Adaptive Smagorinsky Models: Mathematical Theory and Dynamics of Subgrid-Scale Turbulence

Adaptive Smagorinsky Models: Mathematical Theory and Dynamics of Subgrid-Scale Turbulence Rˆomulo Damasclin Chaves dos Santos Santa Cruz State University [email protected] Jorge Henrique de Oliveira Sales Santa Cruz State University [email protected] October 27, 2025 Abstract This paper presents a rigorous mathematical analysis of the adaptive Smagorinsky model in Large-Eddy Simulation (LES) of turbulent flows, addressing the limitations of traditional constant-coefficient approaches. The model introduces a spatially varying dissipation coefficient, CS ( x ) = CS0e−βΦ(x) , designed to reduce excessive near-wall damping while enhancing the regularity and stability of solutions. Using Galerkin approximations, we establish global existence, uniform energy bounds, and the presence of finite-dimensional attractors in weak formulations under suitable assumptions on Φ( x ). The spatially modulated dissipation mechanism improves the physical fidelity of LES by adapting to local flow features, particularly in regions of high shear and near solid boundaries. Our analysis leverages functional analysis, spectral theory, and partial differential equations to derive robust estimates and structural properties of the turbulence operators. We prove the existence and uniqueness of weak solutions, demonstrate enhanced regularity due to the adaptive dissipation, and establish uniform bounds on energy dissipation rates. The model’s ability to stabilize high-Reynolds-number flows is rigorously validated, providing a theoretical foundation for its application in engineering and environmental simulations. The results highlight the model’s capacity to preserve key flow invariants and improve numerical stability, offering a pathway toward more accurate and theoretically justified turbulence modeling. The study bridges historical foundations with contemporary developments, contributing to both the understanding and practical implementation of adaptive subgrid-scale models in computational fluid dynamics. Keywords: Adaptive Smagorinsky model; Turbulence modeling; Energy dissipation; Regularity; Global attractor. 1 Introduction The accurate modeling and prediction of turbulent flows stand as one of the most enduring and fundamental challenges in fluid mechanics, with far-reaching implications for 1 atmospheric science, engineering design, and computational simulation. Since Smagorinsky’s pioneering work [ 1 ], which introduced one of the first Large-Eddy Simulation (LES) models for atmospheric circulation, the need to represent unresolved scales in numerical simulations has spurred the development of increasingly sophisticated turbulence models. Concurrently, the early numerical strategies proposed by Von Neumann [ 3 ] for handling hydrodynamic shocks underscored the importance of stable, consistent computational frameworks capable of capturing the complexity of flow phenomena. Together, these contributions laid the foundation for the interplay between physical modeling, numerical approximation, and mathematical rigor, a synergy that continues to drive modern research in the field. From a theoretical perspective, the mathematical formulation of fluid motion advanced significantly through the work of Ladyzhenskaya [ 2 ], who introduced modified equations for viscous incompressible flows and examined their global solvability under appropriate boundary conditions. This research established a rigorous