A level set based topology optimization for finite unidirectional acoustic phononic structures using boundary element method

https://doi.org/10.1016/j.cma.2021.113776Get rights and content

Highlights

  • Average topological derivative is specified for each design domain.

  • Original and adjoint fields are both computed through a size reduced system matrix.

  • Resluts from different numbers of design domains are obtained.

  • The band gap properties are enhanced with more layers.

Abstract

The paper presents a level set based topology optimization method for unidirectional phononic structures with finite layers of lattice cells. Boundary element method(BEM) is employed as the numerical approach to solve the acoustic problems governed by Helmholtz equation. A sized reduced coefficient matrix is derived due to the iteration forms for input and output quantities on the periodic boundary of unit cells. Topological derivatives are formulated by boundary integral equation combined with adjoint variable method and computed for each layer. An average topological sensitivity of a single design domain is proposed for the updating of the level set function(LSF) governed by an evolution equation. Numerical models with different number of layers are considered and several optimized structures of unit cells are obtained in concerned frequencies. A further investigation into the transmission of acoustic waves is carried out by employing more layers of the periodic structures between the input and output domains. The results demonstrate the effectiveness of the proposed optimization method for the finite unidirectional phononic structures.

Introduction

There are continuing interests in design of phononic crystals and metamaterials which are artificially manufactured periodic structures composed of multi materials [1], [2] . The novel properties of phononic structures such as waveguides [3], [4], negative refraction [5], [6], [7] and negative elastic modulus [8], [9], [10], make them have a promising prospect of applications in manipulation of acoustic or elastic waves. Since the experimental investigation of the propagation characteristics of acoustic waves in a sculpture located in Madrid [11], a great amount of efforts has been devoted to developing numerical methods for simulations of phononic structures [12], [13], [14], [15]. The design of phononic structures is usually carried out by adjusting the configuration of the inclusions and host materials to extend or obtain band gaps in desired ranges [16], [17], [18]. The influence of geometric and material parameters, filling fraction and shapes of inclusions on the width of the lowest band gap is investigated through finite element method (FEM) [16]. A tunable passband is designed by applying external electric field onto the piezoelectric layers in phononic crystals [17]. Unidirectional phononic crystals with functionally graded interlayers are investigated through transfer matrix method [18].

With increased interest in the development of phononic devices, it is necessary to construct effective methodologies for the design of phononic structure/metamaterials. In preliminary studies, an investigation of optimization design for maximizing the band gaps of phononic structures is carried out by using FEM [19], and after that, various methodologies [20], [21], [22] such as genetic algorithm, evolutionary algorithm, and level set method, are adopted for optimization design of 2D phononic crystals. The combination of optimization algorithms and various numerical methods [23], [24], [25] provides more efficient approaches for the creation of unit cells through computational techniques. A topological design on the unit cell optimization is carried out through FEM and bidirectional evolutionary structural method to obtain ultrawide band gap for in-plane and out-of-plane wave modes [23]. Adaptive genetic algorithm is also employed to calculate the mutation and crossover rate for the optimization of topological configurations [24]. In the literature [25], bidirectional evolutionary structure optimization method is applied to the design of phononic crystals for maximizing spatial decay of evanescent waves. The density based-topology optimization methods [26], [27] usually result in a continuous distribution of density which shows grayscale areas. Therefore, it is difficult to specify the boundary conditions in the fixed design domain. The adoption of level set-based structural optimization methods [28], [29] avoids the intermediate state of the boundary definition. However, the boundary conditions on the fixed design domain are approximated through ersatz material approach [30] which employs a domain integral to replace the boundary integral. To avoid the weak material or intermediate material and generate actual boundaries defined by the iso-surface of LSF, we employ BEM as a numerical approach.

