Cover Image

PAPERBACK
$52.00



View/Hide Left Panel

5
Use of Predictive Modeling Packages for Effective Emergency Management*

Nikolai Petrovich Kopylov and Irek Ravilevich Khasanov, All-Russian Scientific Research Institute for Fire Protection (VNIIPO) of the Russian Ministry for Civil Defense Affairs, Emergencies, and Elimination of Consequences of Natural Disasters (EMERCOM)

INTRODUCTION

About 1,000 major disasters and catastrophes occur each year in Russia. As a result of industrial and other technogenic1 accidents alone, more than 200,000 people annually are injured or mutilated and more than 50,000 are killed (including traffic accidents). The economic losses from technogenic and natural disasters total 6 to 7 percent of the country’s gross domestic product.2

An analysis of terrorist acts indicates that providing antiterrorism protection for facilities at risk of fire or explosion is the most urgent and important aspect of guarding against terrorism of a technogenic nature. Given these conditions, the effectiveness of management decisions made in eliminating the consequences of acts of technogenic terrorism largely depends on informational and analytical support and predictions of how fires and emergencies might develop.

The primary goal of the integrated state system for predicting and eliminating the consequences of extreme situations is to integrate the efforts of executive branch agencies at both the federal and the Russian Federation subject levels. The main objectives of activities under the state system are as follows:

*

Translated from the Russian by Kelly Robbins.



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 32
5 Use of Predictive Modeling Packages for Effective Emergency Management* Nikolai Petrovich Kopylov and Irek Ravilevich Khasanov, All-Russian Scientific Research Institute for Fire Protection (VNIIPO) of the Russian Ministry for Civil Defense Affairs, Emergencies, and Elimination of Consequences of Natural Disasters (EMERCOM) INTRODUCTION About 1,000 major disasters and catastrophes occur each year in Russia. As a result of industrial and other technogenic1 accidents alone, more than 200,000 people annually are injured or mutilated and more than 50,000 are killed (includ- ing traffic accidents). The economic losses from technogenic and natural disasters total 6 to 7 percent of the country’s gross domestic product. 2 An analysis of terrorist acts indicates that providing antiterrorism protection for facilities at risk of fire or explosion is the most urgent and important aspect of guarding against terrorism of a technogenic nature. Given these conditions, the effectiveness of management decisions made in eliminating the consequences of acts of technogenic terrorism largely depends on informational and analytical support and predictions of how fires and emergencies might develop. The primary goal of the integrated state system for predicting and eliminat- ing the consequences of extreme situations is to integrate the efforts of executive branch agencies at both the federal and the Russian Federation subject levels. The main objectives of activities under the state system are as follows: *Translated from the Russian by Kelly Robbins. 2

OCR for page 32
 PREDICTIVE MODELING PACKAGES FOR EFFECTIVE EMERGENCY MANAGEMENT • Monitoring and predicting extreme situations • Training specialists in emergency prediction and response • Educating the public on actions to be taken in emergencies • Developing preventive measures to reduce the risks and lessen the con- sequences of emergencies • Improving the management of emergency prediction and response measures Effectively accomplishing these objectives is impossible without utilizing new information and telecommunications technologies.3 The National Crisis Management Center (NCMC) has been created in Russia to unite the information resources and functional capabilities of local subsystems of the integrated state system with the aim of improving the quality and timeli- ness of management decisions on predicting and eliminating the consequences of emergencies. The NCMC is a geographically distributed information man- agement complex with peripheral elements that make it possible to manage the forces, means, and resources of the integrated state system and civil defense entities during crises and emergencies.4 SITUATION MANAGEMENT CENTERS The rapid development of information technologies has led to the appearance of massive amounts of informational, communications, audio, and video data that must be recognized, structured, and analyzed in order to make competent management decisions. Meanwhile, although the rates of information technology development have increased, the amount of time allotted for making management decisions is being reduced, especially for decisions made in crisis situations. The strategy for creating and developing national security support systems by states attests to the fact that information and management centers created at the national and regional levels and in major cities represent the universal foun- dation for the crisis management system. Informational support for such centers is provided by services such as 911, 112, and 01, as well as by scientific and academic centers.5 Intensive efforts are under way to apply modern concepts for the creation of crisis management centers involving high-technology equipment for commu- nications and information exchange, depiction, and processing, which helps in efficiently preparing and making well-founded management decisions. Situation centers have been created in Moscow and regional centers in the various ter- ritorial agencies of EMERCOM. These centers are complexes of programmatic and technical resources housed in special facilities where emergency response officials may assemble if an emergency arises. The situation centers regularly conduct training exercises, some of which involve members of the commission on extreme situations.