basis for analyzing regularized and filtered equations, both of which are central to turbulence modeling. Later, Lilly’s investigations [ 4 ] into the representation of small-scale structures in numerical experiments further emphasized the necessity of subgrid-scale (SGS) parameterizations, bridging phenomenological theories with computational strategies. A major advancement came with Germano’s dynamic subgrid-scale model [ 5 ], which enabled eddy viscosity to adapt locally based on resolved flow features. This dynamic approach markedly improved the physical accuracy of LES and shifted the paradigm for turbulence closure models, grounding them in scale interactions rather than empirical constants. Complementing these developments, Doering’s rigorous estimates [ 6 ] on energy dissipation in body-forced turbulence linked analytical techniques in partial differential equations with fundamental phenomenological insights. These contributions built upon the foundational work of Frisch [ 7 ] and Pope [ 8 ], whose monographs integrated statistical, physical, and modeling perspectives, becoming indispensable references in turbulence research. Recent progress has deepened our understanding of dissipation and enstrophy transfer in turbulent flows. For instance, Layton [ 9 ] derived bounds on energy and helicity dissipation for approximate deconvolution models, providing analytical guarantees for the regularized models widely employed in LES. Similarly, Vassilicos’s comprehensive review [ 10 ] uncovered new insights and unresolved questions about dissipation scaling, challenging conventional interpretations of the Kolmogorov cascade. These findings highlight that even the most fundamental aspects of turbulence theory, such as, the structure of the energy cascade and the universality of dissipation rates, remain only partially understood. Despite these advances, critical challenges persist. Classical turbulence closures often lack rigorous justification, and many models struggle to accurately predict dissipation or preserve essential flow invariants. The interplay between analytical regularization, numerical stability, and physical fidelity remains incompletely characterized. Furthermore, the asymptotic behavior of filtered and deconvolution-based operators, and their influence on the energy spectrum and dissipative scales, demands further exploration from both analytical and computational viewpoints. This work addresses these challenges by developing a mathematically rigorous framework for turbulence modeling and asymptotic analysis. We focus on the structure of dissipation, the stability of subgrid-scale models, and the rigorous validation of regularized and filtered equations. Our approach combines functional analysis, spectral theory, and partial differential equations to derive robust estimates and structural properties of turbulence 2 operators. The significance of this research is twofold. From a theoretical perspective, it strengthens the connection between PDE analysis, operator theory, and turbulence phenomenology, extending classical results, such as, those of Ladyzhenskaya [ 2 ] and Doering [ 6 ] to modern filtered and deconvolution-based frameworks. From an applied standpoint, it informs the design and validation of more accurate LES models, which are critical for high-Reynolds-number simulations in aerospace engineering, atmospheric modeling, environmental prediction, and nuclear reactor dynamics. By integrating physical insights with rigorous mathematical analysis, this study advances both the theoretical understanding and practical modeling of turbulent flows. It bridges historical foundations with contemporary developments, paving the way for more robust, theoretically grounded turbulence models. 