Rapid modeling simulation method for device with double-color scattered material
By applying hybrid finite element fast algorithms and tree tree technology in intelligent wearable devices, the problem of low efficiency of commercial tools when dealing with high-frequency broadband and complex geometric structures is solved, and accurate and efficient electromagnetic simulation of multi-scale models is achieved. , avoid simulation errors.
Patent Information
- Application Number
- CN202411126847.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-29
AI Technical Summary
Existing commercial electromagnetic simulation tools are inefficient in handling high-frequency, broadband and complex geometric structures in smart wearable devices, especially for multi-scale models with two-dispersive material structures, which may cause simulation errors, making it difficult to achieve accurate and efficient electromagnetic fields. simulation.
Using a hybrid finite element fast algorithm, combined with HFSS tree tree technology and second-order Arnoldi modal reduction technology, the three-dimensional structure of smart watches is meshed and electromagnetically simulated, and the tree tree decomposition technology and impedance boundary conditions are reduced. The amount of calculation is used to reduce the matrix dimensions using the Arnoldi algorithm to speed up the calculation.
It significantly improves the accuracy and efficiency of multi-scale models with dual-dispersive material devices in electromagnetic simulation, reduces the computational complexity and time, avoids simulation errors, and improves the efficiency of antenna performance analysis of smart wearable devices.
Smart Images

Figure CN119047251A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of electromagnetic field modeling and simulation, and in particular relates to a rapid modeling and simulation method for a device with a dual-dispersion material. Background Art
[0002] With the rapid development of wireless communication technology and smart wearable technology, the size of chips and antennas in smart wearable devices tends to be integrated and miniaturized. The materials and device structures used are becoming more and more complex and diverse. Structures with dual-dispersion materials are widely used in modern communication equipment.
[0003] Electromagnetic analysis plays an important role in scientific research and engineering design. The finite element method has become an important and widely recognized tool for solving engineering and mathematical problems. As the radio frequency spectrum of electronic equipment becomes increasingly crowded, the difficulty in design and simulation testing has increased dramatically. Existing well-known commercial electromagnetic simulation tools are not efficient enough when dealing with high frequencies, broadbands and complex geometric structures.
[0004] The mainstream commercial software currently uses traditional electromagnetic simulation methods. Frequency domain electromagnetic simulation is to calculate the simulation model frequency by frequency point, which takes a lot of time. For multi-scale models with dual-dispersion material structures, electromagnetic calculations may even result in simulation errors. Therefore, it is particularly important to improve the accuracy and efficiency of electromagnetic field simulation algorithms for multi-scale devices with dual-dispersion materials.
[0005] In view of this, it is very meaningful to propose a fast modeling and simulation method for devices with dual-dispersion materials. Summary of the invention
[0006] In order to solve the problem that the size of chips and antennas in smart wearable devices tends to be integrated and miniaturized, and their material properties and structures are more complex and diverse, the current mainstream commercial software uses traditional electromagnetic simulation methods, which has low solution efficiency and may result in simulation errors when performing electromagnetic calculations on multi-scale devices with dual-dispersion material structures. The present invention provides a rapid modeling and simulation method for devices with dual-dispersion materials to solve the above problems.
[0007] In a first aspect, the present invention provides a rapid modeling and simulation method for a device having a dual-dispersion material, the method comprising the following steps:
[0008] Construct the 3D structure of the smart watch in HFSS and set its material parameters; build the PCB, antenna and structure with dual-dispersion material properties in the watch dial, where the magnetic permeability of the magnetic material is 120 and the magnetic loss is 0.04, adjust the size of the human hand structure and insert it into the watch strap; this model is a multi-scale model, the maximum size of its structure is 320.2mm, the minimum size is 0.02mm, and the multi-scale ratio is 16010;
[0009] Mesh the human tissue and smart watch antenna models in HFSS, and use the hybrid finite element fast algorithm to analyze the impact of magnetic material structure on the performance of smart watch antennas;
[0010] The HFSS discrete mesh data is read and solved using the hybrid finite element fast algorithm. This algorithm is accurate and efficient in calculating multi-scale models with tens of millions of degrees of freedom and dual-dispersion material properties.
[0011] This method first imports the CAD file of the smart watch into HFSS, sets the strap material to plastic, the watch metal shell material to aluminum, and the watch display material to glass.
[0012] Preferably, the PCB, antenna and ferrite structure with dual-dispersion material properties are placed inside the watch dial, and the human hand structure is adjusted and inserted into the watch strap. This model is a multi-scale model, in which the maximum dimension of the structure is 320.2 mm, the minimum dimension is 0.02 mm, and the multi-scale ratio is 16010; the material of the human hand is a dispersive material with an average dielectric loss of 0.5.
[0013] Further preferably, the material properties of the human tissue and the smart watch antenna are set in HFSS, the absorbing boundary condition ABC is added, and the lumped port and electrode excitation information are set.
[0014] Secondly, the hybrid finite element fast algorithm is used to analyze the impact of the structure with or without magnetic material properties on the performance of smart watch antennas, mainly in the following aspects:
[0015] Using the HFSS mesh, the hybrid finite element fast algorithm is used to calculate the S parameters of the smart watch antenna with / without magnetic material property structure.
