A novel laminated power-law model for bending analysis of functionally graded plates with arbitrary material gradients Authors: Shuping Chen, Jinwen Wu, Hongxiao Zhao Published: 2025 ABSTRACT AbstractA novel laminated model for analyzing the bending problem of functionally graded materials with arbitrary characteristic variations is developed. Based on the mathematical principle that an arbitrary curve can be approximated by a series of continuous, piecewise power-law curves, a functionally graded material with arbitrarily continuous material property is modeled as a multi-layer medium. In this model, the elastic modulus within each layer follows a general power-law distribution, ensuring continuity of stress and displacement at the interfaces between adjacent sub-layers. A theoretical method for analyzing the bending problem of a simply supported functionally graded rectangular plates under arbitrary loadings is established using this model. Taking a rectangular plate subjected to double trigonometric loadings as an example, all displacement and stress components are obtained. Parameterized analysis shows that the change in the number of layers has an impact on the convergence of the calculation results, revealing a quantitative relationship between the number of layers and computational accuracy; Under the same layering conditions, the approximation effects of different gradient distribution models on the exact solution are compared. This proposed laminated model has excellent convergence characteristics and computational efficiency, providing a new and effective approach for the mechanical analysis of functionally graded material plate structures, and has important application value in fields such as aerospace.Keywords: Functional graded materials, Bending analysis, Laminated model, Power-law distribution, Convergence analysisSubject terms: Engineering, Materials science, Mathematics and computing Introduction Functionally graded materials (FGMs) are heterogeneous composite materials characterized by a gradual change in properties within the specimen as a function of the position coordinates. The property gradient in the material is typically caused by a position-dependent chemical composition, microstructure1. The advantage of FGMs is to eliminate interfacial stress jumps in composite materials and make the stress distribution smoother. Based on these characteristics, the application of FGMs has become more and more extensive in various engineering fields2–5. With the increasing of the application of FGMs, many studies have focused on the response of functionally graded rectangular plates and shells to different loads. A simplified theory and numerical techniques have been used to study functionally graded plates and shell structures6–19. Another type of method based on elastic theory involves laminated theory and exact solution methods20–33. For functionally graded plates with arbitrarily distributed material properties, the laminated theory is commonly employed for analysis. Since the material parameters of functionally graded plates generally vary along the thickness direction, a composite laminated structure model is used to divide the functionally graded structure into multiple layers along this direction. Assuming that the material parameters are constant in each thin layer, but the material parameters of adjacent layers are slightly different, each thin layer can be analyzed using classical structural theory, and then solved as a whole by combining boundary conditions and interface continuity conditions. When there are enough layers, the analysis results will approach the exact solution of elasticity. Based on the laminated models, Chen and Ding34,35 employed the state-space method to analyze the bending and free vibration of functionally graded rectangular plates. Chen et al.36,37 further investigated the free vibration of functionally graded spherical shells and the three-dimensional vibrations of hollow functionally graded piezoelectric cylinders filled with compressible fluid. Additionally, Liu et al.38 applied a hybrid numerical method to study wave propagation in functionally graded piezoelectric plates. The laminated theory approach offers several advantages: it leverages well-established composite material theories, provides computational simplicity, and demonstrates relatively fast convergence. However, this method also has obvious limitations: firstly, to obtain sufficiently accurate solutions, it often requires a large number of layers, which significantly increases the computational