会员注册 | 二级域名申请 | 我能做什么? | 网站说明书 | 协议书下载 | 广告预定 | 企业邮局 | 标准库 | 关于我们
免费法律咨询
首页 企业目录 产品目录 求购信息 二手设备 备品备件 行业资讯 行业论文 行业标准 技术专利 企业管理 行业书库 人才招聘 专家介绍 技术交流 友情链接 我的交易区
书籍目录索引
您当前的位置: 深冷技术网 → 行业书库 → 在线书籍 → 《气体爆炸手册(GAS EXPLOSION HANDBOOK )》


Introduction to Gas Explosions

 

We have all heard about accidental gas explosions and the destruction they can lead to. Fortunately most of us will not experience accidental explosions. But preventing them from happening requires a good understanding of what a gas explosion is and what one can do to reduce the frequency and consequences of such events.

The objective of this chapter is:

    i) to give an introduction to the field of gas explosions
    ii) to give an overview of loss experience
    iii) to show how we can use our knowledge to improve safety.

This chapter covers these aspects of gas explosions briefly. It is intended to be a first introduction to the field, and should be read the first time the handbook is used.


1.1 What is a Gas Explosion

We define a gas explosion as a process where combustion of a premixed gas cloud, i.e. fuel-air or fuel/oxidiser is causing rapid increase of pressure. Gas explosions can occur inside process equipment or pipes, in buildings or off-shore modules, in open process areas or in unconfined areas. When we are talking about a gas explosion as an event, it is a more general term. It is then common to include the events both before and after the gas explosion process, see the diagram below.


Figure 1.1. An event tree showing typical consequences of accidental releases of combustible gas or evaporating liquid into the atmosphere.

Figure 1.1 shows what can happen if combustible gas or evaporating liquid is accidentally released into the atmosphere. If the gas cloud, formed from the release, is not within the flammability limits or if the ignition source is lacking, the gas cloud may be diluted and disappear. Ignition may occur immediately, or may be delayed by up to tens of minutes, all depending on the circumstances. In case of an immediate ignition (i.e. before mixing with air or oxidiser has occurred) a fire will occur.

The most dangerous situation will occur if a large combustible premixed fuel-air cloud is formed and ignites. The time from release start to ignition can be from a few seconds up to tens of minutes. The amount of fuel can be from a few kilograms up to several tons.

The pressure generated by the combustion wave will depend on how fast the flame propagates and how the pressure can expand away from the gas cloud (governed by confinement). The consequences of gas explosions range from no damage to total destruction. The pressure build-up due to the gas explosion can damage personnel and material or it can lead to accidents such as fires and BLEVE's (domino effects). Fires are very common events after gas explosions.

When a cloud is ignited the flame can propagate in two different modes through the flammable parts of the cloud. These modes are:

    i) deflagration
    ii) detonation

The deflagration mode of flame propagation is the most common. A deflagration propagates at subsonic speed relative to the unburned gas, typical flame speeds (i.e. relative to a stationary observer) are from the order of 1 to 1000 m/s. The explosion pressure may reach values of several barg, depending on the flame speed (see Section 5.1).

A detonation wave is a supersonic (relative to the speed of sound in the unburned gas ahead of the wave) combustion wave. The shock wave and the combustion wave are in this case coupled. In a fuel-air cloud a detonation wave will propagate at a velocity of 1500-2000 m/s and the peak pressure is typically 15 to 20 bar.

In an accidental gas explosion of a hydrocarbon-air cloud (ignited by a weak source - a spark) the flame will normally start out as a slow laminar flame (see sections 2.12 and 4.10) with a velocity of the order of 3-4 m/s. If the cloud is truly unconfined and unobstructed (i.e. no equipment or other structures are engulfed by the cloud) the flame is not likely to accelerate to velocities of more than 20-25 m/s, and the overpressure will be negligible if the cloud is not confined.

Figure 1.2. Gas explosion in a partly confined area with process equipment.