Boundary-only discretization provided by BEM [31] leads to dimension-reduced elements. Therefore, BEM simplifies the modeling process and reduces the DOFs. BEM has been applied to sensitivity analysis for shape optimization problems [32], [33], [34]. With the development of isogeometric analysis (IGA), BEM provides promising applications as a numerical approach which is conveniently combined with the IGA due to the feature of boundary-only discretization. Many research works have been carried out to develop the IGA for various applications [35], [36], [37], [38]. The isogeometric BEM(IGABEM) [39], [40] is developed for acoustic problems and introduced to real-world engineering applications. Moreover, IGABEM is implemented for fracture modeling and crack propagation simulations [41], [42], since knot insertion in B-splines can provide discontinuities in the geometry. For 2D and 3D multi-patch coupling in statics, modal analysis and contact problems, the skew-symmetric Nitsche’s formulation is introduced by the Ref. [43]. Moreover, the IGABEM is also developed for optimization problems [44], [45], [46], [47], [48], [49]. For sound absorbing problems, the IGABEM is adopted for both 2D and 3D cases to optimize the distribution of materials [49], [50]. The research work in the Ref. [51] employs coarse meshes provided by Computer Aided Design (CAD) to reduce the number of design variables and fine meshes for accurate evaluation of sensitivities. Furthermore, Geometry Independent Field approximaTion (GIFT) proposed in the Ref. [52] allows to reduce the number of control points while acquiring a refinement of the variable space for interior regions, which means GIFT can be potentially applied to optimization problems to improve the efficiency.

The combination of BEM and level set method is investigated in various problems [53], [54], [55]. The adoption of level set method leads to clear definition of boundaries which can be directly discretized into boundary elements conveniently. Moreover, due to the features of numerical model from BEM and the properties of fundamental solutions, BEM is also suitable for acoustic or electromagnetic problems defined in infinite domains. The literatures [53], [55] present effective applications of the combination methodology for heat conduction problems and the literature [54] extends it to the optimization design for a coupling problem. The phononic structure composed of periodically distributed scatterers in the infinite domain is investigated through BEM by considering the scattering frequencies [56], [57] and the topology optimization of the scatterers is carried out through scattering matrix method combined with contour integral method [58], [59].

Optimization issues involving eigenvalues arise in many applications related to acoustic or elastodynamic problems [60], [61], [62]. The combination of the contour integral method with BEM requires repeat computations for the numerical evaluation of the contour integrals. The topological optimization algorithm usually leads to high computation cost since the evolution of topology demands the iteration of the solving process for eigenvalues.

However, due to the previous research [63], the finite periodic structures present band gap properties which is determined by the employed unit cell and the reduction of transmission of acoustic waves is enhanced as the number of cells increases. To avoid the high computation cost resulted by the Bloch eigenvalue value analysis which is usually carried out to illustrate the band structure of the unit cell, the topological optimization algorithm for finite unidirectional phononic structures is proposed in this work. Finite layers of design domains are considered and the average topological derivatives which are formulated using adjoint variable method, are adopted for each design domain. The original and adjoint fields are both solved through BEM with a size reduced coefficient matrix by formulating local system matrix of the unit cell. The source code which accompanies this paper is available for downloading from https://sourceforge.net/projects/topology-opt-bem-lsf/ for Linux and Mac OS computers.

The rest parts of the paper are organized as follows: Section 2 explains the derivation of the average topological derivatives using the adjoint variable method. In Section 3, the BEM formulas for the finite unidirectional phononic structures are presented. Several numerical examples with different conditions are investigated to demonstrate the effectiveness of the proposed optimization algorithm. The final parts consist of the conclusions and acknowledgment.