2 Mathematical Background 2.1 Functional Setting and Sobolev Spaces Let Ω ⊂R3 be a bounded domain with sufficiently smooth boundary ∂ Ω (e.g., of class C2). We denote by Lp(Ω), 1 ≤p≤ ∞, the Lebesgue spaces endowed with the norms ∥u∥Lp(Ω) =         ZΩ|u(x)|pdx1/p ,if 1 ≤p < ∞, ess supx∈Ω|u(x)|,if p=∞. (1) For an integer m≥0, the Sobolev space Hm(Ω) is defined as Hm(Ω) = nu∈L2(Ω) : ∂αu∈L2(Ω),∀ |α| ≤ mo,(2) and it is equipped with the norm ∥u∥2 Hm(Ω) =X |α|≤m ∥∂αu∥2 L2(Ω),(3) where α= (α1, α2, α3) is a multi-index and ∂α=∂α1 x1∂α2 x2∂α3 x3. For the analysis of incompressible flows, we introduce the following solenoidal (divergence-free) function spaces: H=nu∈L2(Ω)3:∇ · u= 0, u ·n|∂Ω= 0o,(4) V=nu∈H1 0(Ω)3:∇ · u= 0o,(5) where n denotes the outward unit normal on ∂ Ω. The spaces H and V are real Hilbert spaces endowed with the norms ∥u∥H:= ∥u∥L2(Ω),∥u∥V:= ∥∇u∥L2(Ω).(6) It follows from the Poincar´e inequality that these norms are equivalent to the full H1(Ω)-norm on V: ∥u∥L2(Ω) ≤CP∥∇u∥L2(Ω),∀u∈H1 0(Ω),(7) for some constant CP>0 depending only on Ω. 3 We denote by V′ the dual space of V , and by H′ the dual of H . Since H is a Hilbert space, it can be identified with its dual via the Riesz representation theorem, yielding the standard Gelfand triple V ,→H∼ =H′,→V′,(8) with dense and continuous embeddings. In particular, for u∈V and f∈V′ , the duality pairing ⟨f, u⟩V′,V generalizes the L2-inner product: ⟨f, u⟩V′,V = (f, u)L2(Ω) whenever f∈H. (9) This functional setting provides the variational foundation for the Navier–Stokes equations and their subgrid-scale models, such as the Smagorinsky and dynamic LES formulations, where H represents the space of kinetic energy, V encodes viscous dissipation, and V′ serves as the natural framework for external forces and weak derivatives of the velocity field. 2.2 Navier–Stokes Operator and Weak Formulation The incompressible Navier–Stokes equations in the domain Ω ⊂R3are given by                  ∂tu+ (u· ∇)u−ν∆u+∇p=f, in Ω ×(0, T), ∇ · u= 0,in Ω ×(0, T), u|∂Ω= 0,on ∂Ω×(0, T), u(0) = u0,in Ω, (10) where u : Ω × (0 , T ) →R3 is the velocity field, p : Ω × (0 , T ) →R the pressure, ν > 0 the kinematic viscosity, and fa given body force. The Stokes operator. Let P : L2 (Ω) 3→H denote the Leray–Helmholtz projection onto the divergence-free subspace H. The Stokes operator A:D(A)⊂H→His defined by Au =−P∆u, D(A) = H2(Ω)3∩V. (11) The operator A is self-adjoint, positive, and has a compact inverse in H . Hence, there exists a complete orthonormal basis of eigenfunctions {wk}∞ k=1 ⊂D ( A ) and eigenvalues 0< λ1≤λ2≤ · · · → ∞ such that Awk=λkwk,(wk, wj)H=δkj.(12) This spectral structure is crucial for Galerkin approximations and energy estimates. The trilinear form. The nonlinear convective term is represented by the trilinear form b:V×V×V→R, defined as b(u, v, w) = ZΩ(u· ∇v)·w dx, (13) which satisfies the skew-symmetry and continuity properties b(u, v, v) = 0,for all u, v ∈Vwith ∇ · u= 0,(14) |b(u, v, w)| ≤ Cb∥u∥V∥v∥V∥w∥V,(15) where Cb> 0 depends only on the domain Ω. These properties ensure that the nonlinear term is well-defined as an element of V′. 