Click for next page ( 6


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 5
Chapter 2 INVESTIGATIVE METHODOLOGY NATURE OF EXPLOSIONS The investigation of dust explosions begins with a basic understanding of what 16 involved in the explosion process. The best investigation plan i8 based on prior knowledge of the probable chain of events involved in the explosions. Although there are differences in explosion patterns at different locations, there is a general pattern that is repeated over and over again. Events at different locations often differ only in minor details. Dust explosions occur when structures rupture due to the pressure generated by the very rapid combustion of suspended dust in air after - ignition by a source with energy sufficient to ignite dust. In most cases ignition sources with energies of at least 10 to 20 millijoules are required, but, upon investigation, one finds that the primary ignition sources have generally many times the minimum energy required. When dust burns in a confined space in air, pressures of 80 to 120 psig are attained. In most cases the structure explodes since the most common membrane materials used in the construction of grain elevators fail at pressures ranging from 2 to 3 psig up to 30 psig (Brasie 1979~. As the combustion proceeds, the pressure in the building and all the interconnected spaces increases at a rate that is a function of the type of fuel, the amount of fuel per unit volume, and the size of openings or vents that permits the pressure to be relieved. Whether or not a structure ruptures is then simply a function of the competition between the rate of pressure increase and the ability of the vent spaces to keep the pressure from rising above the failure point. Unfortunately, the basic design in most grain elevators is such that most of the confined spaces (for example, tunnels and legwells) have virtually no vent area. Some headhouses are virtually windowless, and combustions originating in those headhouses generally cause partial if not complete destruction. There is no "explosion" until some part of the structure actually ruptures. The boom or noise heard exterior to the exploding structure is the noise caused by the air returning to the reduced pressure zone created by the explosion. In very large, complex structures, particularly those found in terminal or export elevators, the volumes of the interconnected spaces are very large. Tunnels running underneath silos may be hundreds of 5

OCR for page 5
1 6 feet long. Galleries are of comparable lengths. The interior design of headhouses varies from some in which the bucket elevators are in steel enclosures in open areas to those in which the elevation equipment is inside hollow concrete wells or shafts. All elevators contain storage bins, either for primary storage or for use as "working bins, that are generally within the framework of the headhouse. Since the purpose of the elevator is to move grain into and out of all of these places, fire and pressurized combustion products can go through the various interconnections and reach many of the bins and working spaces. Since the rate of.flame propagation. and movement of pressurized gases is finite, an explosion in a given building complex may actually be a series of explosions with time intervals of fractions to greater than l second between them, depending on the distances and the precise circumstances. In the most general scenario a bucket elevator casing will explode because of a fire inside and the resulting overpressurization. The resulting fireball.that.is emitted causes air movement, which raises the local dust that is lying around. This suspended dust burns, causing a greater fire and the acceleration of the movement of air throughout the facility, thus increasing the combustion zone volume. The hot gases generated can penetrate up, down, or laterally in the elevator complex, depending on the initial location. For example, the combustion front may traverse to the top of the headhouse, down along the gallery and into empty bins where new dust "fuel" may be found.. One or more bins may then explode, generating more combustion gases, which then may enter the tunnel and proceed in both directions down the tunnel, enter more bins through spouting or perhaps proceed to other sections of the elevator complex. The entire process can easily last several seconds. Sturdier parts of the elevator may sustain pressures above 10 or even 20 psi. Other portions of less rugged construction may fail at lower pressures. When self-venting occurs, i.e., rupture of membranes to the exterior, the rate of expansion of the combustion gas is Slowed down substantially and penetration along further interconnected spaces may not occur, particularly if there is insufficient fuel to be suspended to keep the concentration above the lower flammability limit of that particular dust combination. . One cannot predict with any accuracy what the pattern will be in a given elevator complex. However, an investigation of the normal amount of ambient dust in different locations (which is a.function of housekeeping and dust collection procedures) will give some indication of the potential for an explosion and the likely extent of damage. The.amounts and location of ambient dust that survive the fire and explosion process is sometimes a valuable clue in helping to pinpoint the original source. Sometimes flame front edges may be seen along dusty walls or on other surfaces delineating the size and extent of flame propagation. Obviously, if a flame front dies partway down a tunnel the origin of the flame must be at the burned end of the structure. These types of clues may be helpful in pinpointing the . Origin. The above remarks are intended to help clarify what i s a dust ~~explosion" in a grain-handling complex. Knowing that the basic process i s