workload, especially when the gradient changes dramatically; Secondly, due to the use of a layered discrete modeling approach, the material parameters between layers are discontinuous. In recent years, Huang and Wang39–41 have proposed a new laminated model based on piecewise linear continuous approximation, which assumes that the elastic modulus and mass density of the material in each layer vary linearly and are continuous at the sub-layer interface. Using this model, they successfully solved several fracture mechanics problems in functionally graded materials (FGMs), overcoming the limitations of conventional laminated theory in terms of arbitrary gradient simulation, dynamic load adaptability, and computational efficiency. It is noteworthy that this method can also be extended to solve bending problems of functionally graded material plates. It is feasible to assume that the elastic modulus follows a general power function distribution instead of the linear distribution assumption for layered elastic modulus. The proposed power-law model demonstrates significant improvements over piecewise-linear descriptions in two critical aspects: (i) Superior Physical Fidelity: It ensures the continuity of both Young’s modulus E(z) and its derivative dE/dz, thereby eliminating non-physical interfacial stress jumps and providing a more realistic mechanical response. (ii) Enhanced Simulation Capability: The parameters in the model provide great flexibility to simulate a wide range of nonlinear behaviors. This flexibility makes our model universally applicable and highly adaptive for modeling truly arbitrary material gradients. This paper investigates the bending problem of functionally graded rectangular plates with arbitrary gradient changes in material parameters along the thickness direction, and proposes a new laminated theoretical model. The developed model discretizes the functionally graded plate into multiple sub-layers, where the elastic modulus in each sub-layer follows a general power-law distribution while maintaining a constant Poisson’s ratio. By progressively increasing the number of layers, the mechanical response of the laminated plate gradually converges to the exact solution of the original functionally graded plate. Through numerical examples, all displacement and stress components are obtained for a simply supported functionally graded rectangular plate subjected to bi-trigonometric loading. The influence of the number of layers on the calculation accuracy in the proposed model is investigated, and the convergence characteristics of the method are revealed; Additionally, under the same number of layers, the approximation degree of different gradient distribution models to the exact solution is compared. The description of the problem and the basic equations Consider a simply supported functionally graded material (FGM) rectangular plate, with length , width , and thickness , as shown in Fig. 1. (a) simply supported functionally graded rectangular plate. The laminated model of functionally graded material is established as follows, and it is assumed that the Young’s modulus of each layer is of the general power-law distribution form. where and are two material constants which describe the non-uniformity of E(z). The new model can be employed to fit an arbitrary gradient profile, as shown Fig. 2. Fitting curve of gradient variation of functional graded rectangular plate. is the elastic constant for each interface, that is , where. Hence, for any given value of the gradient distribution parameter p, all the constants and can be determined. In the absence of body forces, the equilibrium equations in terms of displacement can be expressed as where ,, are Cartesian coordinates, ,, are displacement components, is the volumetric strain and is the Laplcian operator, and are respectively the Young’s modulus and the Poisson’s ratio, which are functions of the coordinate z. The boundary conditions at the top and bottom surfaces of the plate can be expressed as. where , and , , , represent the transverse and tangential loads acting on the upper and lower surfaces, respectively. On the simply supported edges, the following boundary conditions can be given as For laminated plates, the interface displacement and stress continuity conditions between layers are introduced, that is: The process of solving problems Introduction of Plevako solution According to the Plevako solution42, the components of the displacement can be expressed by two functions and as follows where satisfy the following two partial differential equations with