Universal Particle-Volume Distribution

• Universal particle-volume distribution is a framework that defines how intrinsic volumes of particles are statistically distributed under thermodynamic constraints and geometric influences.
• It is applied in SPH simulations and random geometric systems to optimize particle arrangements and predict lattice tiling and periodicity.
• The theory refines Gibbsian pressure corrections and models cluster formation, aiding practical simulations in granular and fluid systems.

The universal particle-volume distribution encompasses a range of physically and mathematically rigorous descriptions characterizing how the spatial volumes associated with particles, clusters, or insertions are distributed in various statistical and geometric ensembles. It arises in statistical mechanics, continuum simulation (including Smoothed Particle Hydrodynamics, SPH), and random geometric systems, with central frameworks defined by Gibbs statistics, optimization-driven SPH initialization, random cluster formation, and geometric probability for hard convex bodies.

1. Fundamental Definitions and Frameworks

In equilibrium statistical mechanics, the particle-volume distribution describes the probability density of a single particle’s intrinsic volume $v$ within a fixed total system volume $V$. For a Gibbs system in the absence of external fields and short-range, pairwise-additive interactions, the one-particle volume distribution is a truncated exponential: $$p(v) = \frac{a\,e^{-a\,v}}{1 - e^{-a\,V}},\quad 0 \le v \le V,$$ where $a = p/(k_BT)$, with $p$ as the equilibrium pressure at chemical activity $z$ and $T$ the temperature (Ryazanov, 2019). This distribution reflects the fundamental thermodynamic constraints and excluded volume effects.

In continuum simulation methods such as SPH, the “particle-volume” is an emergent property of the kernel-based representation of the volume fraction field: $\alpha(x) = \sum_n V_n\,W(x - x_n),$ where $W$ is a compactly supported kernel (cut-off $h$), $x_n$ are particle coordinates, and $V_n$ the per-particle volumes (Fan et al., 2024).

In geometric probability, the particle-volume distribution generalizes to scale distributions of insertable hard convex bodies (e.g., disks, spheres), connected to the insertion probability $P_{\rm ins}(s)$ of a scaled test particle and its rate of change $F(s)$, yielding (Yang et al., 2023): $$F(s) = -\frac{d}{ds}P_{\rm ins}(s), \quad -\ln P_{\rm ins}(s) = \sum_{j=0}^d a_j\,s^j,$$ with $a_j$ fixed by the host density and Minkowski functionals of the shape.

2. Optimization and Particle Arrangement in SPH

Uniform particle-volume distributions in SPH are generated by minimizing the mismatch between the discrete and analytic volume fractions as an optimization problem: $$J[\{x_n\}, \{V_n\}] = \int_\Omega\left[\sum_n V_n W(x - x_n) - 1\right]^2 dx,$$ with gradient descent implemented via position updates (Fan et al., 2024): $$\frac{dx_i}{dt} = -\frac{\partial J}{\partial x_i}.$$ This evolution replicates the “repulsive relaxation force” used in physics-based SPH initialization. In perfectly periodic arrangements, all $V_i$ converge to a characteristic volume $V_{\rm char}$, determined by the partition-of-unity constraint: $V_{\rm char} = \frac{1}{\sum_j W(r_{ij}, h)},$ where $r_{ij} = \|x_i - x_j\|$ and the sum incorporates kernel support and spatial lattice geometry. Maximizing $V_{\rm char}$ identifies the optimal lattice tiling (hexagonal, square, parallelogram) for a given kernel and cut-off, rendering the distribution both predictable and tunable.

Near boundaries, kernel support is truncated, and boundary correction terms derived from a smoothed characteristic function $P(x)$ maintain uniformity up to $\partial\Omega$.