4 Abstract formulation. Projecting (10) onto H by means of P , we obtain the abstract evolution equation in V′:du dt +νAu +B(u, u) = f, (16) where the bilinear operator B:V×V→V′is given by ⟨B(u, v), w⟩V′,V =b(u, v, w),∀u, v, w ∈V. (17) Weak formulation. We seek u∈L2 (0 , T ; V ) with ∂tu∈L4/3 (0 , T ; V′ ) such that, for all v∈Vand almost every t∈(0, T), ⟨∂tu, v⟩V′,V +ν(∇u, ∇v)L2(Ω) +b(u, u, v) = ⟨f, v⟩V′,V .(18) By taking v = u ( t ) in (18) and using the skew-symmetry property (14) , one recovers the standard energy identity: 1 2 d dt∥u(t)∥2 H+ν∥u(t)∥2 V=⟨f(t), u(t)⟩V′,V ,(19) which forms the basis for existence, uniqueness (in 2D), and energy stability results in incompressible flow theory. 2.3 Nonlinear Dissipation Operator In the adaptive Smagorinsky model, the nonlinear eddy-viscosity term introduces an additional dissipative mechanism governed by a spatially varying coefficient. The nonlinear operator S:V→V′is defined by S(u) = −∇ · h(CS(x)δ)2|∇u| ∇ui,(20) where δ > 0 denotes the characteristic filter width, CS ( x ) is the spatially varying Smagorinsky constant, and Φ( x ) is a smooth, nonnegative boundary-sensitive modulation function. Typically, one considers CS(x) = CS0e−βΦ(x),(21) where CS0> 0 is the reference Smagorinsky constant and β > 0 controls the exponential attenuation of eddy viscosity near solid walls. This adaptive structure mitigates the excessive damping present in constant-coefficient models and preserves near-wall flow dynamics. Constitutive relation. To express (20) in the framework of nonlinear monotone operators, define a(x, ξ)=(CS(x)δ)2|ξ|ξ, x ∈Ω, ξ ∈R3×3.(22) Then, for u, v ∈V, ⟨S(u)−S(v), u −v⟩V′,V =ZΩa(x, ∇u)−a(x, ∇v):∇u− ∇vdx. (23) Monotonicity and continuity. From (22) , the mapping ξ7→ a ( x, ξ ) satisfies the following properties for almost every x∈Ω: a(x, ξ)−a(x, η): (ξ−η)≥0,∀ξ, η ∈R3×3,(24) |a(x, ξ)−a(x, η)| ≤ Ca(|ξ|+|η|)|ξ−η|,∀ξ, η ∈R3×3,(25) 5 where Ca = supx∈Ω ( CS ( x ) δ ) 2 . Thus, S is a bounded, hemicontinuous and monotone operator from V into V′ . By the Minty–Browder theorem, such operators are maximal monotone, ensuring the existence of weak solutions for the associated variational problems. Coercivity. Furthermore, for all u∈V, ⟨S(u), u⟩V′,V =ZΩ(CS(x)δ)2|∇u|3dx ≥c0∥∇u∥3 L3(Ω),(26) where c0 = infx∈Ω ( CS ( x ) δ ) 2> 0. This coercivity condition implies enhanced damping in regions of large velocity gradients (high shear), which plays a stabilizing role in the dynamics of the filtered velocity field. Abstract formulation. By combining the Stokes operator (11) and the nonlinear dissipation operator (20) , the adaptive Smagorinsky model can be expressed in the compact variational form: du dt +νAu +B(u, u) + S(u) = f, (27) which extends the classical Navier–Stokes system (16) by incorporating a spatially varying nonlinear diffusion term. This formulation allows the application of monotone operator theory to prove existence, energy stability, and asymptotic regularity of weak and timeaveraged solutions. 2.4 Energy Inequalities and Uniform Estimates Let u(t) be a weak solution of the adaptive Smagorinsky system (27), that is, ⟨∂tu, v⟩V′,V +ν(∇u, ∇v)L2(Ω) +b(u, u, v) + ⟨S(u), v⟩V′,V =⟨f, v⟩V′,V ,∀v∈V. (28) Energy identity. Taking v = u ( t ) in (28) and using the skew-symmetry property (14) , we obtain 1 2 d dt∥u(t)∥2 H+ν∥∇u(t)∥2 L2(Ω) +ZΩ(CS(x)δ)2|∇u(t)|3dx =⟨f(t), u(t)⟩V′,V .(29) The additional nonlinear dissipation term enhances the energy decay, especially in regions with strong shear, contributing to numerical and analytical stability. A priori estimate. Using Young’s inequality and the coercivity of S , we can estimate the forcing term: ⟨f, u⟩V′,V ≤1 2ν∥f∥2 V′+ν 2∥u∥2 V.