Keeping Disaster at a Distance 

To safely site potentially explosive materials, engineers use computer models to predict where building debris might fall after a catastrophic explosion.

By Patricia Moseley Bowles and P.A. Cox     image of PDF button

Patricia Moseley Bowles served as a principal analyst in the Engineering Dynamics Department of SwRI's Mechanical and Materials Engineering Division. She is a specialist in predicting the dispersion of fragments and debris following internal explosions in structures, and has been the DISPRE program manager since 1990. Bowles joined the staff of Applied Research Associates in San Antonio as a senior analyst in July 2000. 


P.A. Cox is a staff engineer in the Engineering Dynamics Department. A structural analyst with special interests in transient response, limit design, and the finite element method, most of his work has been in applied or basic research related to weapon and vehicle dynamics, ship tank design, and the evaluation and minimization of explosive hazards. 


Whenever ammunition is being stored, there exists the slight but daunting prospect of accidental explosions. These internal explosions can occur through impact, thermal changes, friction, static electric discharge, or the instability of aging munitions. But whatever the cause of such catastrophic accidents, the potential for destruction of property, equipment - or more importantly, human life - is sobering. 

This potentially explosive situation is being addressed by engineers and analysts in the Engineering Dynamics Department of Southwest Research Institute (SwRI), who have been developing and perfecting sophisticated debris prediction tools that can help to minimize the devastation resulting from the accidental explosions of ammunition stored in hardened aircraft shelters or similar magazines. 

Numerous studies have been conducted over the past 20 years to address the effects of accidental explosions. Many were conducted by SwRI for the Norwegian Defence Construction Service (NDCS) and the Klotz Group*, an informal international explosives safety working group. These efforts have led to the development of DISPRE2, an accurate and user-friendly PC-based code that can predict the likely blast field or hazardous debris "throw" outside aircraft shelters or munitions-storage structures. SwRI and its contractors hope it will become the ultimate debris prediction tool for the safe siting of structures containing explosives. 

With this critically important predictive tool, the safe siting of these shelters in relation to surrounding structures, equipment, and personnel can be expedited with a minimum of wasted space and a maximum degree of safety assurance. 

The relationship between launch velocity and scaled cover depth (scaled using the cube root of the charge weight) seems to follow a similar trend for a wide range of loading densities. The determination of similar relationships for debris launch angle and debris mass distribution was not only considered possible, but became a focused goal for explosives safety. These data could be used to develop, computerize, and refine the prediction of debris velocities, angles, and initial mass distribution to facilitate the calculation of debris effects for a variety of structures and to aid in the subsequent siting of the structures. 

Software history 

A 1981 study for NDCS correlated an approximate engineering analysis with the experimental results of 1:20 and 1:100 scale model tests of the explosive effects on a third-generation Norwegian aircraft shelter. The objective was to determine a method to predict the blast field characteristics outside the shelter and the maximum expected debris distances from fragmentation of such a shelter following an accidental explosion of ammunition stored in a chamber beneath the floor. 

In a follow-up phase, the objective was to use these test results to develop a model for better estimating the probability of lethality for humans from such an explosion. The reproducibility and directionality in these tests provided a reliable database, enabling engineers and others to predict how and where building debris might fall after such a catastrophic event. 

On any given site, this information vastly facilitates the placement of such storage magazines in relation to surrounding buildings, equipment, and the personnel who use them.

This goal was achieved only through years of testing using hardened aircraft shelters, which are specifically designed to resist blast and fragment effects. Throughout the 1980s, SwRI participated in several large-scale (1:4 to 1:3) to full-scale tests of these shelters, which were subjected to internal detonations. These programs contributed significantly to the growing database of hazardous debris throw and the air blast distribution following an explosion. 

Early in 1991, SwRI began another test program for NDCS to further study the debris effects from aircraft shelters, this time adding a protective rock rubble berm surrounding the structure. The initial velocities and angles at which concrete debris flew from an aircraft shelter following an internal detonation were carefully documented. This project included 1:15 scale model tests of three shelter designs: combined Norwegian/U.S., third-generation U.S., and third-generation German. 

In 1992, SwRI began developing a code to calculate initial debris parameters and throw from such detonations. The development of a computer code based on new and existing test data, as well as first-principle calculations for such parameters as launch velocity, angle distribution, and mass distribution, would eliminate the need to test each new design or modification. 