3. Gibbsian Distribution and Thermodynamic Refinements

In Gibbs equilibrium, the particle-volume distribution provides the statistical foundation for excluded-volume corrections to equations of state. The average intrinsic volume per particle,

$$\langle v\rangle = \frac{1}{a}\frac{1-(1+y)e^{-y}}{1-e^{-y}},\quad y = aV,$$

generalizes the standard “own” volume parameter $b$ in van der Waals models. Ryazanov’s approach substitutes $b \to \langle v\rangle(p, T)$ and yields refined, self-consistent pressure corrections: $$p = \frac{\rho k_B T}{1 - \rho \langle v\rangle(p,T)} - a_{\rm vdW}\rho^2.$$ In the macroscopic limit ($y \to \infty$), all corrections simplify to the universal exponential law $p(v)=a\,e^{-a\,v}$, with mean and variance $1/a$, $1/a^2$, respectively (Ryazanov, 2019).

Assumptions underpinning this universality include equilibrium conditions, pairwise-additive interactions, and absence of external fields. Deviations arise for finite system size, long-range correlations, or non-Gibbsian statistics.

4. Geometric Probability: Scale Distributions for Hard Convex Particles

Recent work established a universal relation linking the probability of inserting a randomly scaled hard convex particle (shape $K$, scale $s$) to the corresponding scale (volume) distribution $F(s)$ (Yang et al., 2023): $F(s) = -\frac{d}{ds}\ln P_{\rm ins}(s),$ where $P_{\rm ins}(s)$ is a polynomial in $s$ whose coefficients $a_j$ depend on the Minkowski functionals $W_m(K)$: $$-\ln P_{\rm ins}(s) = \sum_{j=0}^d \rho\,\binom{d}{j}\,W_{d-j}(K)\,s^j,$$ for $d$ dimensions and host number density $\rho$. For spheres, disks, triangles, and other convex bodies, explicit forms for $F(s)$ follow directly. Hadwiger’s theorem guarantees that only these geometric invariants appear, rendering the distribution universal across shapes and host configurations.

Empirical confirmation emerges from Monte Carlo sampling: for all investigated shapes, the insertion probability and scale distribution closely match the predicted polynomial forms.

5. Cluster Size Distributions in Random Particle Systems

Universal distributions also describe how particles aggregate into clusters based on spatial overlap. For randomly placed spheres (radius $R$, mean spacing $d$), the order parameter $\alpha=R/d$ governs cluster formation (Khokonov et al., 2018). The rank-based fraction $N(x;\alpha)$ and cluster-size probability $f_s(s;\alpha)$ are

$$N(x;\alpha) = (1-\alpha)x^{-\alpha}, \quad f_s(s;\alpha) = \frac{1}{\alpha}(1-\alpha)^{1/\alpha}s^{-(1+\alpha)/\alpha}.$$

For $\alpha$ in the range $1.5 \lesssim \alpha \lesssim 2.5$, a log-normal approximation applies. Moments of this distribution converge for $n<1/\alpha$ (mean, variance, etc.), and empirically describe clustering in gas condensation and granular assemblies.

This framework provides a parameter-universal description: particle aggregation can be quantitatively fitted and modeled using a single geometric parameter $\alpha$, independent of boundary conditions or detailed interactions.

6. Universality, Limitations, and Applications

The universal particle-volume distribution, by construction, is robust across a diverse range of physical and computational systems. Universality manifests in:

• Exponential laws for Gibbs systems in the thermodynamic limit, with all microscopic details entering solely through $a = p/(k_BT)$ (Ryazanov, 2019).
• Kernel-controlled characteristic volumes and periodic arrangements in SPH, with lattice type transitions predictable from $h/\Delta x$ (Fan et al., 2024).
• Geometric invariance in insertion and scale distributions, indexed only by packing density and Minkowski functionals of the particle shape (Yang et al., 2023).
• Parameter-universal cluster distributions with explicit power-law or log-normal forms (Khokonov et al., 2018).

Limitations include requirement for large system size (continuum approaches), equilibrium conditions, smooth boundaries, and short-range pairwise interactions. Strong correlations, external field effects, non-Gibbs statistics, or non-scalar volume measures result in qualitative changes to the distribution.

Applications span from high-fidelity SPH initialization, thermodynamic modeling, granular and gas-liquid phase clustering, to geometric probability in convex body fluids. The theoretical results guide practical simulation protocols, equation-of-state refinement, and experimental fitting in both physical and engineered systems.