In a building or in an offshore module with process equipment as shown schematically in Figure 1.2, the flame may accelerate to several hundred meters per second. When the gas is burning the temperature will increase and the gas will expand by a factor of up to 8 or 9. The unburned gas is therefore pushed ahead of the flame and a turbulent flow field is generated. When the flame propagates into a turbulent flow field, the effective burning rate will increase and the flow velocity and turbulence ahead of the flame increases further. This strong positive feedback mechanism is causing flame acceleration and high explosion pressures and in some cases transition to detonation.

In a confined situation, such as a closed vessel, a high flame velocity is not a requirement for generation of pressure. In a closed vessel there is no or very little relief (i.e. venting) of the explosion pressure and therefore even a slow combustion process will generate pressure (constant volume combustion, see section 4.9).

The consequences of a gas explosion will depend on:

  • type of fuel and oxidiser
  • size and fuel concentration of the combustible cloud
  • location of ignition point
  • strength of ignition source
  • size, location and type of explosion vent areas
  • location and size of structural elements and equipment
  • mitigation schemes

Gas explosions may be very sensitive to changes in these factors. Therefore it is not a simple task to estimate the consequences of a gas explosion.


1.2 Loss Experience

If we review the annual list of accidents in the Loss Prevention Bulletin (IChemE) we will find that there are many serious explosions each year. In addition there is a large number of minor explosions or near-accidents which are never reported.

Garrison (1988), has reviewed the hundred largest losses in the hydrocarbon process industry, from 1957 to 1986 (see Figure 1.3). He found that 42% of these accidents were caused by vapour cloud explosions. In his classification vapour cloud explosions include gas explosions within buildings as well as outdoors (unconfined explosions). Events classified as explosions constitute 22%. These explosions are probably run-away reactions, explosions in solids, BLEVE's, loss of containment, and gas explosions internally in process equipment.


Figure 1.3. Distribution of types of loss for the 100 largest losses in the hydrocarbon process industry from 1957 to 1986 (Garrison, 1988).


When we look into the details of the individual accidental explosions that have happened, we will find a large variety in size of the explosion and loss experience. From accidental records we can learn that gas explosions have a tendency to repeat themselves in similar conditions. It is therefore important to investigate accidents, report the findings in open literature and take corrective actions.

Flixborough, 1974

The explosion in the Nypro plant at Flixborough on June 1, 1974 is one of the most serious accidents in the history of the chemical industry. At Flixborough, the plant was totally destroyed and 28 persons were killed and 36 others were injured on site. Outside the plant, 53 persons were reported injured and 1.821 houses and 167 shops suffered damage. The cost of the damages was over 100 mill. dollars (Theodore et al., 1989). The cause of the Flixborough explosion was a release of about 50 tons of cyclohexane, probably due to failure of a temporary pipe. The flammable cloud was ignited about 1 minute or so after the release. A very violent explosion occurred. The blast was equivalent to an explosion of about 16 tons of TNT. The characteristic of the gas explosion at Flixborough is that the dense fuel (cyclohexane) was able to form a huge flammable gas cloud and that the confinement and obstructedness within the plant were causing high explosion pressures. From the Flixborough incident we can learn i) if the inventory had been smaller, the flammable gas cloud would have been smaller , i.e. reduce inventory ii) control of plant and process modification is important iii) use blast resistant control rooms and buildings.

Brahegatan, 1983

On March 3, 1983 there was a hydrogen explosion in an open street in Stockholm (Persson, 1984). The event occurred when gas cylinders where unloaded from a lorry and hydrogen suddenly started to leak out. The hydrogen was stored in a bank of 18 cylinders which contained about 10 kg. of hydrogen. The blast wave from the explosion broke windows in a range of about 90 meters. 16 persons were injured. From the Brahegatan incident we can learn that hydrogen is very reactive and even in open areas explosions with hydrogen can be very violent.

West Vanguard, 1985

At the night of October 7, 1985 a blow-out occurred on West Vanguard when the rig was drilling at Haltenbanken in the Norwegian sector (NOU 1986:16). The escaping gas was sucked into the engine room and a very violent gas explosion occurred. The side wall of the engine room was blown open and one man was killed. As is typical for such an accident, a fire followed the explosion. Fortunately, the main integrity of the rig was not seriously damaged and the rest of the crew was rescued. From the West Vanguard incident we can learn that escaping gas can be sucked or diffused into confined areas through ventilation ducts. Location of air intakes should therefore be carefully chosen.