Reproducibility of breakup patterns and debris trajectories in 1:100 and 1:20 scale tests suggested that the prototype shelter would fragment in a manner similar to the rupture patterns observed in the tests, with differences in reinforcement considered. The measured external blast field was also similar. The location of concentrated debris after an explosion appeared to be quite directional in both scales tested and in full-scale tests conducted in the U.S. The reproducibility and directionality of debris throw aided engineers in the development of lethality probability curves for personnel in the vicinity of an accident.

The model 

This predictive model was developed by SwRI several years ago with funding from the U.S. Department of Energy (DOE) and the U.S. Department of Defense Explosives Safety Board (DDESB). The model was called DISPRE, for "dispersion prediction," and was approved as a siting tool for explosives processing and handling facilities in November 1990 by both DOE and DDESB. Various intermediate calculation models and three computer codes developed by the Naval Civil Engineering Laboratory  comprise DISPRE Version 1.0: SHOCK, short for shock loading; FRANG, short for frangible panel; and MUDEMIMP, the acronym for Multiple Debris Missile Impact Simulation. 

The DISPRE model has proven effective in reducing required siting distances for many explosive material processing structures when used within its constraints, based on the limits of the test data used to validate the model. Version 1.0 can be used to predict building debris throw for charge weights up to 120 kilograms (260 pounds) in a rectangular structure. 

Because of a lack of other accurate predictive models at the time, the model was frequently used outside its validated limits. Under funding from the Klotz Group, SwRI modified the DISPRE model, rather than simply extrapolating from it to predict throw. This newest model, known as DISPRE2, compares the differences between the internal loads and breakup of above-ground rectangular structures containing less than 120 kg (260 lb) of TNT-equivalent explosives, and arch-shaped magazines and rectangular above-ground magazines storing up to 5,000 kg (11,000 lb) of such explosives. 

Because the safe separation of magazines also depends on external air blast, DISPRE2 includes the prediction of air blast around the magazine. The software is based on an analysis of existing data and fundamental calculations and has been validated to a level of accuracy consistent with the existing test data - chiefly aircraft shelter breakup, debris, and air blast data. 

Version 1.0 predicted hazardous debris density - defined as no more than one hazardous debris piece per 55.7 square meters - and thus, safe siting distances, reasonably well over a fairly wide range of loading densities. More refined versions have expanded the number of structure types that can be analyzed and has improved the accuracy. Recent efforts have resulted in two versions of the software. The first is specifically for aircraft shelters. The second development version covers seven types of magazines, including aircraft shelters. DISPRE2 Version 2.9 contains both versions of the software. 

Where do we go from here? 

The upper limit of 5,000 kg (11,000 lb) within a storage magazine corresponds to a relatively low loading density for such a structure. At high loading densities - up to 500,000 kg (1.1 million lb) of explosives - the distance from the explosive charge to each structural component and the components' thicknesses also become important in determining failure modes and subsequent debris throw distances. 

When the charge amount is very high, the charge standoff can be quite small, resulting in close-in shock loading of the component. For typical magazine wall thicknesses, this type of loading causes catastrophic failure of the wall or a portion of the wall into small debris particles. Because typical stores for magazines other than aircraft shelters can well exceed 5,000 kg (11,000 lb) of explosives, the DISPRE2 model should be expanded to cover high loading situations and better meet the storage needs of the Klotz Group members. This expansion will require loading realm differentiation. 

SwRI has collaborated with the Ernst Mach Institut of Freiburg, Germany, to develop several program plans for the Klotz Group that include testing and collecting data, as well as performing analyses, to accomplish this goal. In the high-risk world of explosives storage, out of sight can never be out of mind. Mindful of the life and death importance of this mission, SwRI will continue working to develop predictive tools that are as fail-safe as humanly possible. 

*The Klotz Group, formerly known as the Klotz Club, is currently composed of delegates from Norway, Germany, The Netherlands, Sweden, Switzerland, and Singapore. Other participating countries have included the United States, the United Kingdom, and France.  NCEL is now the Naval Facilities Engineering Service Center of Port Hueneme, California.


Comments about this article? Contact Cox at (210) 522-2315 or

Published in the Summer 2000 issue of Technology Today®, published by Southwest Research Institute. For more information, contact Maria Stothoff.

Technics Summer 2000 Technology Today
SwRI Publications SwRI Home