Research paperTwo-scale concurrent topology optimization of lattice structures with connectable microstructures
Introduction
Two-scale structures composed of lattice materials have great potentials in achieving lightweight property and desirable mechanical/multi-physics performances, such as high specific stiffness [1], energy absorption [2], and heat exchange capability [3]. For instance, lattice structures with periodic unit cells can provide both load-bearing and active cooling functionalities [4]; biodegradable scaffolds composed of lattice materials are required to possess certain porosity for nutrient delivery while still maintaining reasonable stiffness to support the tissues [5,6]. Particularly, the geometry and spatial distribution of the microstructures, and the topology of the global structure are crucial to the overall structural performance. However, the systematic design of such multiscale lattice structures is still a challenging task because of their high geometrical complexity.
Topology optimization [[7], [8], [9], [10], [11]] is recognized as a highly effective method to obtain the optimal structural layouts with desired performances. Extensive researches on topology optimization-based lattice microstructural design have been carried out. For instance, Sigmund and Torquato [12] achieved material designs with positive, zero and negative thermal expansion with multiple base materials. Larsen et al. [13] designed and fabricated compliant micromechanisms and materials with a negative Poisson ratio. Sigmund [14] optimized material microstructures to obtain prescribed mechanical constitutive parameters. Guest and Prévost [15] achieved multifunctional material design with maximum stiffness and fluid permeability. Kazemi et al. [16] studied the design of the multi-material truss lattice materials.
A hot topic that has attracted increasing attention is to use topology optimization to design high-performance multiscale lattice structures. Therein, both the macroscale and microscale geometries are to be simultaneously optimized under the assumption of scale separation. Rodrigues et al. [17] proposed a hierarchical optimization procedure to iteratively design macroscale structure with point-wise different microstructures. Liu et al. [18] developed a so-called Porous Anisotropic Material Penalization (PAMP) method to interpolate anisotropic effective properties in the stiffness maximization problem. The PAMP method can be used to design the distribution of structured materials. It decouples the topology description of the macroscale structure and the microstructure and enables simultaneous optimization of the material distribution and the microstructural topologies. On the basis of nonlinear multiscale analysis, Xia and Breitkopf [19] proposed an optimization scheme to obtain microstructure topologies. Design schemes concerning the grouping of lattice materials have been studied by Li et al. [20] and Xu and Cheng [21]. Besides the studies based on the scale separation assumption, some works also dealt with lattice structure design problems accounting for the scale effects [[22], [23], [24], [25]]. Numerical results have proven the performance benefits earned by the optimized multiscale structures.
Because of the rapid advance of additive manufacturing techniques [26,27], novel microstructure configurations with high geometric complexity can be readily realized [[28], [29], [30], [31]]. Though certain manufacturing limits such as the overhang constraints exist (see e.g. Gaynor and Guest [32], Langelaar [33], Liu and To [34]. and Wang et al. [35]), the combination of topology optimization and additive manufacturing [36,37] can greatly enlarge the design space of structures. Topology methods utilizing B-splines (see e.g [38,39]) have been proposed, which can effectively ease the difficulties when rebuilding the CAD models of the optimized designs. However, the above-mentioned inverse homogenization-based approaches may yield optimized microstructural unit cells that are not connectable, as shown in Fig. 1. Though such designs provide meaningful material configurations of a lattice structure, they cannot be manufactured in practice.
There have been a number of studies to resolve the microstructure connectivity issues when designing structures composed of graded lattice materials. Zhou and Li [40] compared three schemes to enforce the connectivity of microstructures. One of these schemes that is convenient to implement is to set up non-design domains in the microstructures. However, designating such non-design domains may pose a limitation to the design space and thus lead to sub-optimal microstructural geometries. Rodman et al. [41] proposed a progressive optimization scheme that optimizes three base unit cells at a stage to keep the connectivity of microstructures. Garner et al. [42] considered the assembly of adjacent cells along with the optimization of individual cells to enforce the microstructure connectivity, which may demand a large computational cost in 3D cases. Besides the above-mentioned methods, some constraints on the design variables have also been proposed to achieve microstructure connectivity: Alexandersen and Lazarov [22] imposed constraints on PDE-filtered densities on the boundaries between layers of graded lattice materials to generate connectable microstructures; Du et al. [43] proposed a so-called “connectivity index” constraint on the basis of the level set method for fixed distributions of lattice materials. For such methods, a large number of constraints, as well as their design sensitivities, need to be evaluated if the number of considered lattice materials is large. Design methods on the basis of metamorphosis techniques to enforce the microstructure connectivity have also investigated. For instance, Wang et al. [44] proposed an optimization framework to obtain a series of connectable microstructures with similar topologies from a prototype. Zhou et al. [45] presented a post-processing scheme to enforce microstructure connectivity, which may introduce a performance deviation between the optimized design and the post-processed one.