[0016] Thirdly, the use of hybrid finite element fast algorithms to accelerate the calculation and analysis of multi-scale models with tens of millions of degrees of freedom and dual-dispersion material devices is mainly:
[0017] The human body tissue and smart watch antenna were meshed discretized in HFSS, and the hybrid finite element fast algorithm was used for electromagnetic simulation;
[0018] A hybrid finite element fast algorithm is suitable for fast modeling and simulation of dual-dispersion material devices. The hybrid finite element fast algorithm includes:
[0019] Based on the differential form of Maxwell's equations, the electric field wave equations to be solved for the human tissue and smart watch antenna model in the frequency domain of the electromagnetic field are obtained:
[0020]
[0021] in, is the Hamiltonian operator, μ is the magnetic permeability, E is the electric field intensity, ∈ is the dielectric constant, σ e represents the conductivity, J i is the current source density, s represents the complex frequency, that is, s=jω, where j refers to the imaginary unit and ω is the angular frequency.
[0022] In the hybrid finite element fast algorithm, the electric field intensity E is decomposed by the edge basis function Φ defined on the cotree edge through the tree-cotree decomposition technique. c and the node basis functions defined at the free nodes The gradient expansion is expressed as follows:
[0023]
[0024] Among them, N c is the number of cotree edges in the grid, N n is the number of free nodes in the grid, refers to the i-th edge basis function defined on the cotree edge, It refers to the gradient of the jth node basis function defined at the free node, and e represents the corresponding unknown coefficient.
[0025] Preferably, a hybrid finite element fast algorithm is used to calculate the conductor loss of human tissue and smart watch antenna. First, the metal structure in the watch is modeled using impedance boundary conditions (IBC), and the expression is as follows:
[0026] n×H=jωY s (n×(E×n))
[0027] Where n is the normal vector of the metal surface, H is the magnetic field, and Y s is the surface admittance of the metal conductor, expressed as The IBC boundary accurately describes the relationship between the tangential electric field and the tangential magnetic field on the metal surface. The use of the IBC boundary avoids solving the interior of the metal, thereby reducing the amount of calculation.
[0028] Substitute the IBC boundary into the matrix equation to be solved and use the Galerkin method to obtain the final system matrix expression as follows:
[0029]
[0030] Where b is the input matrix of the entire calculation area, and the expressions of the mass matrix M, damping matrix C, stiffness matrix K and boundary term matrix Z are as follows:
[0031]
[0032]
[0033]
[0034]
[0035] in<A,B> Ω It means that A and B are inner products within the range of Ω. Let s = s0 (1 + δ), where s0 is the reference complex frequency, generally the complex center frequency. The matrix equation to be solved after expansion is:
[0036] A0e+δA1e+δ 2 A2e=-s0(1+δ)b
[0037] In the above equation, The approximate equivalent is 1+0.5δ, where
[0038] Further preferably, the second-order equation is solved based on the second-order Arnoldi algorithm, in which the dimension of the inverse matrix equation is consistent with the first-order equation, so the computational complexity is almost the same as that of the linear equation, and it is suitable for solving electromagnetic field problems with multiple inputs and multiple outputs. By adding auxiliary equations, the expanded matrix equation to be solved is converted into a linear equation system:
[0039]
[0040] In addition, the above formula can be restated as:
[0041]
[0042] Where A = -A0 -1 A1,B=-A0 -1 A2. Therefore, the system transfer function matrix H(δ) of the matrix equation to be solved after the equation expansion is defined as:
[0043] H(δ)=(I-δF) -1 B0
[0044] in, Here, H(δ) can be further expanded into the following power series:
[0045]
[0046] The specific process of constructing a low-dimensional Krylov subspace to approximate the transfer function is as follows: First, a Krylov subspace is selected as the projection subspace, which is defined as:
[0047] K m (F; B0) = span {B0, FB0, F 2 B0,...,F m-1 B0}
[0048] Second, at each iteration of the Arnoldi algorithm, there is this relationship:
[0049]
[0050] Therefore, the Krylov subspace K can be immediately derived m (F; B0) can be constructed by a standard orthogonal basis, namely:
[0051]
[0052] in, Let Q = {R0, R1, R2, ..., R m-1}
[0053] The projection subspace Q is generated by using a modal order reduction technique based on a second-order Arnoldi algorithm, wherein the second-order Arnoldi algorithm specifically includes:
[0054] S31, Q0=orth(Q0)=[q1,...,q M ], where M represents the number of columns of the right-hand side. P0 is a zero matrix with the same dimension as Q0, and orth represents orthogonalization of the matrix in the brackets;
[0055] S32, perform cyclic assignment,
[0056] S321, define the value of variable j from 1 to m;
[0057] S322, let r=A1q j +A2p j , G T r=0, t=q j ,in I represents the identity matrix, and 0 represents the zero matrix;
[0058] S33, define the value of variable i from 1 to j+M-1;
[0059] S331, order in sequence r=rq i h i,j , t = tp i h i,j ;
[0060] S332, when h j+l,j =norm(r), norm() means calculating the 2-norm of r, q j+l =r / h j+l,j , p j+l =t / h j+l,j When the loop ends;
[0061] S34. The loop ends.
