Concentration Measurements Above an Underwater Release of Natural Gas

Abstract

Summary. Underwater gas releases resulting from blowouts, pipeline or riser breaks, etc., represent a fire hazard for structures and craft near the surfacing zone. The hydrodynamics of the underwater bubble plume has been the subject of several recent investigations, but no method or data that are useful in predicting the gas concentration above the surfacing zone are known to exist. This paper presents measurements from a field experiment using methane flow rates up to 1.3 normal m3/s [46 scf/sec] at a depth of 50 m [164 ft]. In addition to the continuous concentration measurements from four sensors in the surfacing zone (50 cm [20 in.] above the water surface), certain plume parameters (e.g., plume rise time in air and water and plume center wandering and offset) were either measured or deduced from video recordings. The experiments were performed under ideal weather conditions and the data obtained were not influenced by such extraneous environmental factors as wind, waves, currents, or stratification. For comparison with the data, a simple momentum theory as developed for the zone of formation of the light-gas plume. The agreement with the averaged concentration data is very good over the range of flow rates investigated. Introduction Large-scale underwater releases of gaseous hydrocarbons are well known in the offshore oil and gas industry. The releases are sometimes intentional, as in the case of the venting of an underwater flowline, but more often they result from accidents, e.g., blowouts and broken underwater pipelines or risers. A recent incident at Haltenbanken (Norway) was caused by a blowout from a shallow gas pocket (Fig. 1). The most obvious hazard for drilling platforms and craft near the underwater gas release is fire. The buoyancy loss feared by many oil rig crews is important only in very unusual circumstances, such as shallow water, large volumes of gas, and marginal stability. A complete break of a major underwater pipeline at the depths prevailing in the North Sea could cause a buoyancy loss of nearly 10%. The chance that a fig or ship would be situated right in the boil region is naturally small. The buoyancy loss associated with a blowout is much smaller and does not represent a hazard for reasonable depths, i.e., in excess of 100 m [330 ft]. A 1 % buoyancy loss, which is on the order of that caused by changes in salinity, represents a 1 % void fraction (average) in the plume. As will be seen, the surface concentration (gas in air) for the same release could be considerably higher. It follows that a fire hazard is also possible in cases for which the buoyancy loss is well below a critical value. Blowouts in shallow waters have been observed to bum-e.g., Ixtoc field (Mexico, 1979, oil and gas) and the lesser-known Fateh field (Dubai, 1976, gas). Detailed photographs, such as the one of the Fateh field over-water light-gas plume shown in Fig. 2, are useful when the zone of formation of the plume is modeled to determine the gas concentration. The concentration above the water surface is naturally a function of the gas concentration or void fraction in the bubble plume in the surface layer. The void fraction is ideally proportional to the gas flow rate but inversely proportional to the square of the depth. (The effect of compressibility, i.e., gas expansion with reduced depth, changes the flow picture somewhat for large depths.) A blowout comparable in flow rate to that of the Fateh field incident, but set at twice the depth, would have a surface concentration only one-fourth as high and most likely would not bum. For very large depths, no natural gas blowout is likely to bum. Hydrate formation (for depths in excess of 300 m [1,000 ft]) could further reduce the low concentration caused by plume spreading under water. Understanding of the plume-surface interaction in both the impingement zone under the water surface and the zone of formation above the water surface is necessary to predict the gas concentration in the surfacing zone. A description of the hydrodynamic structure of deep-set bubble plumes is given by Fannelop and Sjoen. Their analysis of the impingement region is used herein. The water in the rising plume is deflected outward by the free surface without loss of momentum, whereas the gas (in the form of bubbles or bubble swams) breaks through. This crude interaction model cannot predict the detailed distribution of bubbles in the surfacing zone, and the initial gas velocity is also uncertain. But the gross features of observed plumes, boil area, and outward velocity are in agreement with the model. We found no model or study in the literature for the zone of formation of the light-gas plume above the boil region. The near field of an area source is generally difficult to analyze for both light and heavy releases. The most critical condition (i.e., the highest gas-in-air concentration) would occur in light or zero wind conditions. For the very low velocities that occur in the zone of formation, the flow could easily be laminar with an unstable oscillating plume above. Such a flow would be susceptible to scale effects, and the results from too-small (laboratory-scale) experiments could be misleading with regard to the fire risk above a real blowout. Statoil undertook this field experiment to provide reliable data on an important hazard. The goal was to obtain concentration data as functions of flow rate at a depth comparable to an offshore field. In addition, we hoped to gain further insight into bubble and gas plumes, their interactions with the water surface, and the importance of wind. In a field experiment, the environmental factors are uncontrollable; a single experiment cannot be repeated without some changes in external conditions. We conducted this experiment in May and June 1985 in the Trondheim Fjord at a depth of 50 m [164 ft] under near-ideal (stable and repeatable) weather conditions. Experimental Arrangements Large quantities of gas were required to produce realistic gas concentrations (on the order of 1 vol%) from an underwater releast at 50-m [164-ft] depth. In a typical test, 300 normal m3 [10,800 scf] of gas was released over a period of 5 minutes against a pressure of about 0.6 MPa [6 bar]. The gas supply consisted of a large battery of high-pressure gas bottles that contained pure methane. This gas supply was placed onshore (Fig. 3), and the gas was piped first through a large pressure vessel and then to the release point at 50-m [164-ft] depth through a flexible 10-cm [4-in.] -wide, 110-m [360-ft] -long hose. The pressure-regulation equipment (e.g., the flow-rate control valve) was located onshore. The open end of the pipe was locked in the vertical position and anchored to a heavy steel foundation on the sealbed. Between each release, the pipe would fill with water. An underwater video camera was positioned near the pipe exit to determine the exact time of gas release and to monitor the flow conditions near the exit. We positioned a large welded truss supported by two barges over the plume to monitor, to measure, and to analyze the plume boil region and the zone of formation of the light-gas plume (Fig. 3). This truss was quite rigid in spite of the large span (28.5 m [93.5 ft]). The four gas sensors and a 6-m [20-ft] mast for the overhead video camera were mounted on the truss directly over the geometric plume center.