(30) Substituting (30) into (29) and integrating over (0, t), we obtain ∥u(t)∥2 H+νZt 0∥u(s)∥2 Vds + 2 Zt 0ZΩ(CS(x)δ)2|∇u(s)|3dx ds ≤ ∥u0∥2 H+1 νZt 0∥f(s)∥2 V′ds. (31) Inequality (31) expresses the **uniform boundedness** of the kinetic energy and viscous dissipation over finite time intervals, and the **additional control** of the L3 -norm of ∇udue to the Smagorinsky term. Time-averaged estimates. Dividing (31) by t and taking the limit as t→ ∞ , we obtain the uniform-in-time averaged inequality: lim sup t→∞ 1 tZt 0ν∥∇u(s)∥2 L2(Ω) + 2 ZΩ(CS(x)δ)2|∇u(s)|3dxds ≤1 νlim sup t→∞ 1 tZt 0∥f(s)∥2 V′ds. (32) 6 This result demonstrates that the energy input from the external forcing is balanced by viscous and nonlinear dissipative mechanisms in the long-time regime. Uniform boundedness in the absorbing set. From (31) , the sequence of energy norms satisfies ∥u(t)∥2 H≤ ∥u0∥2 He−νλ1t+1 ν2λ1 sup s≥0 ∥f(s)∥2 V′,(33) where λ1> 0 is the first eigenvalue of the Stokes operator A (cf. (12) ). Thus, the solution u(t) eventually enters and remains in the absorbing ball B0=(u∈H:∥u∥2 H≤1 ν2λ1 sup s≥0 ∥f(s)∥2 V′),(34) which provides a global bound for the energy of all trajectories and serves as the foundation for further analysis of global attractors and asymptotic regularity. Summary. Inequalities (31) – (34) establish the existence of uniform-in-time energy bounds for the adaptive Smagorinsky model. The additional cubic dissipative term yields an L3 -based control that strengthens the classical L2 -viscous dissipation, ensuring improved stability and convergence properties, particularly in high-Reynolds-number regimes. 2.5 Existence and Long-Time Behavior of Weak Solutions Let u0∈H and f∈L2 loc (0 ,∞ ; V′ ). We consider the adaptive Smagorinsky system in the abstract form (27): du dt +νAu +B(u, u) + S(u) = f, u(0) = u0.(35) Existence of weak solutions. The operator F(u) := νAu +B(u, u) + S(u) is bounded, hemicontinuous and monotone on V . In particular, A is linear, self-adjoint and coercive (cf. (11) ), B is trilinear and skew-symmetric (cf. (14) ), and S is maximal monotone and coercive (cf. (24)–(26)). By the Minty–Browder theorem for monotone operators, there exists at least one function u∈L2(0, T;V), ∂tu∈L4/3(0, T;V′),(36) satisfying the weak formulation (28) for all v∈Vand almost every t∈(0, T). Uniqueness in 2D. In two-dimensional domains (Ω ⊂R2 ), the nonlinear term B ( u, u ) satisfies the Ladyzhenskaya inequality, which implies |b(u, u, v)| ≤ C∥u∥V∥u∥1/2 H∥v∥V,(37) allowing the classical Gronwall argument to establish uniqueness of weak solutions. In 3D, uniqueness remains an open problem for general initial data, as in the classical Navier–Stokes theory. Long-time boundedness and absorbing sets. From the energy inequality (31) and the absorbing ball (34) , the dynamical system generated by (35) is dissipative in H : for sufficiently large t≥t0, all trajectories enter the absorbing set B0=(u∈H:∥u∥2 H≤1 ν2λ1 sup s≥0 ∥f(s)∥2 V′).(38) 7 Existence of a global attractor. Let S ( t ) denote the evolution operator (semigroup) associated with (35) . The semigroup S ( t ) is continuous, dissipative, and asymptotically compact in H . By the classical theory of infinite-dimensional dynamical systems, there exists a compact global attractor A ⊂ H, S(t)A=A,∀t≥0,(39) which attracts all bounded sets in Hwith respect to the L2-norm: lim t→∞ distH(S(t)B, A)=0,∀B⊂Hbounded.