[0062] Expand the unknown vector e using the orthogonal basis Q, that is,
[0063]
[0064] Where w=[w0,w1,…,w m-1 ] is the coefficient vector. Substitute the above formula into the matrix equation to be solved, and multiply both sides of the matrix equation to be solved by the transpose Q T , and the reduced matrix equation is obtained:
[0065]
[0066] where m is the number of iterations in the second-order Arnoldi algorithm. and The expressions are By solving the reduced-order matrix equation at each frequency to obtain the coefficient vector w, the electric field at each frequency can be obtained.
[0067] The dimension of the reduced matrix is determined by the number of basis vectors in the projection space, which is usually smaller than the dimension of the original matrix. Therefore, it is easier to obtain the frequency response at each frequency by solving the reduced matrix equation. In addition, the second-order Arnoldi algorithm does not include Schmidt orthogonalization and eigenvalue calculation, which improves the computational efficiency of the projection subspace.
[0068] Further preferably, the dielectric loss and magnetic dielectric loss are modeled. The relative dielectric constant ∈ r and relative magnetic permeability μ r Written in the form of real and imaginary parts respectively, the expressions are as follows:
[0069] ∈ r =∈′-j∈″=∈′(1-jtanD)
[0070] μ r =μ′-jμ″=μ′(1-jtanM)
[0071] Among them, ∈′ is the real part of the relative dielectric constant, tanD is the electric loss tangent, μ′ is the real part of the relative magnetic permeability, and tanM is the magnetic loss tangent. In a wider frequency band, ∈′, tanD, μ′ and tanM appear to be constants or frequency-dependent. The vector fitting method can be used to fit ∈ r and Make a rational approximation, then ∈ r and It can be expressed as:
[0072] ∈ r =k ∈ +χ e (ω)
[0073]
[0074] Among them, k ∈ and k μ is a constant term, χ e (ω) and C m (ω) is a frequency-dependent function, and its expressions are
[0075]
[0076]
[0077] where a p and c p They represent the residue of the fitting function, b p and d p Represent the poles of the fitting function, N e and N m represents the fitting order. Then, r The associated mass matrix M is divided into two parts, one part M1 and the constant term k ∈ The other part M2 is related to the frequency-related term χ e (ω), the expressions of M1 and M2 are as follows:
[0078]
[0079]
[0080] Similarly, The relevant stiffness matrix K is divided into two parts, one part K1 and the constant term k μ The other part K2 is related to the frequency-related term C m (ω), K1 and K2 are expressed as follows:
[0081]
[0082]
[0083] Among them, ∈0 and μ0 are the dielectric constant and magnetic permeability in vacuum, respectively. Generally, for lossy media, the constant terms after fitting the relative dielectric constant and relative magnetic permeability account for a large proportion, resulting in M1 being much larger than M2, and K1 being much larger than K2. Therefore, before reducing the matrix equation, only M1 and K1 can be considered, and the influence of M2 and K2 can be added to the matrix reduction equation.
[0084] Taking conductor loss, dielectric loss and magnetic loss into consideration, the final system matrix equation is obtained, which is expressed as follows:
[0085]
[0086] Through the projection subspace Q, we can obtain e=Qw, and substitute it into the electric field expression. Get the electromagnetic fields of human tissue and smart watch antenna at each solved frequency.
[0087] Compared with the prior art, the beneficial results of the present invention are:
[0088] A human hand and a smart watch are constructed in mainstream commercial software, and a PCB, an antenna, and a three-dimensional structure with dual-dispersion material properties are built inside the dial. The magnetic permeability of the magnetic material is 120, and the magnetic loss is 0.04. The hybrid finite element fast algorithm, which introduces tree-cotree technology and Arnoldi modal reduction technology on the traditional finite element method, is used to analyze the impact of the presence or absence of magnetic materials on the performance of the smart watch antenna. The hybrid finite element fast algorithm is used to solve the smart watch antenna with magnetic materials. This algorithm has both accuracy and efficiency in calculating the multi-scale model results with dual-dispersion material properties. BRIEF DESCRIPTION OF THE DRAWINGS
[0089] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated into and constitute a part of this specification. The accompanying drawings illustrate the embodiments and are used together with the description to explain the principles of the present invention. It will be easy to recognize other embodiments and many expected advantages of the embodiments because they become better understood by reference to the following detailed description. The elements of the drawings are not necessarily to scale with each other. The same reference numerals refer to corresponding similar parts.
