A High-Precision Non-Classical Framework for Fractional Rheology and Viscoelastic Nanobeam Dynamics

Fractional calculus is a more expressive mathematical approach that can represent complex material behavior that traditional models oversimplify or fail to capture. In practice, though, this promise has unfolded unevenly. Many complex solids, whether they are soft polymers, biological tissues, or engineered nanostructures, evolve in time in a way that retains traces of their past. Their deformation does not respond to loading instantaneously but accumulates and redistributes over long periods, making the memoryless assumptions of conventional viscoelastic models feel increasingly inadequate. Although the Maxwell and Kelvin–Voigt viscoelastic models are mathematically neat, they depend on simple, local derivatives and spring–dashpot analogies. Because of this, they often miss or oversimplify the gradual changes and long-range (nonlocal) interactions that real materials actually experience. Fractional derivatives, by contrast, embed memory directly into their kernels and give researchers a way to describe creep, relaxation, and recovery in a manner that aligns more closely with experimental trends. This has become especially relevant as nanoscale systems have moved from abstraction to routine fabrication; at those scales, size-dependent stiffness and nonlinear damping emerge naturally and demand a model capable of accommodating such effects. Yet adopting fractional calculus in computational practice remains challenging. Because the fractional operator reaches back through the full temporal history, straightforward numerical schemes accumulate ever-growing arrays of past values. Both runtime and memory demands can balloon, particularly when solving long-time dynamics or multi-element fractional constitutive equations.

Several research groups have tried to bypass this difficulty with non-classical approaches—memory-free formulations, hierarchical matrices, or compressed operator representations. These methods help, but they often stumble when the fractional kernels decay only algebraically. Slow decay leads to numerical stiffness, and convergence can deteriorate at the very moment high accuracy is required. This ongoing tension between fidelity and feasibility remains a significant obstacle for those hoping to deploy fractional models in realistic simulations of rheology or structural vibration. To this end, new research paper published in International Journal of Solids and Structures and conducted by Dr. Tian-Ming Liu, Dr. Yan-Mao Chen, Dr. Ji-Ke Liu, and led by Professor Qi-Xian Liu from the Sun Yat‐sen University, the researchers developed a high-precision non-classical computational method that reformulates fractional derivatives through interval decomposition and enhanced infinite-state representation. By enforcing exponential decay in the transformed kernels, the method achieves linear computational complexity while retaining the accuracy of classical schemes.

The research team reformulated the Caputo fractional derivative into two complementary integrals—one covering the long-range historical interval and the other confined to a short, recent window. This decomposition provided a lever for controlling the decay properties of the integrand. The long-interval component was transformed into an improper integral and then mapped, through a change of variables, into a form compatible with infinite-state representation. The authors obtained an integrand whose exponential decay allows accurate quadrature using only a modest number of nodes by tuning the transformation parameter. In parallel, the short-interval portion was evaluated through piecewise quadratic interpolation, so as to ensure computational accuracy while balancing computational efficiency. They found the new system no longer required storing the entire solution history. Instead, it evolved a set of auxiliary delay-differential equations whose dimensionality remained linear with respect to the number of quadrature points, preventing the explosive growth characteristic of earlier non-classical methods.

When they applied the new computational structure to multi-element fractional rheology models, the method reproduced creep, relaxation, and recovery behaviors with accuracy comparable to classical predictor–corrector algorithms, yet at a fraction of the computational cost. The convergence estimates associated with the Gauss–Laguerre and Gauss–Jacobi quadratures aligned with the observed behavior: as the integrand decayed exponentially, the truncation error dropped quickly, and only a small subset of quadrature nodes meaningfully contributed to the solution. This selective contribution further reduced the number of auxiliary equations needed in practice. The team then extended the framework to nonlinear viscoelastic vibration of nanobeams. They began by formulating the governing equations using fractional Kelvin–Voigt constitutive relations within a nonlocal strain-gradient framework to capture size-dependent effects. After nondimensionalization and modal discretization through Galerkin’s method, they obtained a reduced-order system containing both integer-order and fractional-order components. Embedding the new fractional solver into this system allowed them to compute forced vibrations with strong geometric nonlinearity. Their simulations revealed distinct fractional-order damping characteristics, including amplitude-dependent decay and shifts in resonant behavior. Finally, the authors performed rheological experiments on resin materials to ensure that the algorithm’s predictions was not numerical artifacts and by fitting the fractional three-element or four-element models to the measured creep and recovery curves, they confirmed that the non-classical method preserved the curvature and long-time behavior observed experimentally.

In conclusion, the new work of Professor Qi-Xian Liu developed a new model that supports both multi-element fractional rheology models and nonlinear viscoelastic vibration of nanobeams, remaining stable over long time windows. This unified framework offers a practical route for incorporating fractional calculus into large-scale simulations and experimentally validated material analyses. Moreover, by producing a formulation whose complexity grows linearly rather than quadratically or exponentially, the authors have effectively transformed fractional modeling from a specialized research tool into an approach that can be realistically embedded within large-scale simulations. The ability to maintain high accuracy while avoiding the weak-decay bottleneck marks a shift in how fractional derivatives can be treated numerically. The implications for rheology are immediate. Many polymeric and biological materials display multiscale relaxation patterns that cannot be reduced to simple exponential modes. Traditional models, even when augmented with multiple spring-dashpot elements, fail to capture these curves without resorting to cumbersome parameterizations. The present approach enables the routine use of multi-element fractional constitutive relations, allowing fits that more faithfully represent experimental measurements. Since the method remains robust over long time windows, it offers a reliable pathway for studying processes such as creep recovery, long-term stress relaxation, and aging phenomena in soft matter.

Moreover, the method becomes even more valuable when applied to systems that move or vibrate. At the nanoscale, structures like tiny beams, membranes, and resonators behave in complicated ways because several effects occur at the same time. Predicting their behavior requires a solver capable of handling fractional damping terms and nonlocal operators without collapsing under computational strain. The authors’ framework demonstrates that such coupling can be simulated efficiently, and can open opportunities to explore regimes that were previously inaccessible including nonlinear vibration suppression, resonance tuning, and the design of nanoscale components whose damping properties depend on fractional mechanisms. Additionally, the method provides a blueprint for tackling other fractional differential equations that arise in diffusion, electrochemistry, biomembrane mechanics, and anomalous transport. The decomposition and quadrature strategies can, in principle, be adapted to any system where fractional kernels exhibit slow decay. As a result, the study may encourage a reevaluation of problems previously considered too costly to simulate in full fractional order.