OCR for page 32
 COUNTERING TERRORISM The activities of a situation center represent the most expedient means of implementing the decision support system based on technologies for numerically simulating and creating visual representations of situations and object behavior. They are the top level in the system for managing the organization, the industry, the region, and the country. The situation center is an information analysis system that makes it possible to assess the real status of the object or event being managed, detect trends as external and internal changes develop, and analyze (simulate) possible conse- quences of management actions.6 From the most general standpoint, the situation center (room or hall) could be called a facility from which ongoing emergencies are observed or possible situations are analyzed. However, such an interpretation fails to take many factors into account. The modern understanding of a situation center focuses on the entirety of programmatic and technical resources, scientific and mathematical methods, and engineering solutions for automating processes for situational depiction, numerical simulation, analysis, and management. 7 All of these means and methods make possible the following: • Providing information on matters where operational decisions are required • Visually depicting management situations to reveal cause-effect relation- ships for events being analyzed • Numerically simulating and conducting situational analyses • Effecting operational control over efforts being carried out by structural subunits • Verifying execution of decisions made The situation centers include various types of analytical support capabilities (programmatic, technical, linguistic, psychological, and so forth). The situation center has four basic levels: (1) scientific-mathematical, (2) engineering, (3) programmatic, and (4) technical. The scientific-mathematical level includes all scientific theories, methods, algorithms, research, and developments necessary for the activities of the other levels. It provides the foundation for determining the expediency of creating the situation center, defines the effectiveness of its operations, integrates various components, and rectifies errors in a correct and timely manner. The engineering level provides concrete solutions in the selection and development of devices and software. It includes the necessary technological and design calculations, numerical simulations, technical equipment, facilities, program specifications, work algorithms, and so forth. The programmatic and technical levels include the appropriate support necessary for the tasks and functions assigned to the higher levels to be carried out.8 The main feature of the situation centers that determines their name is situ- ational (dynamic) simulation. Prediction makes it possible to create scenarios based on analysis of the current situation and existing trends. The situation cen-

OCR for page 32
 PREDICTIVE MODELING PACKAGES FOR EFFECTIVE EMERGENCY MANAGEMENT ter allows managers to see newly arising threats in a timely manner and to take measures to counter them.9 THE VNIIPO SITUATION CENTER The VNIIPO Extreme Situation Modeling Center (Situation Center) was established at VNIIPO in 2006 based on the requirements of the Concept for the Creation of the National Crisis Management Center. Functionally, the center is a part of the NCMC. The VNIIPO Situation Center is designed to provide informational, analytical, and expert support for management decisions by of- ficials from operations management agencies in responding to major fires and technogenic emergencies at critically important sites. The center’s primary tasks are as follows: • Collecting, accumulating, and analyzing information on the status of facilities at risk of fire or explosion and on EMERCOM forces, means, and reserves • Providing informational, analytical, and expert support for management decisions on preventing and eliminating the consequences of fires and techno- genic emergencies • Predicting the development of fires and technogenic emergencies at criti- cally important facilities • Developing, implementing, and supporting software systems for man- agement and modeling at the Situation Center • Providing technical documentation for numerical simulation packages for fires and emergencies and organizing and supporting work to develop models and methodologies to facilitate the Situation Center’s activities • Organizing the operations and information security of the Situation Center • Developing and supporting technical and telecommunications services at the Situation Center and developing and maintaining informational support for data banks and databases Based on its purpose, functions, and tasks, an organizational-technical struc- ture including the following components has been proposed for the VNIIPO Situation Center: • Analysis • Applied software support • Information infrastructure • General-purpose software and hardware environment • Complex of special-purpose software and hardware resources • Information security subsystem