[0090] Figure 1 A schematic flow chart of a rapid modeling and simulation method for a dual-dispersion material device according to an embodiment of the present invention;
[0091] Figure 2 A schematic diagram of a model structure in a rapid modeling and simulation method for a dual-dispersion material device according to an embodiment of the present invention;
[0092] Figure 3 A schematic diagram of the structure of a smart watch in a rapid modeling and simulation method with a dual-dispersion material device according to an embodiment of the present invention;
[0093] Figure 4 The material information of the human hand of the model in the embodiment of the present invention;
[0094] Figure 5 This is a comparison diagram of S parameters obtained by using a hybrid finite element fast algorithm and commercial software HFSS between 1 GHz and 4 GHz for the lumped port 1 of the structure model of a material having no magnetic properties in an embodiment of the present invention;
[0095] Figure 6 This is a comparison chart of S parameters obtained by using a hybrid finite element fast algorithm between 1 GHz and 4 GHz for the lumped port 1 of the model with a double-dispersion (dielectric loss, magnetic dielectric loss) material and a material without magnetic properties in an embodiment of the present invention, wherein the black dotted curve is the S parameter calculated for the material without magnetic properties, and the red solid curve is the S parameter calculated for the material with double-dispersion (dielectric loss, magnetic dielectric loss);
[0096] Figure 7 This is a comparison diagram of S parameters obtained by using a hybrid finite element fast algorithm and commercial software HFSS between 1 GHz and 4 GHz for the lumped port 1 having a dual-dispersion material structure model in an embodiment of the present invention. DETAILED DESCRIPTION
[0097] In the following detailed description, reference is made to the accompanying drawings, which form a part of the detailed description and are shown by illustrative specific embodiments in which the present invention can be put into practice. To this end, directional terms, such as "top", "bottom", "left", "right", "up", "down", etc., are used with reference to the orientation of the figures described. Because the parts of the embodiments can be positioned in several different orientations, directional terms are used for the purpose of illustration and are by no means limiting. It should be understood that other embodiments may be utilized or logical changes may be made without departing from the scope of the present invention. Therefore, the following detailed description should not be adopted in a limiting sense, and the scope of the present invention is defined by the appended claims.
[0098] Figure 1 It is shown that the embodiment of the present invention discloses a fast simulation modeling method for a device with a dual dispersion material, such as Figure 1 As shown, the method comprises the following steps:
[0099] S1. Import the CAD file of the smart watch into HFSS, set the strap material to plastic, the metal shell material to aluminum, and the display material to glass;
[0100] Place the PCB, antenna and ferrite structure with dual-dispersion material properties inside the watch dial, adjust the human hand structure and insert it into the watch strap, where the material of the human hand is a dispersive material with an average dielectric loss of 0.5; this model is a multi-scale model, the maximum size of its structure is 320.2mm, the minimum size is 0.02mm, and the multi-scale ratio is 16010;
[0101] In HFSS, the material properties of the human tissue and the smart watch antenna are set, the absorbing boundary condition ABC is added, and the lumped port and electrode excitation information are set.
[0102] S2. Read the HFSS-sectioned human tissue and smart watch antenna model meshes, use hybrid finite elements to quickly calculate the S parameters of the human tissue and smart watch antenna models, and analyze the impact of magnetic material structure on the performance of smart watch antennas.
[0103] S3. A hybrid finite element fast algorithm is used to solve the human tissue and smart watch antenna models with magnetic material structures. This algorithm is accurate and efficient in calculating multi-scale models with tens of millions of degrees of freedom and dual-dispersion material properties.
[0104] Based on the differential form of Maxwell's equations, the electric field wave equations to be solved for the human tissue and smart watch antenna model in the frequency domain of the electromagnetic field are obtained:
[0105]
[0106] in, is the Hamiltonian operator, μ is the magnetic permeability, E is the electric field intensity, ∈ is the dielectric constant, σ e The conductivity, J i is the current source density, s represents the complex frequency, that is, s=jω, where j refers to the imaginary unit and ω is the angular frequency.
[0107] In the hybrid finite element fast algorithm, the electric field intensity E is decomposed by the edge basis function Φ defined on the cotree edge through the tree-cotree decomposition technique. c and the node basis functions defined at the free nodes The gradient expansion is expressed as follows:
[0108]
[0109] Among them, N c is the number of cotree edges in the grid, N n is the number of free nodes in the grid, refers to the i-th edge basis function defined on the cotree edge, It refers to the gradient of the jth node basis function defined at the free node, and e represents the corresponding unknown coefficient.
[0110] The conductor loss of human tissue and smart watch antenna is analyzed using the hybrid finite element fast algorithm. First, the metal structure in the watch is modeled using the impedance boundary condition (IBC), and the expression is as follows:
[0111] n×H=jωY s (n×(E×n))
[0112] Where n is the normal vector of the metal surface, H is the magnetic field, and Y s is the surface admittance of the metal conductor, expressed as The IBC boundary accurately describes the relationship between the tangential electric field and the tangential magnetic field on the metal surface. The use of the IBC boundary avoids solving the interior of the metal, thereby reducing the amount of calculation.
