Other meanings of Smoothed particle hydrodynamics
COMPUTATIONAL FLUID DYNAMICS
Smoothed-particle hydrodynamics is a mesh-free Lagrangian method for fluid-flow simulations using particles. Each particle carries fluid properties such as mass, density, velocity, and energy, while nearby particles estimate continuous fields through a smoothing kernel. The method was introduced independently for astrophysical calculations in the late 1970s and later became a general tool in computational fluid dynamics.12
SPH replaces a fixed computational grid with moving interpolation points. A field quantity is approximated by a weighted sum over neighboring particles, with the weights supplied by a compact-support smoothing kernel and a length scale usually denoted h. Density can be obtained by summing particle masses, while pressure and other forces follow from discretized continuum equations.1
The method is naturally Lagrangian: particles follow material motion, so advection is handled without repeatedly transporting fluid variables across cell boundaries. This makes SPH attractive for free surfaces, large deformations, multiphase interfaces, and problems in which a conventional mesh would become severely distorted. Conservation of mass is especially direct when particle masses remain fixed.
SPH is valuable when topology changes or moving boundaries dominate the calculation. It has been used for stellar and galactic dynamics, impact and fragmentation, coastal and ocean flows, violent sloshing, granular materials, and some biomedical and engineering problems.23
Its principal strengths are geometric flexibility, automatic concentration of computational effort in regions containing particles, and the absence of mesh tangling. Its weaknesses include comparatively expensive neighbor searches, sensitivity to kernel and smoothing-length choices, and difficulties near boundaries. Pressure oscillations, tensile instability, artificial surface tension, and inaccurate behavior across strong density contrasts have motivated many revised formulations.
Modern implementations often combine SPH with domain decomposition, tree or hash-based neighbor searches, and parallel computing. Resolution is local rather than fixed globally, although particle disorder can reduce the expected accuracy.
Reliable SPH requires consistent discretization, a suitable equation of state or pressure solver, and controls for viscosity and particle disorder. Artificial viscosity is widely used to capture shocks and stabilize converging flows, but excessive viscosity can smear shear layers and suppress physically important instabilities.1
Variants include density-based and pressure-based formulations, incompressible or weakly compressible schemes, Godunov-type SPH, and formulations with improved consistency or transport velocity corrections. Adaptive smoothing lengths allow particles to represent changing resolution, but they introduce additional terms that must be treated consistently to preserve conservation.4
Validation commonly uses standard tests such as shock tubes, hydrostatic equilibria, dam breaks, Kelvin–Helmholtz instability, and rotating-fluid problems. Agreement with analytic solutions or controlled experiments matters because visually plausible particle motion can conceal substantial pressure or convergence errors.
SPH was not originally conceived as a general-purpose engineering method; its first influential formulations targeted self-gravitating astrophysical fluids, where a mesh would poorly follow expanding, collapsing, or fragmenting matter. The same particle representation later made the method useful for extreme free-surface events and highly deformable solids.
A subtle distinction separates particle resolution from smoothing resolution. The interparticle spacing, kernel support, and physical length scales need not coincide, and insufficient neighbors can produce noise even when many particles are present. Boundary treatment is another specialized subject: ghost particles, boundary forces, mirror constructions, and integral corrections each alter accuracy and conservation.
In astrophysics, SPH is often coupled to gravity solvers and radiative or magnetohydrodynamic physics. In engineering, hybrid particle–mesh methods can retain SPH's geometric flexibility while reducing long-range force costs or improving pressure projection.
SPH denotes a family of discretizations rather than one single algorithm; accuracy and stability depend substantially on the kernel, boundary treatment, pressure formulation, and time integration scheme.
Help improve the encyclopedia. Reports go straight to the site manager.