OCR for page 32
 COUNTERING TERRORISM SUBSYSTEM FOR INFORMATION SUPPORT FOR MANAGEMENT DECISION MAKING The Situation Center’s subsystem for information support for management decision making must be responsible for preparing guidelines and statistical information needed for making command decisions. Databases that have been developed and are being utilized successfully lie at the foundation of the func- tional complexes and tasks of the information support subsystem at the Situation Center. For example, VNIIPO has created a user version of the informational database “Fire and Explosive Hazards of Substances and Materials and Means of Extinguishing Them” (see Figure 5-1), which is used in more than 100 of EMERCOM’s State Fire Service branches. The database contains information on more than 12,000 substances and materials, including data on the fire and explo- sive hazards of substances and materials, means of extinguishing them, and the potential reactions of substances and materials if they should come into contact. VNIIPO has developed and is using several regions a geographic information system for decision support in operations management by local fire and rescue units involved in responding to fires and eliminating the consequences of emer- gency situations. This system provides informational support for the following types of activities: FIGURE 5-1 Screenshot from the database “Fire and Explosive Hazards of Substances Fig 5-1.eps and Materials and Means of Extinguishing Them.” bitmap

OCR for page 32
 PREDICTIVE MODELING PACKAGES FOR EFFECTIVE EMERGENCY MANAGEMENT • Reception and processing of fire (emergency) calls, including location and formulation of orders for dispatching personnel and equipment to handle them • Accounting and control of the status and deployment of equipment and weapons • Redeployment of units, depending on their operating regimes • Management of operations at the fire (emergency), establishment ac- cording to proper procedure of accounting of situation changes and use of person- nel and equipment, and registration of necessary information • Implementation of other measures aimed at ensuring service delivery according to established procedure and increasing the effectiveness of firefighters’ actions An automated decision support system for use by fire captains at the scene has been developed to provide operational information and analytical support for decision makers. This system automates the following processes: • Accumulation and storage of site data • Presentation in convenient form of information used by the fire captain in preparing operational decisions on managing firefighters’ actions at the scene • Calculation of potential fire situations • Calculation of personnel and equipment needed to extinguish fires • Calculation of delivery systems for means of extinguishing fires, includ- ing calculation of pump-hose system parameters • Preparation of typical command decisions • Preparation of operational documents • Creation and correction of databases SUBSYSTEM OF ANALYTICAL SUPPORT FOR MANAGEMENT DECISIONS The subsystem for analytical support of management decisions must facilitate numerical simulation and prediction of the development of fires and emergency situations. With the aim of studying major fires at dangerous production facilities or in population centers, a series of studies has been conducted to simulate major fires in open spaces.10 Based on this research, a numerical simulation has been proposed for the aerodynamics of the environment. It is based on nonstationary Navier-Stokes differential equations, taking into account the effects of turbulence, atmospheric stratification, smoke aerosol diffusion, and phase transitions caused by the presence of moisture in the surrounding air. Figure 5-2 depicts a smoke cloud formed over a fire with a radius of 5 kilometers and a maximum heat trans- fer of qm = 4.7·104 W/m–2. For several decades, the institute has been working to develop and apply

OCR for page 32
 COUNTERING TERRORISM FIGURE 5-2 Isolines of smoke aerosol concentrations (a) and convective flow structure Fig 5-2.eps (b) over a fire with a 5-kilometer radius 1 hour after ignition. bitmap mathematical modeling of fires in structures and buildings. Mathematical models are widely used in resolving questions of ensuring the safety of people during fires, designing evacuation paths, and creating fire alarm systems. The various mathematical models of fire development in structures (interior fires) fall into the following three categories:11 1. Integral mathematical models (first-generation models) 2. Zone mathematical models (second-generation models) 3. Field (computational fluid dynamics) mathematical models (third-gen- eration models) Integral fire models are limited to recording physical heat parameters at the level of average values (by volume or by heat-absorbing surfaces).12 Equations on the development of a fire describe the change in average volume parameters for the situation over time. The system of differential equations for the balance in the structure includes equations on the material and oxygen balance, equations on the balance of combustion products and inert gas, and an energy equation. An example of the successful use of the integral modeling method would be the study conducted by institute specialists of possible development scenarios for the fire caused by the crash of a Boeing-767 aircraft into the World Trade Center in New York City. Several fire scenarios were considered. The first group of scenarios simulated the combustion of jet fuel spilled from the plane’s fuel tanks, the second group of scenarios covered the burning of office furniture, and the third group focused on the combined burning of jet fuel and furniture. Figure