Ice cream factory

In an ice cream factory the refrigerating system containing ammonia was leaking during a fire. Suddenly the whole basement exploded and the building was partly destroyed. From this incident we can learn that even substances like ammonia, burning very slowly, can cause severe gas explosions if they explode inside a confined area.

Devnya, 1986

There was a very serious explosion and fire in a PVC factory in Devnya, Bulgaria, 1986 (NTB, 1986). The accident was caused by a pipe failure. The pipe had not been X-rayed. 17 persons were killed in this accident, among them 8 women working in the laboratory. From the Devnya incident we can learn that: i) inspection is a key factor to safe operation ii) all activities not absolutely necessary for the operation of the plant (such as laboratory work) should be removed from potentially hazardous areas.

"Berge Istra "

On December 30, 1975 the oil/ore carrier M/S "Berge Istra" (Solberg, 1981) sank in the Molucca Sea. Two of the crew were rescued . They reported a rapid series of three massive explosions followed by immediate sinking of the ship. In October 1979, the sister ship M/S "Berge Vanga" disappeared in the Atlantic Ocean. Practically nothing is known about that incident. No-one was rescued.

The rapid sinking of "Berge Istra" indicates that a gas explosion in the double bottom of the ship ripped the ship structure open and water flooded the double deck and the engine room.

From the "Berge Istra" event we can learn that: i) gas explosions can damage the integrity of large constructions, like a supertanker, and ii) a flammable gas cloud in a confined volume, like the double bottom of a ship, will easily generate damaging pressures.

Road accident

Two persons were driving in a car with a plastic bag filled with oxygen/acetylene. After 4 to 5 km the bag exploded. The two persons in the car were fairly lucky, their only injuries were ear drum ruptures and some hair burnt off. The car suffered damages for NOK 60 000. The two persons intended to have some fun by making a "bang". One of them, a car mechanic, filled acetylene and oxygen from an acetylene torch in a plastic bag, when he dropped by the garage where he was working. This episode may sound to be an uncommon one, but it is not. In Norway, during the last five years, we have heard about two other explosions due to people carrying or trying to make "bangs" in a similar way as the two persons in the car. From this road accident we can learn: i) one should not play around with premixed combustible gas; it is very dangerous ii) most people do not have the slightest idea of how dangerous combustible gas can be if it is not handled with care.

Piper Alpha, 1988

Piper Alpha is the "Flixborough accident" in the off-shore industry. At Piper Alpha a rather small gas explosion in a compressor module caused fires which subsequently resulted in rupture of the riser. The main part of the platform burned down. 167 people were killed. The gas explosion overpressure was calculated by the FLACS code (Lord Cullen, 1990) to be about 0.3 bar for the most likely gas cloud. From the Piper Alpha incident we can learn that a gas explosion can easily result in domino effects and loss of control. Installations should be designed to avoid domino effects.

Deodorant factory

Due to environmental concerns, freon as a driver gas in deodorant atomisers, was changed to butane. After a short time of production, the main part of the factory was destroyed by a gas explosion (Anon 1986). From the deodorant factory accident we can learn that process modification must be controlled.

Port Hudson, Missouri, 1970

In this incident, liquid propane was released from a pipeline. The gas cloud flowed into a valley and about 20 min after the release started, the gas cloud exploded violently. The explosion was probably a detonation. The explosion started as an internal explosion in a pump house and this triggered the unconfined cloud to detonate. From Port Hudson we can learn that explosions in confined areas can initiate detonations causing high pressures in unconfined areas.

Rafnes, 1988

The incident at Rafnes, Norway, in 1988 is known as a large fire. However, the first incident was actually a gas explosion. The persons sitting in the blast resistant control room, felt that the whole building was shaking. There was no major damage due to the explosion, and no one was injured. The explosion pressure was likely in the order of 100 mbar or so.

The Rafnes plant was designed with blast resistant buildings. If the release that occurred at Rafnes had happened in an old-fashioned plant, it is very likely that the consequences of the gas explosion would have been quite different. This is an example of protection that worked as was intended. From the Rafnes incident we can learn: i) control rooms and buildings should be blast resistant, ii) fire is a common event after an explosion, and iii) it is possible to build a blast barrier that can mitigate and protect against the consequences of a gas explosion.