[0113] Substitute the IBC boundary into the matrix equation to be solved and use the Galerkin method to obtain the final system matrix expression as follows:
[0114]
[0115] Where b is the input matrix of the entire calculation area, and the expressions of the mass matrix M, damping matrix C, stiffness matrix K and boundary term matrix Z are as follows:
[0116]
[0117]
[0118]
[0119]
[0120] in<A,B> Ω It means that A and B are inner products within the range of Ω. Let s = s0 (1 + δ), where s0 is the reference complex frequency, generally the complex center frequency. The matrix equation to be solved after expansion is:
[0121] A0e+δA1e+δ 2 A2e=-s0(1+δ)b
[0122] In the above equation, The approximate equivalent is 1+0.5δ, where
[0123] Further preferably, the second-order equation is solved based on the second-order Arnoldi algorithm, in which the dimension of the inverse matrix equation is consistent with the first-order equation, so the computational complexity is almost the same as that of the linear equation, and it is suitable for solving electromagnetic field problems with multiple inputs and multiple outputs. By adding auxiliary equations, the expanded matrix equation to be solved is converted into a linear equation system:
[0124]
[0125] In addition, the above formula can be restated as:
[0126]
[0127] Where A = -A0 -1 A1,B=-A0 -1 A2. Therefore, the system transfer function matrix H(δ) of the matrix equation to be solved after the equation expansion is defined as:
[0128] H(δ)=(I-δF) -1 B0
[0129] in, Here, H(δ) can be further expanded into the following power series:
[0130]
[0131] The specific process of constructing a low-dimensional Krylov subspace to approximate the transfer function is as follows: First, a Krylov subspace is selected as the projection subspace, which is defined as:
[0132] K m (F; B0) = span {B0, FB0, F 2 B0,...,F m-1 B0}
[0133] Second, at each iteration of the Arnoldi algorithm, there is this relationship:
[0134]
[0135] Therefore, the Krylov subspace K can be immediately derived m (F; B0) can be constructed by a standard orthogonal basis, namely:
[0136]
[0137] in, Let Q = {R0, R1, R2, ..., R m-1}
[0138] The projection subspace Q is generated by using a modal order reduction technique based on a second-order Arnoldi algorithm, wherein the second-order Arnoldi algorithm specifically includes:
[0139] S31, Q0=orth(Q0)=[q1,...,q M ], where M represents the number of columns of the right-hand side. P0 is a zero matrix with the same dimension as Q0, and orth represents orthogonalization of the matrix in the brackets;
[0140] S32, perform cyclic assignment,
[0141] S321, define the value of variable j from 1 to m;
[0142] S322, let r=A1q j +A2p j , G T r=0, t=q j ,in I represents the identity matrix, and 0 represents the zero matrix;
[0143] S33, define the value of variable i from 1 to j+M-1;
[0144] S331, order in sequence r=rq i h i,j , t = tp i h i,j ;
[0145] S332, when h j+l,j =norm(r), norm() means calculating the 2-norm of r, q j+l =r / h j+l,j , p j+l =t / h j+l,j When the loop ends;
[0146] S34. The loop ends.
[0147] Expand the unknown vector e using the orthogonal basis Q, that is,
[0148]
[0149] Where w=[w0,w1,…,w m-1 ] is the coefficient vector. Substitute the above formula into the matrix equation to be solved, and multiply both sides of the matrix equation to be solved by the transpose Q T , and the reduced matrix equation is obtained:
[0150]
[0151] where m is the number of iterations in the second-order Arnoldi algorithm. and The expressions are By solving the reduced-order matrix equation at each frequency to obtain the coefficient vector w, the electric field at each frequency can be obtained.
[0152] The dimension of the reduced matrix is determined by the number of basis vectors in the projection space, which is usually smaller than the dimension of the original matrix. Therefore, it is easier to obtain the frequency response at each frequency by solving the reduced matrix equation. In addition, the second-order Arnoldi algorithm does not include Schmidt orthogonalization and eigenvalue calculation, which improves the computational efficiency of the projection subspace.
[0153] The dielectric loss and magnetic dielectric loss of human tissue and smart watch antenna are analyzed using a hybrid finite element fast algorithm. The relative dielectric constant ∈ r and relative magnetic permeability μ r Written in the form of real and imaginary parts respectively, the expressions are as follows:
[0154] ∈ r =∈′-j∈″=∈′(1-jtanD)
[0155] μ r =μ′-jμ″=μ′(1-jtanM)
[0156] Among them, μ' is the real part of the relative dielectric constant, tanD is the electric loss tangent, μ' is the real part of the relative magnetic permeability, and tanM is the magnetic loss tangent. In a wider frequency band, μ', tanD, μ' and tanM appear to be constants or frequency-dependent. The vector fitting method can be used to ∈ r and Make a rational approximation, then ∈ r and It can be expressed as:
[0157] ∈ r =k ∈ +χ e (ω)
[0158]
[0159] Among them, k ∈ and k μ is a constant term, χ e (ω) and C m (ω) is a frequency-dependent function, and its expressions are
[0160]
[0161]
[0162] where a p and c p They represent the residue of the fitting function, b p and d p Represent the poles of the fitting function, N e and N m represents the fitting order. Then, r The associated mass matrix M is divided into two parts, one part M1 and the constant term k ∈ The other part M2 is related to the frequency-related term X e (ω), the expressions of M1 and M2 are as follows:
[0163]
[0164]
[0165] Similarly, The relevant stiffness matrix K is divided into two parts, one part K1 and the constant term k μ The other part K2 is related to the frequency-related term C m (ω), K1 and K2 are expressed as follows:
[0166]
[0167]
[0168] Among them, ∈0 and μ0 are the dielectric constant and magnetic permeability in vacuum, respectively. Generally, for lossy media, the constant terms after fitting the relative dielectric constant and relative magnetic permeability account for a large proportion, resulting in m1 being much larger than M2 and K1 being much larger than K2. Therefore, before reducing the matrix equation, only m1 and K1 can be considered, and the influence of M2 and K2 can be added to the matrix reduction equation.