OCR for page 5
7 one of pressurization with outward movement of gases towards available openings helps the investigator understand from which direction the pressurization occurred. In the cases in which an exploalon reaches shock-speed levels, the propagation of the shock wave radially away from the source may be quite evident by the deformation of structural members in the path of the expanding high-speed shock wave. Hence, it is fairly important that the investigator have a mental picture of what the physics of the explosion process were in order to help locate the origin and identify the directions of propagation. , METHODOLOGY The purpose of the investigation is to identify the original source of ignition and couple this with information concerning the available dust loading (i.e., quantity of layered dust per unit volume) to describe the explosion event in appropriate detail. Many different sources of facts must be investigated in order to put together a coherent, logical and defendable scenario for the explosion. There are several basic mechanisms: - 1. The direct investigation of the site and examination of all of the areas and identif ication of those clues and pieces of information that help in describing the basic explosion process at this site. Interviewing and talking with anyone who can provide some information on the event. These people include operating personnel on the site, people in nearby buildings such as offices, homes, or other industrial facilities, and even passers-by. 3. The post-explosion examination of individual items in a . laboratory. This step might include, for example, the investigation of marks, fracture surfaces, deformation of parts, or things related to electrical components. In this phase, care must be taken to determine whether the damage to the individual components being examined resulted from the explosion process being investigated or from some earlier event. Individual judgment must be used in these examinations. The general approach to investigation is to have a team of experienced investigation people. The tasks of the specific investigation are logically divided into the areas discussed above. Depending on the circumstances and the availability of witnesses, it may be desirable to have one or two team members do all the appropriate interviewing. Interviewing is not without its-difficulties since witnesses may assume that an investigator is an adversary. In some cases it may not be possible to determine anything because of the potential for litigation. Site Investigation Those directly involved in looking at the structure and surroundings have a complex task. It is desirable to meet first with the owners and

OCR for page 5
8 ~ managers of the facility and local officials Deco apprise them of the mission and intent of the investigating team. The team can establish rapport with management by stating who they are, why they are at the facility, and what they plan to do and accomplish. After their preliminary investigation the team should once again meet with management to share information and findings. This will be helpful in the event that the team wishes to revisit the damaged site to witness, for example, salvage operations. . . If possible, the team should obtain a building or drawing plan showing the various spaces and interconnecting design. The layout of the complex is sometimes not obvious if destruction is severe. Needless to say the investigating team should be properly prepared from the standpoint of safety. Hardtoed.shoes, protective clothing, and hardhats are mandatory. Gloves should be worn. Excellent, portable lighting is extremely desirable. A powerful six-cell focusing flashlight is invaluable since there is often no power at the site and at night extra illumination is always needed. Furthermore, tunnels, the inside of bins, and galleries may be dimly lighted, if at all. The investigative team should have.its own liability insurance and should assure management that the team is participating in the investigation at its own risk. Rules and restrictions of local officials, such as fire marshals or other safety people, should be followed. it may be necessary, however, to negotiate with these officials if the restrictions are arbitrary and unnecessarily severe. In some cases there is a preoccupation with immediate removal of rubble and wreckage to clean up the site. It is desirable to photograph and examine the wreckage before rubble and wreckage are removed, unless there is a time consideration involved (for example, injured personnel in the wreckage). The site investigative team should go through the entire structure systematically from top to bottom.and from end to end as accessibility permit s. Photographs should be taken of any places on ache site that have any possible c rues. A record of the location of the photographs should be maintained as they are obtained. It is desirable, if possible, to indicate the location of photographs on a site or elevation drawing. There are many things for the investigating team to seek. Some of the important clues include the direction rubble is thrown and the extent.of damage of various structural elements such as I-beams, concrete walls, reinforcing bars, or other items. In locations such as tunnels or galleries, where there may be light construction items such as spouting, particular attention should be paid to the extent of deformation and indications of the direction of the pressure wave. Even small details should not be overlooked, such as which pieces of rubble or wreckage are on top of which other pieces, whether glass is blown in or out, and whether the roof has been lifted.and resettled. Generally, after several tours through the accessible parts of the complex the basic explosion pattern should be discernible and the point of origin of the combustion can be determined. It is then important to go back

OCR for page 5
9 to the area of origin and look for possible sources of ignition. Unfortunately, the damage in the area of origin may be so bad that it is not possible to find the critical evidence before clearing away the debri 8. Bucket elevator casings, buckets, belts, etc. may be buried under tons of wreckage. For example, the bottom bearing in the boot, which might be suspected as an ignition source, frequently is not accessible. In some Cases the boot not only may be under wreckage, it may be under water from the firefighting action. In such cases, cooperation with those involved in removing the wreckage or demolishing the facility is required to be sure that at the time the critical elements are exposed someone is available to obtain them for investigation before they are thrown away. It may not always be easy to obtain this cooperation. The second phase of the initial investigation can be described as looking for small details. This phase may include the collection of pieces and parts, (for example, components of electrical systems) or other things that may require detailed off-site examination. At the site the team should meticulously go through the wreckage and note minute details of the placement and location of wreckage with respect to the sequence in which the items are piled on top of each other. The team should carefully look for and examine evidence of flame fronts and the direction in which the flame fronts advanced. For example, frequently lights in tunnels will show char markings on one side of the light only. The deformation, bending or twisting of light-gauge metal objects is often a clue to the direction from which forces were applied. Photographs of these details are extremely valuable and sufficient records should be kept to indicate the location of these items. - - Frequently the failure of electric components may be judged to be the cause of the initial ignition. However, one must be careful to,be sure ~.~- broken or damaged electric components show positive evidence that they failed prior to the explosion. It is easy to misjudge a smashed, shorted-out section of conduit or wire as being responsible for the ignition when in fact the damage was done during the explosion process. The placement and layering of rubble, as mentioned before,-can provide important clues as to'the sequence in which different portions of structures failed. Damage exterior to the main structure may also provide clues in some cases. Differences in damage adjacent to various openings in the main elevator structure can provide clues to the intensity of the pressure wave coming from the various apertures. This in turn can provide insight into the direction and strength of the pressurization process during the main explosion events. Although these comments are broad and general, once some field experience is developed, it becomes relatively easy to sort out the various factors and to begin to pinpoint the probable initial zone of the explosion process. It is then important to try to estimate what was the single initial f ire-explosion event that started the whole sequence. The experience of the investigating subpanel was that this was a discernible event in almost all of the cases investigated.