Section snippets

Level set method

Let us introduce a scalar function ϕ(x), called level set function defined in a design domain D. With a given range [1,1], ϕ(x) is defined as follows: 0<ϕ(x)1,xΩΓ,ϕ(x)=0,xΓ,1ϕ(x)<0,xDΩ,where Ω is material region and Γ denotes the boundary of Ω. Obviously it can be seen that the distribution of the material is related to the value of the level set function ϕ(x). Thus an evolution of ϕ(x) can be carried out to achieve the optimum topology configuration of the material. As shown in Fig. 1

BEM for finite unidirectional phononic structures

To formulate the system equations for the finite periodic structures, Eq. (2) can be transformed into expression of integral equation forms with the Green identity: p(y)+Γq(x,y)p(x)dΓ(x)=Γp(x,y)q(x)dΓ(x)inΩ,and its normal directional derivative with respect to point y is written as q(y)+Γqˆ(x,y)p(x)dΓ(x)=Γpˆ(x,y)q(x)dΓ(x)inΩ,where the p(x,y) is the fundamental solution for Eq. (2), y, x are collocation point and source point, respectively, and r=|xy| is the distance between x and y. p

Verification of topological derivative

A model with input and output domains is depicted in Fig. 5. The left boundary of input domain DI is specified with Dirichlet boundary condition p=p̄0=1 [Pa] and the upper and lower boundaries are considered as rigid wall with q=0 (Neumann boundary condition). The upper, lower and right boundaries of the output domain DO are also considered as rigid wall with q=0. Furthermore, the objective function f=|ppobj|2 is defined and evaluated on the 2 elements in the center of right boundary of DO. To

Conclusions

This paper presents a level set based topology optimization method for finite unidirectional phononic structures using BEM. The adjoint method is employed to derive the topological derivative associated with the objective functional defined on a portion of boundary in the output domain and the acoustic problems formulated in adjoint and original fields are solved through boundary element method. The boundary elements are generated by discretizing the zero contour line of the level set function.

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.

Acknowledgment

This work is supported by the Applied Basic Research Programs of Shanxi Province of China [grant number 201801D221022], the National Natural Science Foundation of China [grant numbers 11872168, 52075361], the Key R & D Projects in Shanxi Province of China [grant number 201903D421030], the Major Special Projects of Science and Technology in Shanxi Province of China [grant number 20201102003], and the China Scholarship Council [grant number 201908140142].

References (66)

  • LiF.L. et al.

    Boundary element method for band gap calculations of two-dimensional solid phononic crystals

    Eng. Anal. Bound. Elem.

    (2013)
  • KrushynskaA. et al.

    Towards optimal design of locally resonant acoustic metamaterials

    J. Mech. Phys. Solids

    (2014)
  • FomenkoS. et al.

    In-plane elastic wave propagation and band-gaps in layered functionally graded phononic crystals

    Int. J. Solids Struct.

    (2014)
  • HyunC.J. et al.

    A genetic algorithm for multiple objective sequencing problems in mixed model assembly lines

    Comput. Oper. Res.

    (1998)
  • Garcia-LopezN. et al.

    An improved robust topology optimization approach using multiobjective evolutionary algorithms

    Comput. Struct.

    (2013)
  • JiangY. et al.

    Topology optimization under design-dependent loads with the parameterized level-set method based on radial-basis functions

    Comput. Methods Appl. Mech. Engrg.

    (2020)
  • WangK. et al.

    Ultrawide band gap design of phononic crystals based on topological optimization

    Physica B

    (2019)
  • HanX. et al.

    Bandgap design of three-phase phononic crystal by topological optimization

    Wave Motion

    (2020)
  • ChenY. et al.

    Maximizing spatial decay of evanescent waves in phononic crystals by topology optimization

    Comput. Struct.

    (2017)
  • AllaireG. et al.

    Structural optimization using sensitivity analysis and a level-set method

    J. Comput. Phys.

    (2004)
  • BrebbiaC.A.

    The birth of the boundary element method from conception to application

    Eng. Anal. Bound. Elem.

    (2017)
  • AithalR. et al.

    Shape sensitivity analysis in thermal problems using BEM

    Eng. Anal. Bound. Elem.

    (1995)
  • SianoD. et al.

    Sensitivity analysis and correlation experimental/numerical FEM-bem for noise reduction assessment of an engine beauty cover

    Energy Procedia

    (2015)
  • ModaveA. et al.