[0169] Taking conductor loss, dielectric loss and magnetic loss into consideration, the final system matrix equation is obtained, which is expressed as follows:
[0170]
[0171] Through the projection subspace Q, we can obtain e=Qw, and substitute it into the electric field expression. Get the electromagnetic fields of human tissue and smart watch antenna at each solved frequency.
[0172] In this embodiment, the three-dimensional structure of the human hand and the smart watch is constructed in HFSS, such as Figure 2 As shown, and set its material parameters; build PCB, antenna and structure with dual dispersion material properties in the watch dial. The key structure inside the smart watch dial is as follows Figure 3As shown, the material information of the human hand is as follows Figure 4 shown.
[0173] Electromagnetic simulation of human tissue and smart watch antenna with non-magnetic material structure is performed in HFSS to obtain S parameter simulation results of human tissue and smart watch antenna in the range of 1-4GHz, such as Figure 5 As shown in Table 1, the simulation results calculated by the fast hybrid finite element algorithm (MFEM-fast), HFSS interpolation frequency sweep method (HFSS-interpolate) and HFSS fast frequency sweep method (HFSS-fast) are in good agreement. As shown in Table 1, the efficiency of the algorithm of the present invention is improved by 3.55 times and 8.79 times respectively compared with HFSS-interpolate and HFSS-fast, which verifies the stability and efficiency of the hybrid finite element fast algorithm.
[0174]
[0175] Table 1
[0176] The hybrid finite element fast algorithm is used to perform electromagnetic calculations on the performance of smart watch antennas with / without magnetic material structures, such as Figure 6 As shown, the S-parameter simulation results of human tissue and smart watch antennas with / without magnetic material structures in the range of 1-4GHz are obtained. Analysis shows that the smart watch antenna with magnetic material structure has more resonant working points and a wider frequency band.
[0177] The smart watch antenna model with a dual-dispersion material structure is meshed in HFSS, where the magnetic permeability of the magnetic material is 120 and the magnetic loss is 0.04. The patented algorithm is used to read the HFSS mesh for electromagnetic simulation to obtain the S-parameter simulation results of human tissue and smart watch antenna in the range of 1-4GHz, such as Figure 7 As shown in the figure, the S parameters calculated by the fast hybrid finite element method (MFEM-fast) are in good agreement with the simulation results calculated by the HFSS interpolation frequency sweep method (HFSS-interpolate), while the results calculated by the HFSS fast frequency sweep method (HFSS-fast) show obvious errors. The observed inaccuracies can be attributed to the inaccurate calculation of the eigenvectors by the HFSS fixed solver. This further verifies the accuracy of the hybrid finite element fast algorithm.
[0178] Furthermore, the test of the smart watch antenna model with a dual-dispersion material structure found that the hybrid finite element fast algorithm is 3.40 times faster than the HFSS interpolation frequency sweep method (HFSS-interpolate) simulation in terms of calculation time, and 9.38 times faster than the HFSS fast frequency sweep method (HFSS-fast) simulation, proving that the hybrid finite element fast algorithm can ensure accuracy and efficiency in electromagnetic analysis of multi-scale models with dual-dispersion materials. The specific test parameters are shown in Table 2:
[0179]
[0180] Table 2
[0181] In the specific implementation, this patent imports the smart watch CAD file in step format into HFSS, and builds the human hand, PCB, antenna and dual-dispersion material structure inside the dial in HFSS. This model is a multi-scale model, and the maximum size in its structure is 320.2mm, the minimum size is 0.02mm, and the multi-scale ratio is 16010; the material parameters, lumped port excitation, radiation boundary and sweep frequency range of the three-dimensional structure are set in HFSS. The magnetic permeability of the magnetic material is 120 and the magnetic loss is 0.04. The hybrid finite element fast algorithm introduces tree-cotree technology and Arnoldi mode reduction technology on the traditional finite element method, which can quickly and accurately analyze the frequency response problem of the multi-scale model with dual-dispersion material properties. The hybrid finite element fast algorithm is used to calculate the simulation results of the smart watch antenna with / without magnetic material structure. The hybrid finite element fast algorithm is used to solve the smart watch antenna with magnetic material. This algorithm has both accuracy and efficiency in electromagnetic simulation of multi-scale models with dual-dispersion materials. Modal reduction and hybrid finite element method have obvious advantages in dealing with conductor loss, dielectric loss and magnetic loss. The hybrid finite element fast algorithm is simpler in electromagnetic simulation of such models, and can also significantly reduce the calculation time and space complexity of the model.
[0182] The present invention proposes a rapid modeling and simulation method for devices with dual-dispersion materials. Human tissue, smart watches, PCB inside the dial, antenna and other structures are built in HFSS. HFSS is used to analyze the influence of the presence or absence of magnetic materials on the performance of the smart watch antenna, and a hybrid finite element fast algorithm is used to solve the problem. The hybrid finite element fast sweep frequency method introduces tree-cotree technology and Arnoldi modal reduction technology on the traditional finite element method. This algorithm maintains accuracy and stability in electromagnetic simulation of multi-scale models with tens of millions of degrees of freedom and dual-dispersion material devices, and has advantages in calculation speed. Therefore, the hybrid finite element fast algorithm can speed up the iterative design scheme and quickly approach the optimal design, which can significantly reduce the product development cycle and improve development efficiency.
