtile) that occupies the volume between the extrusion die wall and the flow channel. The spatial distribution of the microtile strut-thickness and the deformation of a straight lattice to produce a morphed infill that conforms to the die cavity are represented as fields that are parameterized using B-splines. Taking advantage of the low dimension of the design representation, the optimization is performed with a gradient-less algorithm. The microtiles are thus arranged in a way that, even with homogeneous heating via a heating band, an inhomogeneous temperature distribution can be achieved within the flow channel by passive heat transfer regulation based on the characteristics of the die. Two case studies are carried out to demonstrate the effectiveness of this approach and validate it in an industrial setting.
Design Theory and Methodology Considering Manufacturability and Its Fixation
Trautschold and Dong in the paper titled “Additive Manufacturability Analysis of Multiscale Aperiodic Structures: A Statistical Mechanics Approach” present heuristics that are based upon statistical mechanics to assist the manufacturability analysis of additive manufacturing processes. Specifically, this approach focuses on multiscale aperiodic structures. The heuristics, which can associate structural properties at statistical level with manufacturability, are derived from four topological properties of the complex network representations of multiscale aperiodic geometry. The effectiveness of the heuristics is evaluated via a cross-model validation and external validation. The results demonstrate that the heuristics are effective and can provide insight into the additive manufacturability of multiscale aperiodic structures.
In the paper titled “Manufacturing Fixation in Design: Exploring the Effects of Manufacturing Fixation During Idea Generation,” Brennan et al. carried out a workshop-based study to explore the impact of fixation on the use of advanced manufacturing technologies. A particular type of design fixation called manufacturing fixation in design (MFD) is identified to represent an unconscious and often unintentional adherence to a limited set of manufacturing processes and/or constraints and capabilities during the design ideation process. The stated next steps in this research are to develop measurement techniques of MFD and explore mitigation techniques.
Design of Manufacturing Processes
In the paper titled “Decentralized and Centralized Planning for Multi-Robot Additive Manufacturing,” Poudel et al. present a decentralized scheduling technique for multi-robot corporative additive manufacturing processes. Since many decisions in manufacturing process planning are dictated by design features, this paper focuses on the process planning aspects of additive manufacturing processes. The decentralized approach is compared with the centralized approach through two case studies. The results show that the centralized approach provides a better solution compared to the decentralized approach for small-scale design or small-scale manufacturing setups, while in large-scale design or manufacturing setups, the decentralized approach was shown to outperform the centralized approach for the considered cases.
Fillingim and Fu in the paper titled “Framework for the Evolution of Heuristics in Advanced Manufacturing” present a study to understand how heuristics are developed, retrieved, employed, and modified by designers. A unique process is used to extract the heuristics and their characteristics for both designers and operators of a hybrid manufacturing setting. Eight participants performed a series of two design journals, two interviews, and one survey. The results show statistically significant correlations between heuristic reliability, evolution, and frequency of use. The future development of the heuristics collected in this research depends on machine quality and current technology levels, experience of the user of these heuristics, or the objectives for the use of their respective manufacturing machines.
Uncertainty in Design and Manufacturing
Chen et al., in the paper titled “GAN-DUF: Hierarchical Deep Generative Models for Design Under Free-Form Geometric Uncertainty” present a generative adversarial network-based design under uncertainty framework (GAN-DUF) that uses a deep generative model to simultaneously learn a compact representation of ideal designs and the conditional distribution of fabricated designs. Two case studies are carried out, an airfoil design and an optical metasurface absorber design, to evaluate the effectiveness of the GAN-DUF approach. The studies demonstrate that the design solutions found by this novel approach are more likely to perform better after manufacturing. Further opportunities of the GAN-DUF approach include building a universal uncertainty quantification model compatible with both shape and topological designs, modeling freeform geometric uncertainties without the need to make any assumptions on the distribution of geometric variability and allowing fast prediction of uncertainties for new nominal designs.
In the paper titled “Iterative Uncertainty Calibration for Modeling Metal Additive Manufacturing Processes Using Statistical Moment-Based Metric,” Dehaghani, Tang, and Wang present an improvement of a previously reported calibration and validation framework for metal additive manufacturing processes that uses both physics-based models and a multi-fidelity model to predict final properties of the fabricated component. The improvement detailed in the paper is an interactive calibration process that uses the second-order statistical moment-based metric. This method is evaluated in a four-variable porosity modeling case study and the results show improved accuracy.
Deng et al. in the paper titled “Data-Driven Calibration of Multi-Fidelity Multiscale Fracture Models via Latent Map Gaussian Process,” present a data-driven framework that integrates a mechanistic reduced-order model (ROM) with a calibration scheme based on random processes to support fracture-aware design of multi-scale materials. The ROM model significantly accelerates direct numerical simulations by using a stabilized damage algorithm and systematically reducing the degrees-of-freedom through clustering. The ROM is calibrated by constructing a multi-fidelity random process based on latent-map Gaussian processes. The results show the validity of the approach in predicting the damage behavior of a multiscale metallic component with spatially varying porosity. The microstructure porosity is found to have a significant effect on the performance of macro components and that this feature must be considered in design process.
The papers in this Special Issue encompass a broad range of topics related to design for advanced manufacturing. This suggests that many open research questions remain on the topic of advanced manufacturing, such as additive manufacturing, as related to design methods, design processes, design representation, and the treatment of manufacturing constraints and uncertainties in design. Although additive manufacturing ended up being the focus of the submissions for this Special Issue, many of these same research challenges apply in the context of other advanced manufacturing process. It is our hope that the contents of this Special Issue will continue to stimulate new advances in design for advanced manufacturing.