variable coefficients where . The stress components can also be expressed in terms of the functions and , as follows Write and in terms of the separated variables as follows where m and n are positive integers. Substituting Eqs. (11) into Eqs. (8) and (9) yields two governing differential equations for unknown functions and where The general solutions of and can be obtained by solving the two ordinary differential equations of the variable coefficients respectively, and then and can be given, and the all displacement and stress components of the rectangular plate of the functionally graded material can be obtained by substituting into Eqs. (7) and (10). General power-law distribution Substituting (1) into (13) and (14), we have The solutions of Eq. (16) are expressed as. for and for or The solution of Eq. (17) is expressed as with. where are Whittaker functions, and are the gth order modified Bessel functions of the first and second kind respectively, and are the order modified Bessel functions of the first and second kind respectively. The coefficients , , , , , can be determined from boundary conditions (4). Detailed solution Now, We present a detailed solution for the case where Young’s modulus follows a general power-law distribution. By substituting Eqs. (11), (18), (19), and (20) into Eqs. (7) and (10), and performing summation over different m and n, we have All displacement and stress components can be expressed in terms of six undetermined constants 、、、、、, as follows where both () and () are function of . Written as. with By substituting the last three equations of (24) into the boundary conditions (4) at the upper and lower surfaces, we obtain a system of linear algebraic equations for the undetermined constants ,,,,,. Once these constants are determined, all displacement components and stress components of the functionally graded rectangular plate under arbitrary loading can be obtained. For the i-th layer, we have At the upper and lower surfaces of each layer, we have The transmission relations between each state quantity of the upper and lower surfaces of each layer are obtained as follows By using the interfacial continuity conditions (6), we have with Thus, the state variables at any interface can be fully determined, and the undetermined constants for each layer can be correspondingly obtained through the following recursive relations: Consequently, all state variables at arbitrary points within the plate can be expressed analytically. Numerical examples and discussion Consider a simply supported functionally graded material (FGM) plate with length m, width m, and a thickness of m. The plate is subjected to a double trigonometric distributed loading MPa on the top surface, while the bottom surface remains . Since the loading is expressed in the form of a double trigonometric series, it suffices to consider only the first term () in the previously derived solution. To examine the convergence of the laminated plate model, we assume the Young’s modulus of the material follows a power-law distribution through the thickness direction: where represents the Young’s modulus at the midplane () of the functionally graded plate, taken as MPa. Tables 1,2,3 and 4 list the approximation of displacement components, stress components, and the precise solution at a point (, , ) within the functionally graded plate as the number of layers increases under different gradient models. To quantify the convergence of the laminated model for the functionally graded plate, a relative error is introduced and defined as follows: Table 1Analysis of calculation accuracy of laminated plate model().,N = 1−1.6290e-011−5.9031e-011−0.039278−0.24569−0.523680.95569N = 2−7.5435e-012−2.4289e-011−0.019359−0.12876−0.336940.61743N = 3−6.2132e-012−1.9807e-011−0.017228−0.12124−0.383370.70458N = 5−5.4714e-012−1.7476e-011−0.017022−0.12790−0.426880.78547N = 10−5.1623e-012−1.6476e-011−0.017008−0.12690−0.422170.77674N = 20−5.0836 e-012−1.6224e-011−0.017003−0.12689−0.428790.78903Exact solution26−5.0468e-012−1.6109e-011−0.017003−0.12676−0.425690.78328 Analysis of calculation accuracy of laminated plate model(). Table 2Analysis of calculation accuracy of laminated plate model().,N = 1−8.9823e-013−3.8795e-012−0.026547−0.22753−0.821671.51457N = 2−3.2274e-012−1.0894e-011−0.019648−0.15481−0.514070.94628N = 3−4.1186e-012−1.3339e-011−0.018079−0.13566−0.420110.77244N = 5−4.6794e-012−1.4956e-011−0.017109−0.12669−0.439440.80876N = 10−4.9548e-012−1.5816e-011−0.017010−0.12698−0.431050.79323N = 20−5.0311 e-012−1.6057e-011−0.017004−0.12689−0.424360.78081Exact