CFD analysis of gas explosions vented through relief pipes

Abstract

Vent devices for gas and dust explosions are often ducted to safe locations by means of relief pipes. However, the presence of the duct increases the severity of explosion if compared to simply vented vessels (i.e. compared to cases where no duct is present). Besides, the identification of the key phenomena controlling the violence of explosion has not yet been gained. Multidimensional models coupling, mass, momentum and energy conservation equations can be valuable tools for the analysis of such complex explosion phenomena. In this work, gas explosions vented through ducts have been modelled by a two-dimensional (2D) axi-symmetric computational fluid dynamic (CFD) model based on the unsteady Reynolds Averaged Navier Stokes (RANS) approach in which the laminar, flamelet and distributed combustion models have been implemented. Numerical test have been carried out by varying ignition position, duct diameter and length. Results have evidenced that the severity of ducted explosions is mainly driven by the vigorous secondary explosion occurring in the duct (burn-up) rather than by the duct flow resistance or acoustic enhancement. Moreover, it has been found out that the burn-up affects explosion severity due to the reduction of venting rate rather than to the burning rate enhancement through turbulization.

Introduction

Vent devices for gas and dust explosions are often ducted to safe locations by means of relief pipes, for the discharge of hot combustion products – even toxic – or blast waves [1], [2]. Relief pipes are specifically required when the hot jet flowing violently from the vent area has to be avoided, e.g. within buildings. On the other hand, the presence of a duct is likely to increase the severity of the explosion with respect to simply vented vessels [2], [3], [4], [5].

In the last two decades a number of works have addressed the issue of gas explosions vented through relief pipes [4], [5], [6], [7], [8], [9]. Several phenomena were identified as affecting the increase of the overpressure with respect to simply vented vessels such as secondary explosion in the duct (burn-up), frictional drag and inertia of the gas column in the duct, acoustic and Helmholtz oscillations.

Acoustic oscillations are deemed to be involved in the generation of later strong pressure peaks in simply vented vessels [10]. Kordylevski and Wach [7] detected acoustic oscillations in the pressure records of their experiments on small-scale ducted vent explosions suggesting this phenomenon could be somehow responsible for the unusual pressure rise in the vessel. Actually the link was quite loose as they observed stronger pressure rises (with respect to simply vented vessels) also in the absence of oscillatory behaviour. Moreover, Ponizy and Leyer [5] reported an increased violence of the duct-vented explosion on a small scale with stoichiometric propane–air mixtures which does not constitute a suitable condition for acoustic enhancement of explosion [11]. Helmholtz oscillations in small explosion chambers fitted to venting ducts were observed early by Cubbage and Marshall [12]. Also, McCann et al. [13] clearly detected Helmholtz oscillations in the final stages of explosions in small-scale vessels fitted to duct of different lengths. They pointed out that such oscillations could play an important role in triggering Taylor instability. Nevertheless, this effect cannot be considered necessary condition for the increase of the explosion violence, as the occurrence of this flame front instability is well acknowledged even in simply vented vessels [14]. Besides, as pointed out by the same authors, such oscillations should have more pronounced effects on larger scale explosions as confirmed by Kumar et al. [15], who detected severe Helmholtz type oscillations for middle scale lean hydrogen–air explosions vented through a duct.

The turbulent mixing of hot and fresh gases in the initial section of the duct after the flame entrance promotes a violent burning therein (an explosion-like combustion or “burn-up”). Hence, the pressure impulse in the duct induces the backflow of gases from the duct to the vessel with the possible consequent turbulization of residual combustion in the vessel and the blockage of the gas efflux [4], [5]. This violent explosion in the duct named as burn-up was addressed by some authors as the main responsible for the dramatic increase of the pressure in the vessel [4], [5], [9], [16].

Other authors indicated additional pressure drops due to the resistance of the gas flow in the vessel–duct assembly as main responsible for the higher pressure rise in the vessel with respect to simply vented vessels [17], [18]. Mechanical, steady-state type pressure drops for the present configuration can be substantial due to the very high flow velocities attained in the duct and the concentrated losses in the sudden flow area changes (respectively, at the duct entrance and exit).

On the theoretical side, some efforts have been devoted to get insights into the phenomenon by developing mathematical models. Zero-dimensional models proved effective in advancing the comprehension and the formalization of data gained for simply vented explosions [19], [20], [21]. Zero-dimensional and one-dimensional models [4], [17] were also proposed to represent an explosion vented through a duct but result in a scarce predictive capability as they rely strongly on empirical parameters. In fact, it could not be expected that such models provided a sound description of the phenomenon due to the assumption of a spherical flame propagation whatever the geometric complicacy (here included the presence of a discharging duct). According to these models the enhancement of the burning rate through turbulization [4] and the friction losses [17] are the most important phenomena affecting overpressure.

The available guidelines for the design of ducted vents for gas explosions are those proposed by Bartknecht [3], also reported in NFPA 68 [1], which gives barely an empirical correlation based on simply vented vessels indications presented in the same reference. Due to their empirical foundation, NFPA correlations are to be used very carefully as they can lead to gross errors [20].

Prior to any correlation development, a sensible approach should be trying to understand the relative contribution of all above reported events to the whole phenomenon. This step appears to be preparatory (if not mandatory) for the selection of the ruling parameters and for the making of the necessary approximations aiming at developing sound engineering correlations.

To this regard it must be noticed that, when care is used in analyzing their results, computational fluid dynamic (CFD) models can be valuable tools in assessing explosion scenarios provided that a satisfactory validation is carried out [22], [23], [24], [25], [26]. CFD models can in principle take into account much more physics than zero-dimensional models as they can, for example, relieve the severely restrictive hypothesis on the geometry of the system and of the propagating flame.

In this work a theoretical model for the gas explosion vented through a duct has been developed and numerically solved by means of the CFD-ACE+ code by CFDRC [27]. The model has been validated and tested against the comprehensive set of experimental data of Ponizy and Leyer [5], [9]. Comparison of model predictions with experimental data is presented and explanations of experimental observed trends are proposed on the basis of model results. More specifically, the effects of varying the duct geometry (length and diameter) and the ignition position have been studied and analysed with the aid of the numerical computations. Detailed field data gained from the calculations have allowed to evaluate selectively the relative contribution of the mechanical (hindered gas efflux) and chemical (enhanced combustion) contributions to the final pressure load recorded in the system.