(40) Implications for adaptive Smagorinsky LES. The presence of the nonlinear eddyviscosity operator S ( u ) enhances the dissipative mechanism, enlarges the effective viscosity in high-shear regions, and stabilizes the long-time dynamics. Consequently, the adaptive Smagorinsky model admits a well-defined weak solution semigroup with a bounded global attractor, providing a rigorous foundation for numerical simulations and long-time averaging of turbulent flows. 2.6 Energy Identity and Estimates Let u∈L2 (0 , T ; V ) be a weak solution of the adaptive Smagorinsky model (35) . Testing the weak formulation (28) with v=u(t) yields the energy identity 1 2 d dt∥u(t)∥2 L2(Ω) +ν∥∇u(t)∥2 L2(Ω) +ZΩ(CS(x)δ)2|∇u(t)|3dx =⟨f(t), u(t)⟩V′,V ,(41) where the convective term vanishes due to the skew-symmetry property: b(u, u, u)=0,∇ · u= 0.(42) Application of Poincar´e inequality. From (7) , there exists a constant CP> 0 such that ∥u(t)∥2 L2(Ω) ≤C2 P∥∇u(t)∥2 L2(Ω),∀u∈H1 0(Ω)3.(43) Estimate of the forcing term. Applying Young’s inequality to ⟨f, u⟩V′,V , we obtain, for any ϵ > 0, ⟨f, u⟩V′,V ≤1 2ϵ∥f∥2 V′+ϵ 2∥u∥2 V.(44) Choosing ϵ=νleads to ⟨f, u⟩V′,V ≤1 2ν∥f∥2 V′+ν 2∥∇u∥2 L2(Ω).(45) Differential energy inequality. Substituting (45) into (41) and absorbing the ν 2∥∇u∥2 L2 term into the left-hand side gives d dt∥u(t)∥2 L2(Ω) +ν∥∇u(t)∥2 L2(Ω) + 2 ZΩ(CS(x)δ)2|∇u(t)|3dx ≤1 ν∥f(t)∥2 V′.(46) Integrated energy estimate. Integrating (46) over [0 , t ] leads to the a priori energy estimate ∥u(t)∥2 L2(Ω) +νZt 0∥∇u(s)∥2 L2(Ω) ds + 2 Zt 0ZΩ(CS(x)δ)2|∇u(s)|3dx ds ≤ ∥u0∥2 L2(Ω) +1 νZt 0∥f(s)∥2 V′ds. (47) This estimate provides: 8 •Control of kinetic energy: ∥u(t)∥2 L2remains uniformly bounded for finite times. •Viscous dissipation control: Rt 0∥∇u(s)∥2 L2ds is finite. • Nonlinear eddy-viscosity control: Rt 0RΩ ( CSδ ) 2|∇u|3dxds quantifies the enhanced damping in high-shear regions. Together, these inequalities form the foundation for existence, uniqueness (in 2D), and long-time stability of weak solutions to the adaptive Smagorinsky LES model. 2.7 Compactness and Weak Convergence Let {un} be a sequence of Galerkin approximations to the adaptive Smagorinsky system (35) , constructed in finite-dimensional subspaces Vn⊂V . Assume that {un} satisfies the uniform energy bounds: sup n≥1 ∥un∥L∞(0,T;H)≤C, sup n≥1 ∥un∥L2(0,T;V)≤C, (48) for some constant C > 0 independent of n. Weak convergence. By the Banach–Alaoglu theorem, there exists a function u∈ L2(0, T;V) such that, up to a subsequence, un⇀ u weakly in L2(0, T;V),(49) and ∂tun⇀ ∂tuweakly in L4/3(0, T;V′).(50) Strong convergence via compactness. To pass to the limit in the nonlinear terms, strong convergence in L2 (0 , T ; H ) is required. By the Aubin–Lions compactness lemma [11], we have the compact embedding L2(0, T;V)∩H1(0, T;V′),→,→L2(0, T;H),(51) which implies the existence of a subsequence (still denoted un) such that un→ustrongly in L2(0, T;H).(52) Consequence for the nonlinear operator. The strong convergence (52) together with the pointwise monotonicity and continuity of the Smagorinsky operator S ( u ) (cf. (24) , (25)) allows passing to the limit in ⟨S(un), v⟩V′,V → ⟨S(u), v⟩V′,V ,∀v∈V, (53) thereby establishing that the limit u is indeed a **weak solution** of the original adaptive Smagorinsky system. The combination of uniform energy bounds (48) , weak compactness (49) – (50) , and strong convergence (52) ensures rigorous passage to the limit in the Galerkin scheme, guaranteeing the existence of weak solutions. 