solution26−5.0468e-012−1.6109e-011−0.017003−0.12676−0.425690.78328 Analysis of calculation accuracy of laminated plate model(). Table 3Analysis of calculation accuracy of laminated plate model().,N = 1−1.8163e-012−6.6760e-012−0.016192−0.14656−0.632361.16744N = 2−4.0047e-012−1.3091e-011−0.017708−0.13911−0.487830.89838N = 3−4.5500e-012−1.4612e-011−0.017416−0.13140−0.429210.78964N = 5−4.8620e-012−1.5524e-011−0.017042−0.12667−0.431380.79383N = 10−5.0055e-012−1.5976e-011−0.017006−0.12692−0.428840.78912N = 20−5.0441e-012−1.6099e-011−0.017004−0.12689−0.425460.78285Exact solution26−5.0468e-012−1.6109e-011−0.017003−0.12676−0.425690.78328 Analysis of calculation accuracy of laminated plate model(). Table 4Analysis of calculation accuracy of laminated plate model().,N = 1−1.1405e-011−3.6579e-011−0.028827−0.18889−0.464210.84975N = 2−6.2893e-012−1.9955e-011−0.017728−0.12382−0.367320.67456N = 3−5.6215e-012−1.7893e-011−0.016940−0.12318−0.409540.75330N = 5−5.2618e-012−1.6795e-011−0.016996−0.12731−0.425180.78233N = 10−5.1096e-012−1.6307e-011−0.017004−0.12688−0.434320.78073N = 20−5.0704e-012−1.6182e-011−0.017003−0.12689−0.427680.78696Exact solution26−5.0468e-012−1.6109e-011−0.017003−0.12676−0.425690.78328 Analysis of calculation accuracy of laminated plate model(). where x is computed value, is exact solution. Table 5 presents the convergence analysis of the Laminated Power-Law Model for functionally graded rectangular plates. It can be observed that displacement components (u, v, w) have the slowest convergence and are highly sensitive to the number of layers(N), requiring N = 10 or even N = 20 to achieve extremely high accuracy below 1%. The convergence speed of Stress components (σ, τ) is much faster than the displacement, with the majority of errors below 1% at N = 5, and many achieving this accuracy at N = 3. The convergence speed and accuracy of high-order models (p = 1,2) are generally better than those of low order models (p = −1, −2), particularly for displacement components, which verifies the effectiveness of the high-order modulus gradient distribution. At N = 20 layers, all physical quantities are fully converged, demonstrating the final convergence and reliability of the model. Table 5Convergence Analysis of the Laminated Power-Law Model.Physical QuantityPower-Law Exponent (p)N = 1 relative ErrorN = 2 relative ErrorN = 3 relative ErrorN = 5 relative ErrorN = 10 relative ErrorN = 20 relative Erroru, v−1223%49%23%8.4%2.3%0.7%−282%36%18%7.3%1.8%0.3%194%21%10%3.7%0.8%0.05%2126%25%12%4.3%1.2%0.5%w−1267%51%23%8.5%2.3%0.7%−2141%32%17%7.2%1.8%0.3%1315%19%9%3.6%0.8%0.06%2127%24%11%4.3%1.2%0.5%σz−1131%14%1.3%0.1%0.03%0.0%−256%16%6%0.6%0.04%0.06%14.7%4.1%2.4%0.2%0.02%0.06%270%4.3%2.4%0.04%0.006%0.0%τzx τzy−194%1.6%4.3%0.9%0.1%0.1%−280%22%7%0.05%0.2%0.1%180%9%7%0.07%0.1%0.1%249%2%4%0.4%0.09%0.1%τxy−123%21%10%0.3%0.8%0.8%−293%21%1.3%3.2%1.3%0.4%149%21%1%1.3%0.7%0.4%29%17%1%0.1%2.0%0.5%σx σy−122%21%10%0.3%0.8%0.7%−293%21%1%3.3%1.3%0.5%149%21%1%1.3%0.8%0.4%28%14%1%0.1%0.3%0.5% Convergence Analysis of the Laminated Power-Law Model. Figure 3 show the distributions of stress and displacement components along the thickness direction of the functionally graded plate at a chosen position (,) with different layers for . The convergence of various physical quantities can be visually observed from the graph. As the number of layers (N) increases, the calculated value curve gradually approaches the exact value curve. It is worth mentioning that the normal stress inside the plate is continuously distributed, unlike the results of the uniform layered model which are discrete. This characteristic represents a significant advantage of the new laminated model. Comparison of various physical quantities and exact solutions under different layers(). For , the calculated values of in-plane displacement u and v are lower than the exact values at most points along the thickness of the plate, except for positions near the bottom of the plate where they exceed the exact values. As the number of layers (N) increases, the calculated displacement u and v at points near the top surface of the plate gradually increase, while those near the bottom surface decrease, moving closer to the exact solution. In contrast, the calculated value of the out-of-plane displacement w are consistently lower than the exact values along the entire thickness of the plate, but they gradually increase towards the exact value as the number of layers increases. Figure 4 and 5 show the distribution of stress and displacement components along the thickness direction of the functionally graded