Correlation between particle size distribution and explosion intensity of aluminum powder

Abstract

To successfully reduce the occurrence of aluminum powder explosion accidents, this study uses a statistical analysis approach to examine the relation between 14 particle size parameters (D₁₀–D₉₀, D_3,2, D_4,3, σ_D, PSS, and SSA) and 5 explosion intensity parameters (P_max, (dP/dt)_max, K_st, t_{1 min}, and t_{2 min}). Note that D₁₀–D₉₀ represent the 10th–90th percentiles of particle size distribution, respectively, unit: μm; D_3,2 is the average diameter of surface area momentum, unit: μm; D_4,3 represents the average diameter of volume momentum, unit: μm; σ_D represents the particle size polydispersity; PSS is the particle size span, unit: μm; and SSA is the specific surface area, unit: m²·kg⁻¹; P_max is the maximum explosion pressure, unit: MPa; (dP/dt)_max is the maximum explosion pressure increase rate with time, unit: MPa·s⁻¹; K_st is the explosion index, unit: MPa·m·s⁻¹; t₁ represents the time from the start of ignition to the time when the maximum pressure is attained in the 20-L explosion sphere cavity, unit:·s, t₂ represents the time from the start of ignition to the time when the explosion pressure increase rate reaches the maximum value, unit:·s. The t_{1 min} and t_{2 min} represent the minimum value of t₁ and t₂, respectively, unit:·s. The stepwise regression method is used to screen the effective particle size parameters that can describe the effect of particle size distribution on the explosive intensity of aluminum powder. The results indicate that σ_D and PSS are not recommended to characterize the particle size distribution in examining the explosion intensity. In addition, among the particle size percentiles, D₁₀, D_20, and D₃₀ have the strongest correlation with the explosion intensity parameters. SSA has a strong correlation with (dP/dt)_max and K_st, and D_3,2 has a strong correlation with P_max and t_{2 min}. The correlation of the small particle size percentiles D₁₀ and D₂₀ with P_max as obtained by the stepwise regression method is statistically significant. Moreover, the correlations of SAA with (dP/dt)_max and K_st; D₁₀, D_3,2, and D₃₀ with t_{1 min}; and D₁₀ and SSA with t_{2 min} are statistically significant. Therefore, it is advised to use particle size percentiles (D₁₀–D₃₀) and D_3,2 to describe the particle size distribution when examining the explosion intensity. Furthermore, we should focus on the proportion of small-particle-size aluminum powder components (D₁₀–D₃₀).