2.8 Functional Inequalities and Interpolation In order to control intermediate norms that appear in the nonlinear terms, we employ the Gagliardo–Nirenberg interpolation inequality. For u∈H1 0 (Ω) and 2 ≤p≤ 6, there exists a constant C=C(Ω, p)>0 such that ∥u∥Lp(Ω) ≤C∥u∥1−θ L2(Ω)∥∇u∥θ L2(Ω), θ =3(p−2) 2p.(54) Remarks: 9 ∥u(t)∥2 H1 0(Ω) +Zt τ∥∆u(s)∥2 L2(Ω) ds ≤C1 + ∥u0∥2 H1 0(Ω) +∥f∥2 L2(0,T;H), t ≥τ > 0. This result demonstrates the model’s capacity to control higher-order derivatives, ensuring numerical stability and physical fidelity. 4.3 Finite-Dimensional Global Attractor The adaptive Smagorinsky system generates a semiflow S ( t ) that possesses a compact global attractor A ⊂ H1 0 (Ω) with finite fractal dimension. The attractor’s dimension is bounded by: dimF(A)≤C∗L δ31 + 1 ν2max x∈ΩCS(x)2Re3/2, where C∗ depends on domain geometry and Sobolev constants. The existence of a finite-dimensional attractor confirms the model’s long-time stability and its suitability for numerical simulations of turbulent flows. 5 Conclusions The adaptive Smagorinsky model represents a significant advancement in the mathematical modeling of turbulent flows, addressing critical limitations of classical LES closures. By introducing a spatially modulated dissipation coefficient, the model enhances both the regularity and stability of solutions, particularly in high-shear regions and near solid boundaries. Our rigorous analysis establishes the existence and uniqueness of weak solutions, demonstrates uniform energy dissipation bounds, and proves the existence of a finite-dimensional global attractor. The model’s adaptive dissipation mechanism provides a robust framework for capturing the complex dynamics of turbulent flows, offering improved physical fidelity and numerical stability. The enhanced regularity and finite-dimensional dynamics make it particularly suitable for high-Reynolds-number simulations in engineering and environmental applications. This study bridges theoretical advancements with practical implications, contributing to the development of more accurate and reliable turbulence models. Future research directions include the extension of uniqueness results to three-dimensional domains, the exploration of optimal modulation functions Φ( x ), and the application of the model to real-world flow problems. The theoretical foundations laid herein pave the way for further innovations in adaptive subgrid-scale modeling and computational fluid dynamics. Acknowledgements Santos gratefully acknowledges the support of the PPGMC Program through the Postdoctoral Scholarship PROBOL/UESC, No. 218/2025. Sales expresses his sincere gratitude to CNPq for the financial support provided under Grant 308816/2025-0. 16 Table 1: Notation, Symbols, and Nomenclature Symbol Description Domain/Context Ω Bounded domain in R3Physical space ∂Ω Boundary of Ω Domain geometry Lp(Ω) Lebesgue space with norm pFunc. analysis Hm(Ω) Sobolev space of order mFunc. analysis HSolenoidal functions in L2(Ω)3,u·n|∂Ω= 0 Fluid mech. VSolenoidal functions in H1 0(Ω)3Fluid mech. V′Dual space of VFunc. analysis H′Dual space of HFunc. analysis uVelocity field Navier-Stokes eqs. pPressure field Navier-Stokes eqs. νKinematic viscosity Fluid prop. fExternal force Navier-Stokes eqs. AStokes operator Spectral analysis PLeray-Helmholtz projection Helmholtz dec. b(u, v, w) Trilinear form, convective term Variational form. B(u, v) Bilinear operator associated with b(u, v, ·) Abstract form. S(u) Nonlinear dissipation, adaptive Smagorinsky model Turbulence model CS(x) Spatially varying Smagorinsky: CS(x) = CS0e−βΦ(x)Adaptive model CS0Reference Smagorinsky