[0183] The above description is only a preferred embodiment of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present invention (but not limited to) to form a technical solution.
Claims
1. A rapid modeling and simulation method for a device with dual-dispersion materials, characterized in that: The method comprises the following steps: Construct the 3D structure of the smart watch in HFSS and set its material parameters; build the PCB, antenna and 3D structure with dual-dispersion material properties in the watch dial, where the magnetic material has a magnetic permeability of 120 and a magnetic loss of 0.04, adjust the size of the human hand structure and insert it into the watch strap; this model is a multi-scale model, the maximum size of its structure is 320.2mm, the minimum size is 0.02mm, and the multi-scale ratio is 16010; Mesh the human tissue and smart watch antenna models in HFSS, and use the hybrid finite element fast algorithm to analyze the impact of magnetic material structure on the performance of smart watch antennas; The HFSS discrete mesh data is read and solved using the hybrid finite element fast algorithm. This algorithm is accurate and efficient in calculating multi-scale models with tens of millions of degrees of freedom and dual-dispersion material properties.
2. The rapid modeling and simulation method of a dual-dispersion material device according to claim 1, characterized in that: Import the CAD file of the smartwatch into HFSS, set the strap material to plastic, the watch metal case material to aluminum, and the watch display material to glass.
3. The rapid modeling and simulation method of a dual-dispersion material device according to claim 2, characterized in that: The PCB, antenna and structure with dual-dispersion material properties are placed inside the watch dial, and the human hand structure is adjusted and inserted into the watch strap, wherein the material of the human hand is a dispersive material with an average dielectric loss of 0.5; the material properties of the human tissue and the smart watch antenna are set in HFSS, ABC absorption boundary conditions are added, and the lumped port and electrode excitation information are set.
4. The rapid modeling and simulation method of a dual-dispersion material device according to claim 3, characterized in that: Using the HFSS mesh, the hybrid finite element fast algorithm is used to calculate the S parameters of the smart watch antenna with / without magnetic material property structure, and the influence of magnetic materials on the performance of the smart watch antenna is analyzed.
5. The rapid modeling and simulation method of a device with dual-dispersion material according to claim 4, characterized in that: The human tissue and smart watch antenna are meshed discretized in HFSS, and the hybrid finite element fast algorithm is used to perform electromagnetic simulation of multi-scale human tissue and smart watch antenna with dual-dispersion material structure.
6. The rapid modeling and simulation method of a device with dual-dispersion material according to claim 5, characterized in that: Based on the differential form of Maxwell's equations, the electric field wave equations to be solved for the human tissue and smart watch antenna model in the frequency domain of the electromagnetic field are obtained: in, is the Hamiltonian operator, μ is the magnetic permeability, E is the electric field intensity, v is the dielectric constant, σ e represents the conductivity, J i is the current source density, s represents the complex frequency, that is, s=jω, where j refers to the imaginary unit and ω is the angular frequency. in, is the Hamiltonian operator, μ is the magnetic permeability, E is the electric field intensity, ∈ is the dielectric constant, σ e represents the conductivity, J i is the current source density, s represents the complex frequency, that is, s=jω, where j refers to the imaginary unit and ω is the angular frequency. In the hybrid finite element fast algorithm, the electric field intensity E is decomposed by the edge basis function Φ defined on the cotree edge through the tree-cotree decomposition technique. c and the node basis functions defined at the free nodes The gradient expansion is expressed as follows: Among them, N c is the number of cotree edges in the grid, N n is the number of free nodes in the grid, refers to the i-th edge basis function defined on the cotree edge, It refers to the gradient of the jth node basis function defined at the free node, and e represents the corresponding unknown coefficient.
