Variational Projection of Navier Stokes (VPNS) : A Novel Optimization Based Solver for Incompressible Fluid Flows

This paper introduces the Variational Projection of Navier Stokes (VPNS), a novel computational framework that reformulates the simulation of incompressible flows by shifting from a traditional Newtonian formulation to a variational mechanics perspective. The method is grounded in the Principle of Minimum Pressure Gradient (PMPG)—the continuum analog of Gauss' Principle of Least Constraint—which asserts that flow evolution minimizes the $L^2$-norm of the pressure gradient force required to maintain incompressibility. By spatially discretizing this principle, we transform the fluid mechanics problem into a Convex Quadratic Programming (QP) problem. Leveraging the Moore-Penrose theory of generalized inverses, we derive a closed-form analytical solution to this constrained optimization, yielding an explicit system of Ordinary Differential Equations (ODEs) amenable to direct time integration. A defining advantage of VPNS is the elimination of the computationally expensive Pressure Poisson Equation (PPE) from the time-marching loop. Instead, the global constraint is satisfied via a pre-computed algebraic projection operator, reducing the enforcement of mass conservation to a matrix-vector multiplication. We validate the solver against the lid-driven cavity benchmark, demonstrating that VPNS captures the flow physics with high fidelity while maintaining divergence-free constraints to machine precision ($10^{-14}$). Furthermore, perturbation analyses quantitatively verify that the computed flow evolution corresponds strictly to the global minimum of the pressure gradient cost landscape. VPNS bridges analytical mechanics and computational fluid dynamics, offering a physically insightful and robust solver that precludes the need for iterative convergence within time steps.