OCR for page 32
 PREDICTIVE MODELING PACKAGES FOR EFFECTIVE EMERGENCY MANAGEMENT 1200 1100 1000 900 800 700 600 T, C 500 Standard fire 90 t of kerosene 400 HC-curve 375 t of furniture 300 kerosene+furniture (x=0.1) kerosene+furniture (x=0.25) 200 kerosene+ furniture (x=0.25) 100 0 0 10 20 30 40 50 t, min FIGURE 5-3 Calculated dynamics of average volume temperature in a structure with Fig 5-3.eps various amounts of jet fuel. 5-3 shows the calculated dynamics of the average volume temperature in the structure given various amounts of jet fuel assumed in the fuel load. Based on the results of calculations of the joint combustion of spilled jet fuel and furniture, a quantitative estimate was made of the amount of fuel involved in the fire at the World Trade Center. This assessment agrees with the data from American researchers on the quantity of fuel onboard the planes just before impact. The development of a fire may be described in more detail with the help of zone models, which are based on the premise of the formation of two layers in a burning structure: (1) the upper level of combustion products (smoke-filled zone) and (2) the lower level of undisturbed air (free zone). Thus, the status of the gaseous environment in zone models is evaluated through the use of average thermodynamic parameters from not one but several zones, and the zone boundar- ies are generally considered movable. Zone models became widespread in simulating local fires in structures and

OCR for page 32
0 COUNTERING TERRORISM systems of structures with relatively simple configurations and having com- parable linear sizes.13 However, creating zone models requires making a large number of simplifications and omissions based on a priori suppositions on flow structure. Such a method is inapplicable in cases where information on this struc- ture that might be obtained experimentally is lacking, so consequently, there are no grounds for zone modeling. Furthermore, more detailed information is often required on the fire than just average parameter values for each layer (zone). Field (computational fluid dynamics) models are more powerful and univer- sal tools than zone models, inasmuch as they are based on a completely different principle. Several computer programs are currently available for field modeling, and they are fairly accurate in describing the rate, temperature, and concentration fields at the initial stage of a fire.14 Therefore, the field model is the best means of approaching fire modeling in complex and unique structures, for example, in transport tunnels. Figure 5-4 pres- ents optical density fields for smoke in a central vertical section of the Lefortovo Tunnel, which is shallowly situated in Moscow’s third transport ring. A study was carried out using a three-dimensional field model to predict the distribution of fire hazard factors in the tunnel both with and without antismoke ventilation.15 A traffic accident involving a truck and several passenger vehicles in this 18.2 × 5.2- meter tunnel served as the emergency situation for the purposes of the model. In this scenario, maximum theoretical heat exchange intensity of 100 megawatts was reached 15 minutes after the start of the fire. It was supposed that the fire would break out at the center of the tunnel; therefore, given the symmetry, one-quarter of the actual tunnel volume was modeled. Calculations in the model covered 750 meters of the tunnel’s length. The temperature fields in the horizontal section at the height of 1.7 meters are presented in Figure 5-5. It is clear that despite the smoke filling the evacuation paths, the temperature in the working zone up to the 240-second mark does not exceed the critical level of 343 kelvins. Smoke with a temperature of 343 kelvins reaches the height of 1.7 meters at the 300-second mark (Figure 5-5e). At this moment, the distance from the center of the fire at which the evacuation path is blocked because of increased temperature is 90 meters. Work on simulating fires at various types of facilities holds a significant place in prediction efforts at VNIIPO. Facilities involved in extracting, processing, and storing flammable and highly flammable liquids face a high risk of fire. In this regard, the institute has developed a software package to calculate fire and explo- sion hazard factors at such facilities. This software is intended for quantitative calculation of hazard factors and their consequences; visualization of calculation results in map format; and electronic communication of the results in the form of graphs, maps, and tables. Figure 5-6 presents a sample screenshot from this program. In addition to its work on predicting the development of fires and emergency situations, the VNIIPO Situation Center is also developing numerical simulations