1.3 Analysis and Management of Gas Explosion Safety

Loss experience shows that prevention of gas explosions by reducing the risk of accidental releases, formation of explosive clouds and ignition only, is not sufficient. The frequency of gas explosions is still not low and the consequences of a gas explosion can be dramatic. Therefore, we have to build in a last barrier against gas explosions in our facilities. That can be done by performing safety analyses and by following good engineering practice. By doing so, the risk of gas explosions can be strongly reduced. The objective of this section is to discuss how we can apply the knowledge of gas explosions and the tools for predicting such incidents to make this last barrier effective.

Pappas (1990) discusses a Norwegian model for managing safety in off-shore development projects. Some examples of the activities in safety management presented by Pappas, are shown in Figure 1.4.


Conceptual Study
  1. Main layout settled / Separation checked with regard to loads.
  2. Module shape settled / Explosion venting.
  3. Estimate Accidental loads.
  4. Degree open - closed wall / Consistent with HVAC.
Detailed Engineering
  1. Final calculation of blast loads to be used in Accidental Load Specification.
  2. Ensure that blast loads are included in relevant specifications such as HVAC package, fire walls, gratings, closed decks etc.
  3. Blast vented from one module may affect other modules. Check interfacing contractors.
Fabrication and Installation
  1. Check that fabrication contractors have understood the reasons for design and functional requirements.
  2. Check blast panel specifications and installation.
  3. Check vent panel specifications and installation, i.e. weight, fastening mechanism and blocking of opening area.
  4. Check if deformation of walls will have unacceptable consequences for unplanned pipes
Figure 1.4. Examples of activities in safety management in an offshore development project. (Pappas, 1990).

In development projects, gas explosion hazards should be taken into consideration from day one. It is in the early phase of the development project (i.e. conceptual study) that major decisions such as location of different areas, separation of areas and overall layout (that will influence the vent arrangement and the process itself) are made. In the detailed engineering phase the final calculation of gas explosion loads is one important activity. In the fabrication and installation phase, checking that design is followed is a main activity. In all these stages, it is important to have good understanding of gas explosion hazards and to apply simple guidelines and good engineering practise.

To estimate consequences and loads from gas explosions is often part of a risk analysis. As shown in Figure 1.5, a typical risk analysis consists of 5 elements. The risk consists of the frequency and the consequences of an event. (Risk = frequency * consequence.)


Figure 1.5. Risk analysis (Ramsay 1990).


The elements in consequence evaluation are shown in Figure 1.6. When we are using FLACS for simulation of gas explosions, the FLACS simulation is one part of the consequence evaluation. Definition of scenario (i.e. size of gas cloud, ignition location, vent arrangement etc.) is a very important part of the consequence analysis. Definition of scenario is also related to the frequency estimates. One example is that the larger the gas cloud, the lower is the frequency of its occurrence. In gas explosion analysis, results of gas explosion simulations are very sensitive to certain parameters. One of them is the ignition point location. In some cases, by moving the ignition point and keeping the other parameters constant, the explosion may change by orders of magnitude in pressure. This is a fact that should be recognised when consequence and risk analyses are made.


Figure 1.6. Consequence evaluation of gas explosions.


The benefits we can obtain from a consequence analysis are:

  • assessment of risk in formal risk assessment studies
  • improved design and operation
  • supporting decision making
  • transfer of knowledge
  • cost benefit
  • safety
注意:本栏目所有书籍均来至于网络,如果其中涉及了您的版权,请及时联系我们,我们在核实后将在第一时间予以删除!

Copyright©2001版权所有_杭州汉皇网络科技有限公司 联系我们:webmaster@cngspw.com  浙ICP备10209442号-1 ICP经营许可证 浙B2-20100450
服务热线:0571-85065806  传真:0571-85065896 地址:杭州下城区高新技术产业基地电子商务园区费家塘路588号4号楼402-403室
主办单位:杭州汉皇网络科技有限公司  本网站法律顾问:汪卓君律师(浙江杭天信律师事务所)
cngspw.com(hangzhou_china),Ltd;Allrights Reserved 版权声明  

execute:39.063