Besides, some works based on the so-called de-homogenization concept aim to achieve near-optimal stiffness of lattice structures without restricting the repeatability of unit cells. Pantz and Trabelsi [46,47] and Allaire et al. [48] projected the designs obtained with the homogenization method to connectable solid/void designs with a finite length scale. Groen et al. [49,50] and Groen and Sigmund [51] simplified and further improved the method [46,47] and implemented control of the size and shape of the projected designs. With their method, connectable lattice designs with near-optimal stiffness can be obtained with relatively high efficiency, on the basis of the homogenization results and a postprocessing step to project them onto a finer mesh. The optimized designs obtained with these de-homogenization approaches usually have spatially varying porosities ranging from 0 to 1.
The present study focuses on the microstructure connectivity in the two-scale concurrent topology optimization of structures with multiple lattice materials. Specifically, we consider lattice materials that are composed of repetitive unit cells with specified porosity (which is beneficial for possible applications with multi-physics requirements such as active cooling). In this context, as the distribution of the lattice materials on the macroscale is not known a priori but is updated during the optimization process, the locations (i.e. the macroscale multi-material interfaces) to enforce microstructure connectivity keep changing. To resolve the microstructure connectivity issue under this circumstance, this paper proposes a designable connective region method, which ensures the microstructural connectivity by keeping the connective regions in all types of microstructures sharing a common topology. This is conveniently achieved through microscale design variable linking and avoids imposing any connectivity constraints. Therein, the microstructure topologies are described with the Solid Isotropic Material with Penalization (SIMP) model [52], and their effective material properties are computed with the asymptotic homogenization method. The macroscale material distribution is described with the Discrete Material Optimization (DMO) method [53]. For both the macroscale and microscale design problems, the density filter technique [54] is adopted to avoid the checkerboard and other numerical instability in the optimization process. The sensitivities of the objective function with respect to the design variables are derived using the adjoint method, and the optimization problem is solved with the mathematical programming method MMA [55].
The present paper is organized as follows: first, the proposed microstructure connectivity method is introduced in Section 2. Then the asymptotic homogenization method employed to compute the effective material properties, the topology optimization formulation, and sensitivity analysis procedures are discussed in Section 3. Numerical examples are given in Section 4, which is followed by conclusions.
Section snippets
Designable connective region method for microstructure design
In this section, based on the density model, we present a designable connective region method to ensure the connectivity of the optimized lattice material microstructures with prescribed porosity in a two-scale concurrent design framework. A structure that is composed of lattice materials is considered. Here, sets of macroscale design variables each with a dimension of (total number of designable elements in the macrostructure) are defined to describe the
Homogenization method for evaluating effective properties of lattice microstructures
In the present study, the separation of scales is assumed, i.e. the sizes of the microstructures are assumed to be much smaller in comparison with that of the macrostructure. Considering the periodic arrangement of microstructures within each lattice material, the effective mechanical properties are computed with the homogenization method. This method is based on the asymptotic expansion of the macroscale quantities. For the sake of completeness, a brief introduction of the homogenization
Numerical examples
In this section, the proposed designable connectivity region method to ensure the microstructure connectivity is examined through 2D and 3D structural design problems. We use linear quadrilateral elements and cubic elements to discretize the design domain in 2D and 3D, respectively. The microscale penalization parameter is assigned to be 4, and the macroscale one is set to be 3. If not otherwise specified, the radii for density filtering are respectively set as and times
Conclusions
In the present study, we propose a method to ensure the connectivity of lattice materials that are optimized with the inverse homogenization method, in a two-scale concurrent structural topology optimization problem for stiffness maximization. The macroscale distributions of lattice materials and their microstructural topologies are determined simultaneously. In the framework of the density-based topology optimization method, the DMO method is adopted to penalize the intermediate densities and
CRediT authorship contribution statement
Pai Liu: Conceptualization, Methodology, Formal analysis, Writing - review & editing. Zhan Kang: Supervision, Conceptualization, Writing - review & editing. Yangjun Luo: Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The authors acknowledge the support of the National Science Foundation of China (11902064, 11872140).