constant Model const. βExponential modulation parameter Adaptive model Φ(x) Smooth, nonnegative modulation function Adaptive model δCharacteristic filter width LES λkStokes eigenvalues Spectral analysis wkStokes eigenfunctions Orthonormal basis ⟨·,·⟩V′,V Duality product V′, V Variational form. (·,·)L2(Ω) Inner product in L2(Ω) Hilbert space ∥·∥Lp(Ω) Norm in Lp(Ω) Lebesgue space ∥·∥Hm(Ω) Norm in Hm(Ω) Sobolev space ∇Gradient operator Vector calc. ∆ Laplacian PDEs ∂tPartial derivative w.r.t. time Evol. eqs. Re Reynolds number: Re = UL/ν Dimensional analysis UCharacteristic velocity: U=∥u0∥L2/p|Ω| Dimensional analysis LCharacteristic length: L= diam(Ω) Dimensional analysis ACompact global attractor Dyn. sys. dimF(A) Fractal dimension of ADyn. sys. theory S(t) Semigroup (evolution operator) Dyn. sys. εSEnergy dissipation rate Asymptotic analysis CPPoincar´e constant Func. inequalities γExponential decay rate: γ=ν/C2 P Asymptotic stability θInterpolation exponent, Gagliardo-Nirenberg Func. analysis unGalerkin approximation of uNum. methods VnFinite-dimensional Galerkin subspace Num. approx. TFinal time Time domain u0Initial condition Initial value prob. 17 References [1] Smagorinsky, J. (1963). General circulation experiments with the primitive equations: I. The basic experiment. Monthly weather review, 91 (3), 99-164. https://doi.org/ 10.1175/1520-0493(1963)091%3C0099:GCEWTP%3E2.3.CO;2. [2] Ladyzhenskaya, O. A. (1967). New equations for the description of the motions of viscous incompressible fluids, and global solvability for their boundary value problems. Trudy Matematicheskogo Instituta Imeni VA Steklova, 102, 85-104. [3] VonNeumann, J., & Richtmyer, R. D. (1950). A method for the numerical calculation of hydrodynamic shocks. Journal of applied physics, 21(3), 232-237. https://doi. org/10.1063/1.1699639. [4] Lilly, D. (1966). The representation of small-scale turbulence in numerical simulation experiments. NCAR Report. https://ui.adsabs.harvard.edu/link_gateway/ 1966ncar.reptq05b0L/doi:10.5065/D62R3PMM. [5] Germano, M., Piomelli, U., Moin, P., & Cabot, W. H. (1991). A dynamic subgridscale eddy viscosity model. Physics of fluids a: Fluid dynamics, 3(7), 1760-1765. https://doi.org/10.1063/1.857955. [6] Doering, C. R., & Foias, C. (2002). Energy dissipation in body-forced turbulence. Journal of Fluid Mechanics, 467, 289-306. https://doi.org/10.1017/ S0022112002001386. [7] Frisch, U., & Kolmogorov, A. N. (1995). Turbulence: the legacy of AN Kolmogorov. Cambridge university press. [8] Pope, S. B. (2001). Turbulent flows. Measurement Science and Technology, 12(11), 2020-2021. 10.1088/0957-0233/12/11/705. [9] Layton, W. (2007). Bounds on energy and helicity dissipation rates of approximate deconvolution models of turbulence. SIAM journal on mathematical analysis, 39 (3), 916-931. https://doi.org/10.1137/06066223X. [10] Vassilicos, J. C. (2015). Dissipation in turbulent flows. Annual review of fluid mechanics, 47(1), 95-114. https://doi.org/10.1146/annurev-fluid-010814-014637. [11] Simon, J. (1986). Compact sets in the space L p (O, T; B). Annali di Matematica pura ed applicata, 146 (1), 65-96. https://doi.org/10.1007/BF01762360. [12] Temam, R. (2012). Infinite-dimensional dynamical systems in mechanics and physics (Vol. 68). Springer Science & Business Media. [13] Constantin, P., & Doering, C. R. (1995). Variational bounds on energy dissipation in incompressible flows. II. Channel flow. Physical Review E, 51 (4), 3192. https: //doi.org/10.1103/PhysRevE.51.3192. 18