plate at a chosen position (,), and the comparison with the exact solution. The analysis considers two cases with different layer numbers ( and ) for and , respectively. Comparison of various physical quantities and exact solutions under different gradient models(N = 2). Comparison of various physical quantities and exact solutions under different gradient models(N = 10). From Fig. 4 and 5, it can be seen that the calculated values of in-plane displacement and out-of-plane displacement for are closer to the accurate values than . This indicates that the high-order form of the modulus gradient distribution in each layer of the proposed laminated model is more effective in simulating the actual arbitrary modulus gradient distribution of the functionally graded plate compared to the low-order form or linear form. Furthermore, displacement components exhibit greater sensitivity to the total number of layers (N) than stress components. Notably, the convergence rate of displacement is significantly slower, with out-of-plane displacement w displaying the slowest convergence. This implies that while accurate stress predictions can be achieved with relatively few layers, further increasing N yields only marginal improvements in stress accuracy. In contrast, obtaining precise displacement results (particularly for w) demands a substantially finer layer discretization. In summary, the proposed power-law laminated model provides a new and effective method for the mechanical analysis of functionally graded material plate structures, and has important application value in fields such as aerospace. It has the following characteristics: (1) By adopting a general power-law distribution for the Young’s modulus within each sub-layer and ensuring its continuity (including the function value and its first derivative) at the interfaces, the new model effectively eliminates non-physical stress jumps. This results in a continuous distribution of stress components through the thickness, which aligns more closely with the actual mechanical behavior of Functionally Graded Materials (FGMs). (2) Compared to traditional laminated theory, the proposed power-law laminated model achieves high computational accuracy and rapid convergence, with results deviating from the exact solution by less than 0.7% even with a small number of layers (e.g., N = 5 or 10). Especially suitable for scenarios with drastic gradient changes. (3) The new model examines various power-law distribution forms (such as p = −1, −2, 1, 2) and can adapt to different material gradient requirements. By adjusting the number of layers and power-law parameters, complex material gradient distributions can be flexibly simulated. (4) The convergence characteristics of displacement and stress components differ significantly. Accurate stress predictions can be achieved with a relatively small number of layers (N), and further refinement yields only marginal gains. In contrast, obtaining precise displacement results, especially for the out-of-plane displacement (w), requires a substantially finer layer discretization due to its notably slower convergence rate. The high-order power-law form (e.g., p = ± 2) demonstrates superior accuracy in approximating the exact solution compared to the low-order form (e.g., p = ± 1), confirming its enhanced capability to simulate complex, arbitrary material gradients effectively. (5) The analysis mainly focuses on the static bending problem of four sided simply supported functionally graded plates, but the performance in more complex situations, such as dynamic loads, other boundary constraints, or coupled thermo mechanical effects, has not been explored yet. In addition, the assumption of a constant Poisson’s ratio may limit the applicability of the model to certain real-world FGM structures. Author contributions Contributions of author S.C: methodology, numerical simulation, and preparation of the first draft of the manuscript. Contribution of author J. W: assist in numerical simulation and conclusion analysis. Contribution of author H.Z: assist in numerical simulation and conclusion analysis. Funding This work was supported by the Fundamental Research Program of Shanxi Province (No. 202403021221119), and the National Natural Science Foundation of China (No.51504220). Data availability The datasets used and/or analysed during the current study available from the corresponding author on reason-able request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analysed during the current study available from the corresponding author on reason-able request.