OCR for page 5
10 Sometimes the ignition source is a factor external to the basic operation of the terminal itself. For example, the leakage of propane into parts of structures associated with grain elevators and the subsequent ignition of the propane were the first steps in the destruction of a facility. The elevator, primarily of wood, was burned to the ground leaving no evidence of blast effects that may have occurred due to an initial dust explosion. However, from later examination and information provided by people associated with the operation, it was determined that there was an initial leakage of propane--from a line recently installed to provide fuel for a dryer--into an adjacent structure where ignition occurred during grain loading and unloading operations. The propane explosion propagated into the elevator proper and the rapid expansion of the propane fireball initiated a dust explosion. Indeed a dust explosion occurred and did heavy damage to the elevator, which, being made of wood, was soon totally engulfed in flames and destroyed. In a case like this, one must look at the operation of a total elevator complex to include not just those things associated with the movement or processing of grain but all those elements that are required to operate the complex. Interviewing of Witnesses . _ Another major portion of the investigation is the interviewing of all witnesses who may have information leading to the determination of the cause of the explosion. Those who should be interviewed include operational people directly on the site? people in nearby offices or other functional buildings, truck drivers or railroad employees who may have been in the vicinity, maintenance people, people who may have been driving by or who live or work nearby. When an approximate timetable of the events leading up to the explosion has been determined, it may be advisable to -id ~ -- some of the people in order to fill in details as the total pit: ~ -:_ to emerge. Hence, several different interviews may be desirable with various people. It is particularly desirable to interview people one at a time and to the extent possible not reveal statements of one individual to other witnesses. After the first round of interviewing, it is important to try to reconcile the stories and viewpoints expressed by those interviewed. Discrepancies in descriptions of what happened will frequently occur and need to be resolved. The resolution of discrepancies generally requires very tactful reinterviewing. Witnesses may have hazy recollections and may exaggerate or present faulty opinions. They may collaborate on a "safe" scenario (i.e., nonincriminating). Therefore, to get the best cooperation from all concerned, it is important to convey to the interviewees that the interviews are nonpunitive and nonthreatening. Other sources of information are those involved in the emergency processing after the explosion. They include firefighters, police or other local officials, and those who have treated the injured in cases where injuries occur. Frequently information is passed from those injured to those helping, such as ambulance drivers. The subpanel noted that at later

OCR for page 5
11 times injured witnesses have a tendency to be nontalkative, particularly there is any perceived liability threat. Witnesses tend to be vocal and cooperative 600n after the event but tend to become more reticent as time pa 8Se8 ~ - After interviewing, an integration session should be held in which all of the investigators pool the information obtained. At this time, it may be possible to construct a scenario that appears to be reasonably verifiable. However,-discrepancies frequently exist and it may be that no specific scenario can be stated.- It is then necessary to go back for additional site investigation or investigation of pieces or parsecs that may supply ache missing clues and to conduct additional interviews to try to f ill in the gaps. In some cases it may not be possible to pinpoint the actual initial event until rubble and wreckage are removed from key parts of the equipment that may yield clues. The removal may take place within hours or may be delayed for weeks for various reasons. If the original investigators cannot be present when the rubble is removed, the cooperation of some local person in segregating key evidence is important. Management personnel, if cooperative, may serve this function, or OSHA field people may be available. Insurance people may also want to be involved. Laboratory Investigation In some cases the critical evidence from items obtained from the site just after the explosion or when the rubble is cleared cannot be ascertained by visual examination alone. Laboratory tests may be necessary to determine fracture mechanisms, burn sequence, etc. It is important that appropriate laboratory facilities be available to the investigating team. The interpretation of laboratory data, of course, requires professional Judgment and correlation with the rest of the information from the explosion site. A report is the logical output of an investigation. Such a report should document the data and logic that led to the conclusion on probable chain of events. Photographs, site drawings, or related graphical information are valuable portions of such reports. Reports issued by the National Transportation Safety Board (NTSB) on transportation disasters, pipeline explosions, etc. are good models to follow. The format of the NTSB report is described in an earlier report by the panel (National Materials Advisory Board 1980~.

OCR for page 5