Static Hazards of the VAST

Abstract

The vertical atmospheric storage tank (VAST) is a large, mechanically weak tank often used to store static accumulating flammable liquids. Explosion of a single VAST can destroy an entire tank farm via rocketing and domino effects. This article provides a new methodology for assessing electrostatic hazards. Appendix A describes flammability aspects, including a new analysis of flash point measurement errors. Appendix B discusses two recent VAST explosions. Recommendations are made for improving the safety of VAST operations. Electrostatic analysis starts with our previously published model expressing the surface potential at the center of the liquid surface “Vmax” in the form of an infinite series of Bessel and hyperbolic functions. A truncated one-term approximation with an improved scale factor “κ” is provided to facilitate spreadsheet calculation. Using at least the first three terms of the series expansion, thousands of individual calculations were made for a wide range of tank dimensions, filling rates and liquid relaxation times, to determine the maximum potential “V∗max” attained during filling, which occurs at some critical depth “Dcrit”. It’s shown there exists a critical width “wcrit”, typically 0.75-1.5 m, at which V∗max reaches a maximum for a given flow rate and relaxation time. At widths above the critical width, V∗max becomes independent of flow rate and, provided the aspect ratio (H/w) of the VAST exceeds 2/3, V∗max becomes independent of height (H) as well. The worst case for VAST filling is a width near “wcrit” plus an effective liquid relaxation time ∼30 s (measured conductivity ∼1 pS/m). We derived very simple expressions for the critical width “wcrit”, the critical liquid depth “Dcrit”, the V∗max yielded by a given charging current “IC”, and the charging current needed to yield a particular V∗max. Of particular interest were the conditions for V∗max = 25 kV, the recognized threshold for incendiary brush discharges. Assuming a worst case 30 s relaxation time and a maximum allowable V∗max of 25 kV, we examined three different expressions for the charging current “IC” and determined the minimum “vd” product (flow velocity x inlet pipe diameter) that could theoretically produce this potential. Equations and graphs are presented showing that the maximum allowable vd product increases with increased VAST width (beyond the critical width) and varies with the assumed charging equation. It’s shown that the classic “Schon” equation isn’t conservative and that charging currents are likely to be much higher than predicted, especially when the diameter of the inlet pipe is less than about 0.1 m (4-inch) and the liquid has a high charging tendency. The “vd” expressions and graphs should assist in the development of Codes of Practice. The uniform charge density assumption used in our model is most valid as conditions approach worst-case, since for widths near the critical width the jet mixing time constant at typical flow rates is small compared with the assumed “worst case” 30 s relaxation time constant. However, in practical cases, jetting from bottom side entry inlets may greatly increase charge densities at the liquid surface. This plus other non-idealities such as water and sediment in tank bottoms are discussed with reference to practical VAST operations.