References (59)
- et al.
Topology optimization for concurrent design of layer-wise graded lattice materials and structures
Int. J. Eng. Sci.
(2019) - et al.
Optimal active cooling performance of metallic sandwich panels with prismatic cores
Int. J. Heat Mass Transf.
(2006) - et al.
Active cooling by metallic sandwich structures with periodic cores
Prog. Mater. Sci.
(2005) - et al.
A level set method for structural topology optimization
Comput. Methods Appl. Mech. Eng.
(2003) - et al.
Structural optimization using sensitivity analysis and a level-set method
J. Comput. Phys.
(2004) - et al.
Design of materials with extreme thermal expansion using a three-phase topology optimization method
J. Mech. Phys. Solids
(1997) Materials with prescribed constitutive parameters: an inverse homogenization problem
Int. J. Solids Struct.
(1994)- et al.
Optimizing multifunctional materials: design of microstructures for maximized stiffness and fluid permeability
Int. J. Solids Struct.
(2006) - et al.
Multi-material topology optimization of lattice structures using geometry projection
Comput. Methods Appl. Mech. Eng.
(2020) - et al.
Optimum structure with homogeneous optimum truss-like material
Comput. Struct.
(2008)
Concurrent topology optimization design of material and structure within FE2 nonlinear multiscale analysis framework
Comput. Methods Appl. Mech. Eng.
Topology optimization for concurrent design of structures with multi-patch microstructures by level sets
Comput. Methods Appl. Mech. Eng.
Topology optimisation of manufacturable microstructural details without length scale separation using a spectral coarse basis preconditioner
Comput. Methods Appl. Mech. Eng.
Topology optimization of hierarchical lattice structures with substructuring
Comput. Methods Appl. Mech. Eng.
High-stiffness and strength porous maraging steel via topology optimization and selective laser melting
Addit. Manuf.
Design methodology for porous composites with tunable thermal expansion produced by multi-material topology optimization and additive manufacturing
Compos. Part B Eng.
Natural frequency optimization of 3D printed variable-density honeycomb structure via a homogenization-based approach
Addit. Manuf.
Topology optimization of 3D self-supporting structures for additive manufacturing
Addit. Manuf.
Deposition path planning-integrated structural topology optimization for 3D additive manufacturing subject to self-support constraint
Comput. Des.
Level set-based topology optimization with overhang constraint: towards support-free additive manufacturing
Comput. Methods Appl. Mech. Eng.
Role of anisotropic properties on topology optimization of additive manufactured load bearing structures
Scr. Mater.
Cellular level set in B-splines (CLIBS): a method for modeling and topology optimization of cellular structures
Comput. Methods Appl. Mech. Eng.
Compatibility in microstructural optimization for additive manufacturing
Addit. Manuf.
Design of graded lattice structure with optimized mesostructures for additive manufacturing
Mater. Des.
Topology optimization of modulated and oriented periodic microstructures by the homogenization method
Comput. Math. Appl.
Homogenization-based stiffness optimization and projection of 2D coated structures with orthotropic infill
Comput. Methods Appl. Mech. Eng.
De-homogenization of optimal multi-scale 3D topologies
Comput. Methods Appl. Mech. Eng.
Topology optimization of non-linear elastic structures and compliant mechanisms
Comput. Methods Appl. Mech. Eng.
Topology optimization of coated structures and material interface problems
Comput. Methods Appl. Mech. Eng.
Cited by (73)
Multiscale topology optimization of structures by using isogeometrical level set approach
2024, Finite Elements in Analysis and DesignA multivariate level set method for concurrent optimization of graded lattice structures with multiple microstructure prototypes
2024, Computer Methods in Applied Mechanics and EngineeringMulti-scale topology optimization of structures with multi-material microstructures using stiffness and mass design criteria
2024, Advances in Engineering SoftwareTopology optimization of differentiable microstructures
2024, Computer Methods in Applied Mechanics and EngineeringRational designs of mechanical metamaterials: Formulations, architectures, tessellations and prospects
2023, Materials Science and Engineering R: ReportsMFSE-based two-scale concurrent topology optimization with connectable multiple micro materials
2023, Computer Methods in Applied Mechanics and Engineering