7. The rapid modeling and simulation method of a device with dual-dispersion material according to claim 6, characterized in that: The conductor loss of human tissue and smart watch antenna is analyzed using the hybrid finite element fast algorithm. First, the metal structure in the watch is modeled using the impedance boundary condition (IBC), and the expression is as follows: n×H=jωY s (n×(E×n)) Where n is the normal vector of the metal surface, H is the magnetic field, and Y s is the surface admittance of the metal conductor, expressed as The IBC boundary accurately describes the relationship between the tangential electric field and the tangential magnetic field on the metal surface. The use of the IBC boundary avoids solving the interior of the metal, thereby reducing the amount of calculation. Substitute the IBC boundary into the matrix equation to be solved and use the Galerkin method to obtain the final system matrix expression as follows: Where b is the input matrix of the entire calculation area, and the expressions of the mass matrix M, damping matrix C, stiffness matrix K and boundary term matrix Z are as follows: in<A,B> Ω It means that A and B are inner products within the range of Ω. Let s = s0 (1 + δ), where s0 is the reference complex frequency, generally the complex center frequency. The matrix equation to be solved after expansion is: A0e+δA1e+δ 2 A2e=-s0(1+δ)b In the above equation, The approximate equivalent is 1+0.5δ, where 8. The rapid modeling and simulation method of a device with dual-dispersion materials according to claim 7, characterized in that: The expanded second-order matrix equation to be solved is reduced in order. The second-order equation is solved based on the second-order Arnoldi algorithm, where the dimension of the inverse matrix equation is consistent with the first-order equation, so the computational complexity is almost the same as that of the linear equation, and it is suitable for solving electromagnetic field problems with multiple inputs and outputs. By adding auxiliary equations, the expanded matrix equation to be solved is converted into a system of linear equations: In addition, the above formula can be restated as: Where A = -A0 -1 A1,B=-A0 -1 A2. Therefore, the system transfer function matrix H(δ) of the matrix equation to be solved after the equation expansion is defined as: H(δ)=(I-δF) -1 B0 in, Here, H(δ) can be further expanded into the following power series: The specific process of constructing a low-dimensional Krylov subspace to approximate the transfer function is as follows: First, a Krylov subspace is selected as the projection subspace, which is defined as: K m (F;B0)=span{B0,FB0,F 2 B0,...,F m-1 B0} Second, at each iteration of the Arnoldi algorithm, there is this relationship: Therefore, the Krylov subspace K can be immediately derived m (F; B0) can be constructed by a standard orthogonal basis, namely: in, Let Q = {R0, R1, R2, ..., R m-1 } The projection subspace Q is generated by using a modal order reduction technique based on a second-order Arnoldi algorithm, wherein the second-order Arnoldi algorithm specifically includes: S31, Q0=orth(Q0)=[q1,...,q M ], where M represents the number of columns of the right-hand side. P0 is a zero matrix with the same dimension as Q0, and orth represents orthogonalization of the matrix in the brackets; S32, perform cyclic assignment, S321, define the value of variable j from 1 to m; S322, let r=A1q j +A2p j , G T r=0, t=q j ,in I represents the identity matrix, and 0 represents the zero matrix; S33, define the value of variable i from 1 to j+M-1; S331, order in sequence r=rq i h i,j , t = tp i h i,j ; S332, when h j+l,j = norm(r), norm() means calculating the 2 norm of r, q j+l =r / h j+l,j , p j+l =t / h j+l,j When the loop ends; S34. The loop ends. Expand the unknown vector e using the orthogonal basis Q, that is, Where w=[w0,w1,…,w m-1 ] is the coefficient vector. Substitute the above formula into the matrix equation to be solved, and multiply both sides of the matrix equation to be solved by the transpose Q T , and the reduced matrix equation is obtained: where m is the number of iterations in the second-order Arnoldi algorithm. and The expressions are By solving the reduced-order matrix equation at each frequency to obtain the coefficient vector w, the electric field at each frequency can be obtained. The dimension of the reduced matrix is determined by the number of basis vectors in the projection space, which is usually smaller than the dimension of the original matrix. Therefore, it is easier to obtain the frequency response at each frequency by solving the reduced matrix equation. In addition, the second-order Arnoldi algorithm does not include Schmidt orthogonalization and eigenvalue calculation, which improves the computational efficiency of the projection subspace.
9. The rapid modeling and simulation method of a dual-dispersion material device according to claim 8, characterized in that: The dielectric loss and magnetic dielectric loss of human tissue and smart watch antenna are analyzed using a hybrid finite element fast algorithm. The relative dielectric constant ∈ r and relative magnetic permeability μ r Written in the form of real and imaginary parts respectively, the expressions are as follows: ∈ r =∈′-je″=e′(1-jtanD) m r =μ′-jμ″=μ′(1-jtanM) Among them, ∈′ is the real part of the relative dielectric constant, tanD is the electric loss tangent, μ′ is the real part of the relative magnetic permeability, and tanM is the magnetic loss tangent. In a wider frequency band, ∈′, tanD, μ′ and tanM appear to be constants or frequency-dependent. The vector fitting method can be used to fit ∈ r and Make a rational approximation, then ∈ r and It can be expressed as: ∈ r =k ∈ +x e (oh) Among them, k ∈ and k μ is a constant term, χ e (ω) and C m (ω) is a frequency-dependent function, and its expressions are where a p and c p They represent the residue of the fitting function, b p and d p Represent the poles of the fitting function, N e and N m represents the fitting order. Then, r The associated mass matrix M is divided into two parts, one part M1 and the constant term k ∈ The other part M2 is related to the frequency-related term χ e (ω), the expressions of M1 and M2 are as follows: Similarly, The relevant stiffness matrix K is divided into two parts, one part K1 and the constant term k μ The other part K2 is related to the frequency-related term C m (ω), K1 and K2 are expressed as follows: Among them, ∈0 and μ0 are the dielectric constant and magnetic permeability in vacuum, respectively. Generally, for lossy media, the constant terms after fitting the relative dielectric constant and relative magnetic permeability account for a large proportion, resulting in M1 being much larger than M2, and K1 being much larger than K2. Therefore, before reducing the matrix equation, only M1 and K1 can be considered, and the influence of M2 and K2 can be added to the matrix reduction equation. Taking conductor loss, dielectric loss and magnetic loss into consideration, the final system matrix equation is obtained, which is expressed as follows: Through the projection subspace Q, we can obtain e=Qw, and substitute it into the electric field expression. Get the electromagnetic fields of human tissue and smart watch antenna at each solved frequency.