    Corner treatments for high-order local absorbing boundary conditions in high-frequency acoustic scattering

    J. Comput. Phys.

    (2020)
  • WuH. et al.

    Isogeometric finite element analysis of interior acoustic problems

    Appl. Acoust.

    (2015)
  • SimpsonR. et al.

    Acoustic isogeometric boundary element analysis

    Comput. Methods Appl. Mech. Engrg.

    (2014)
  • CooxL. et al.

    A performance study of NURBS-based isogeometric analysis for interior two-dimensional time-harmonic acoustics

    Comput. Methods Appl. Mech. Engrg.

    (2016)
  • PengX. et al.

    Isogeometric boundary element methods for three dimensional static fracture and fatigue crack growth

    Comput. Methods Appl. Mech. Engrg.

    (2017)
  • ChenL.L. et al.

    Structural shape optimization of three dimensional acoustic problems with isogeometric boundary element methods

    Comput. Methods Appl. Mech. Engrg.

    (2019)
  • LianH. et al.

    Shape optimization directly from cad: An isogeometric boundary element approach using t-splines

    Comput. Methods Appl. Mech. Engrg.

    (2017)
  • LiuC. et al.

    Shape optimization of sound barrier using an isogeometric fast multipole boundary element method in two dimensions - sciencedirect

    Eng. Anal. Bound. Elem.

    (2017)
  • ChenL. et al.

    Acoustic topology optimization of sound absorbing materials directly from subdivision surfaces with isogeometric boundary element methods

    Comput. Methods Appl. Mech. Engrg.

    (2020)
  • ChenL. et al.

    An isogeometric approach of two dimensional acoustic design sensitivity analysis and topology optimization analysis for absorbing material distribution

    Comput. Methods Appl. Mech. Engrg.

    (2018)
  • Cited by (13)

    • A BEM-based topology optimization for acoustic problems considering tangential derivative of sound pressure

      2022, Computer Methods in Applied Mechanics and Engineering
      Citation Excerpt :

      Jiang et al. [45] propose the topology optimization method based on IGABEM and the piecewise constant level method for the distribution design of the sound absorbing materials. Gao et al. [46]derive the average topological derivative for the topology optimization problem for finite unidirectional phononic structures using the BEM and level set method. Oliveira et al. [47] employ IGABEM and the level set method for topology optimization.

    • Enhanced ultrasonic wave generation using energy-localized behaviors of phononic crystals

      2022, International Journal of Mechanical Sciences
      Citation Excerpt :

      One mechanism of phononic band-gap formation is Bragg-scattering; its periodic nature brings about repeated destructive interferences [6]. Several research topics on the phononic band gap include the modeling [7], experimental validation [8], unit-cell designs for lowering [9] or broadening [10] frequency ranges, and design-optimization execution [11]. Promising engineering applications have been presented, including wave mitigation [12], isolation [13], and attenuation [14].

    • Partitioned path-following strategy for nonlinear structural analyses using the boundary element method

      2022, Computer Methods in Applied Mechanics and Engineering
      Citation Excerpt :

      The Boundary Element Method (BEM) is recognized as a robust and efficient numerical tool, characterized by a one-order reduction in mesh dimensionality and its ability to accurately reproduce high gradient stress fields [9–11]. It finds applications in several areas of engineering, including isogeometric analysis [12,13], biomechanics [14,15], fracture mechanics [16–19], acoustics [20–22], shape [23,24] and topology optimization [25–29]. Some reviews [30–32] provide more details on the development and application of the BEM.

    • Analytically-integrated radial integration PBEM for solving three-dimensional steady heat conduction problems

      2021, Engineering Analysis with Boundary Elements
      Citation Excerpt :

      First, just the 2D boundary of the physical model should be discretized which facilitates the modeling of complex geometries. Second, semi-analytical solutions to problems make the results have high accuracy [13–16]. These merits promote the application of BEM in heat conduction problems, so BEM becomes one of the most popular numerical methods.

    View all citing articles on Scopus
    View full text