OCR for page 32
1 PREDICTIVE MODELING PACKAGES FOR EFFECTIVE EMERGENCY MANAGEMENT a myu: 0.1 0.12 0.238 0.3 0.5 1 6 4 xc 2 0 0 20 40 60 80 100 120 140 160 180 200 zc yc b myu: 0.1 0.12 0.238 0.3 0.5 1 6 4 xc 2 0 0 20 40 60 80 100 120 140 160 180 200 zc yc c myu: 0.1 0.12 0.238 0.3 0.5 1 6 4 xc 2 0 0 20 40 60 80 100 120 140 160 180 200 zc yc d myu: 0.1 0.12 0.238 0.3 0.5 1 6 4 xc 2 0 0 20 40 60 80 100 120 140 160 180 200 zc yc e myu: 0.1 0.12 0.238 0.3 0.5 1 6 4 xc 2 0 0 20 40 60 80 100 120 140 160 180 200 zc yc FIGURE 5-4 Optical density fields for smoke in the central vertical section of the Lefor- tovo Tunnel at 60 (a), 120 (b), 180 (c), 240 (d), and 300 (e) seconds after combustion. FIGURE 5-4

OCR for page 32
2 COUNTERING TERRORISM a (3D) | 12 Aug 2003 | t: 300 343 400 500 600 10 5 cz 00 8 6 4 2 0 20 40 60 80 100 120 140 160 180 200 c x c y b (3D) | 12 Aug 2003 | t: 300 343 400 500 600 10 5 c z 00 8 6 4 2 0 20 40 60 80 100 120 140 160 180 200 c x c y c (3D) | 12 Aug 2003 | t: 300 343 400 500 600 10 5 c z 0 8 6 4 2 0 0 20 40 60 80 100 120 140 160 180 200 cx c y d (3D) | 12 Aug 2003 | t: 300 343 400 500 600 10 5 c z 00 8 6 4 2 0 20 40 60 80 100 120 140 160 180 200 cx c y e (3D) | 12 Aug 2003 | t: 300 343 400 500 600 10 5 c z 00 8 6 4 2 0 20 40 60 80 100 120 140 160 180 200 cx c y FIGURE 5-5 Temperature fields (in degrees kelvin) in a horizontal section at a height of Fig 5-5.eps 1.7 meters at 60 (a), 120 (b), 180 (c), 240 (d), and 300 (e) seconds after combustion.

OCR for page 32
43 PREDICTIVE MODELING PACKAGES FOR EFFECTIVE EMERGENCY MANAGEMENT FIGURE 5-6 Sample screenshot from program to calculate fire and explosion hazard Fig 5-6.eps factors. bitmap and software for process simulation of firefighting and emergency response. For example, the institute has developed a software package for calculating the per- sonnel and resources needed to extinguish fires involving oil, petroleum products, chemicals, and stable gas condensate in storage tanks, during pour-offs to storage ponds or transfers to railway tankers, and at technical pumping stations. The pro- gram takes into account the volumes and structures of the combustion sites, the properties of the flammable liquids, tactical and technical characteristics of the foam and water delivery equipments used in extinguishing fires involving oil and petroleum products, and the characteristics of stationary and mobile firefighting equipment. CONCLUSION Despite existing developments in the numerical simulation of fires and emer- gency situations, serious issues remain to be resolved with the use of mathemati- cal models in the work of the VNIIPO Situation Center. Based on an analysis of possible fire and emergency scenarios, a list of models in need of further refinement should be drawn up, and the need for creating new models should

OCR for page 32
 COUNTERING TERRORISM also be evaluated. After the models are selected, a series of studies needs to be carried out to verify them. A significant volume of work is also needed to adapt mathematical models for use in the Situation Center and to create algorithms and software packages. Making calculations to forecast fires and emergency situations is impossible without reliable data inputs on facilities. Data collection efforts must be organized and carried out, the information must be processed, and modern technologies and geographic information systems must be used to create a database on facilities at risk of fire and explosion. In using mathematical models of fires and emergency situations based on nonlinear, nonstationary, three-dimensional model systems (for example, field models of fires), it should be taken into account that numeri- cal solution of such systems requires tens of hours of computer time even using high-output processing technologies. Introducing new modern technologies for numerical simulation of emergency situations requires the following: • Improving the reliability of predictions to prevent and eliminate the consequences of emergency situations • Organizing comprehensive monitoring and information-processing ef- forts regarding the status of facilities, the environment, and natural and techno- genic phenomena that cause emergency situations • Developing mathematical models of the development of fires and emer- gency situations • Optimizing and facilitating timely correction of action plans and measures for preventing emergency situations as well as eliminating their consequences • Providing a modern level of technical capabilities to support the work of operations personnel, including network communications technologies and means of collecting, analyzing, and presenting information on emergency situations NOTES 1. Technogenic is used to refer to phenomena arising as a result of the development or deploy- ment of technology. 2. Vorobyov, Yu. L. 2005. Safety in Daily Activities (Aspects of State Policy). Moscow: Busi- ness Express, 376 pp. 3 . Faleev, M. I. 2002. Computer technologies in creating an information space for dealing with disasters and catastrophes. iBUSINESS 6:19-21. 4 . Concept for the Creation of the National Crisis Management Center. 2005. Moscow: Min- istry of the Russian Federation for Civil Defense, Emergencies, and Elimination of Consequences of Natural Disasters, 35 pp. 5 . National Crisis Management Center.

OCR for page 32
 PREDICTIVE MODELING PACKAGES FOR EFFECTIVE EMERGENCY MANAGEMENT 6 . Shatrov, V. F., and A. Yu. Silantyev. 2003. Situational centers: Information support for high-level management decisions. Pp. 8-17 in Systems Problems of Quality, Mathematical Model- ing, and Information and Electronic Technologies—Part II: Imitative Modeling and Conflictology. Materials from an International Conference and the Russian Scientific School. Moscow: Radio and Communications. 7. Filippovich, A. Yu. 2003. Integration of Situational, Imitative, and Expert Modeling. Mos- cow: Radio and Communications, 310 pp. 8. Filippovich. Integration. 9 . Romanov, V. V., and D. D. Shulga. 2003. Conceptual description of conflicting interactions. Strategic Stability 2:16-21. 10 . Kopylov, N. P., A. M. Ryzhov, and I. R. Khasanov. 2000. Major fires and their modeling. Pp. 170-187 in Modeling fires and explosions, N. N. Brushlinsky and A. Ya. Korochenko, eds. Moscow: Pozhnauka [Fire Science]. Kopylov, N. P., and I. R. Khasanov. 2001. Predicting the fire situation at sites under demo- lition. Pp. 101-102 in Extreme Situations: Prevention and Elimination. Collected Materials from a Scientific-Practical Conference. Minsk: Belarus State University. 11. Ryzhov, A. M, I. R. Khasanov, A. V. Karpov, et al. 2003. Application of a Field Method for Mathematical Modeling of Fires in Structures: Methodological Recommendations. Moscow: VNIIPO, 35 pp. 12 . Astapenko, V. M., Yu. A. Koshmarov, I. S. Molchadsky, and A. N. Shevlyakov. 1988. Ther- modynamics of Structure Fires. Moscow: Stroiizdat [Construction Publishers], 448 pp; Molchadsky, I. S. 2005. Fire in a Structure. Moscow: VNIIPO, 456 pp. 13. See, for example, Cooper, L. Y., J. A. Rockett, H. E. Mitler, and D. W. Stroup. 1989. A pro- gram for the development of a benchmark compartment fire model computer code. Fire Technology 25(4):116-127. Takeda, H. 1988. Transient model of early stages in compartment fires. Pp. 21-34 in Math- ematical Modeling of Fires, J. R. Mehaffey, ed. Philadelphia: ASTM; Merkushkina, T. G., and V. V. Romanov. 1981. Use of mathematical modeling in studying fire hazard factors. Pp. 34-43 in Safety of People in Fires. Moscow: VNIIPO. 14. Ryzhov et al. Mathematical Modeling of Fires. Ryzhov, A. M. 2000. Field models of fires. Pp. 25-88 in Modeling Fires and Explosions, N. N. Brushlinsky and A. Ya. Korolchenko, eds. Moscow: Pozhnauka. Yang, K. T., J. R. Lloyd, A. M. Kanury, and K. Satoh. 1984. Modeling of turbulent buoyant flows in aircraft cabins. Combustion Science and Technology 39:107-118. Raycraft, J., M. D. Kelleher, H. Q. Yang, and K. T. Yang. 1990. Fire spread in a three-di- mensional pressure vessel with radiation exchange and wall heat losses. Mathematical and Computer Modeling 14:795-800. Cox, G. 1995. Combustion Fundamentals of Fire. London: Academic Press, 476 pp. Welch, S., and P. Rubini. 1996. SOFIE—Simulations of Fires in Enclosures: User Guide. Bedford: Cranfield University, 127 pp. 15. Ryzhov et al. Mathematical Modeling of Fires.