Skip to main content

Full text of "Long-range plan for a research project on carbon monoxide production and prediction"

See other formats

A 1 1 1 □ 3 154L70 



NISTIR 89-4185 


PREDICTION William M. Pitts 

National Institute of Standards 
and Technology 

National Engineering Laboratory 
Center for Fire Research 
Gaithersburg, MD 20899 


- U 5 6 

Robert A. Mosbacher, Secretary 


Raymond G. Rammer, Acting Director 


Research Information Center 
Gaithersburg, MD 20899 

NISTIR 89-4185 

flti >TC 




National Institute of Standards 
and Technology 

National Engineering Laboratory 
Center for Fire Research 
Gaithersburg, MD 20899 

May 1989 

Issued October 1989 

Robert A. Mosbacher, Secretary 


Raymond G. Kammer, Acting Director 



This document presents a five-year plan for the Center for Fire Research 
(CFR) Priority Research Project on Carbon Monoxide Production and Prediction. 
Sections of the report provide a justification for the priority project, 
assess the current state of knowledge, summarize current relevant CFR research 
efforts, discuss specific research needs, list major assumptions utilized in 
formulating the research program, outline a research plan designed to meet the 
goals of the project and address the specific research needs, provide a rough 
timetable and budget, and present a discussion of the project philosophy and 

management . 




This document presents a five-year plan for the Center for Fire Research 
(CFR) Priority Research Project on Carbon Monoxide Production and Prediction. 
The overall goal of this project is to improve the understanding of and 
predictive capability for the formation of carbon monoxide (CO) in fires. The 
principal emphasis is on fires located within compartments since the ultimate 
motivation for the project is the large number of fire fatalities which result 
from the generation and transport of high concentrations of this toxic 
species . 

In preparation for formulating this plan the author has performed an 
extensive review of the existing literature and relevant intramural and 
extramural CFR research programs . An important component of this review is 
the findings of a Workshop on Developing a Predictive Capability for CO 
Formation in Fires which was held in Clearwater, Florida on December 3-4, 

1988. Leading experts in the field have provided recommendations and 
justifications for the principal areas in which research is required in order 
to achieve the ultimate project goal. An executive summary of this workshop 
is available [1]. 

In the following sections a justification for the priority project is 
provided, the current state of knowledge is briefly assessed, current relevant 
CFR research efforts are summarized, specific research needs are discussed, 
major assumptions utilized in formulating the research program are listed, a 
research plan designed to meet the goals of the project and address the 
specific research needs is outlined along with a rough timetable and budget, 
and a discussion of the project philosophy and management is presented. 



One of the mandates of the Center for Fire Research is a reduction in the 
number of fire deaths and injuries. To achieve this goal, CFR has led the 
effort to characterize and model fire behavior. The development of realistic 
fire models allows strategies to be developed for ameliorating its effects. 

Investigations have shown that a large percentage of fire deaths and 
injuries can be attributed to products of combustion. Even in cases where 
burn injuries and death occur, incapacitation of victims by fire gases often 
plays a pivotal role since escape from flames is hindered or rendered 
impossible. Careful studies (e.g., [2-3]) have shown that more than one half 
of all fire victims have fatal levels of carboxyhemoglobin in their 
bloodstreams . 

Even if a complete understanding of the physiological effects of CO (in 
combination with effects due to other fire products, e.g., reduced oxygen and 
increased respiration resulting from elevated carbon dioxide concentrations) 
is available, accurate predictions of fire toxicity require reliable estimates 
for the concentrations of toxic gases generated by a fire. 


Systematic investigations of CO formation in fires date from at least the 
1960s. However, despite nearly three decades of research the current 
understanding of CO formation in fires must be characterized as poor. This is 
true despite the fact that it has long been recognized that CO produced in 
fires is the principal gaseous species responsible for fire deaths. 


Numerous large-scale enclosure fire tests have been reported in which CO 
concentrations have been measured (e.g., [4-5]). These tests show 

unequivocally that CO concentrations high enough to be life-threatening often 
occur in this type of fire. On the other hand, no systematic investigations 
of CO formation in full-scale tests have been identified in the literature, 
and the conditions necessary for the generation of high levels of CO are not 
well characterized. The uncertainties are compounded by variations in 
experiment protocols (e.g., probe placement) and the utilization of inadequate 
experimental instrumentation. 

In general, high CO concentrations are associated with either smoldering 
or underventilated fires. Smoldering fires will not be considered here since 
other Center projects deal with this topic. Some crude correlations of CO 
formation and the ventilation parameter (Ah* ) have been attempted [6], 

However, the fire behaviors in these studies have not been adequately 
assessed, and the utility of these correlations for actual fires is limited. 
For locations removed from the room of origin, CO concentrations are believed 
to depend critically on whether or not additional combustion of product gases 
occurs after exiting the enclosure. The understanding of this process is very 
poorly characterized, and the controlling parameters have not been identified. 

George Mulholland has developed a zero-order model for CO formation in 
room fires [7] which assumes very low concentrations of CO for preflashover 
conditions and significant CO concentrations for postf lashover fires. CO 
concentrations following f lashover are estimated to be 0.5 times the carbon 
dioxide concentration. This model is based on measurements recorded in three 
series of full-scale fire tests performed at CFR. 


In recent years some carefully performed experiments have been carried 
out which offer the promise of an improved understanding of CO formation in 
terms of engineering correlations. Workers at Harvard [8] and Cal Tech [9] 
have generated controlled two -layer combustion systems which closely mimic the 
simple two-layer model often used to represent enclosure fires. Note that, in 
general, these experiments are steady state while actual fires are highly 
dynamic. A remarkable finding of these experiments has been that the major 
products of combustion (including CO) for a given fuel can be correlated in 
terms of the global equivalence ratio in the upper layer containing the 
combustion products. The correlations do depend on fuel type. Recent work 
[9] has shown that the same correlation holds even for the cases where 
additional air is injected into the upper layer. It should be noted that 
C0/C0 2 ratios measured in these experiments do not achieve the level of 0.5 
assumed by Mulholland based on results in full-scale studies [7]. 

Very recently, the Cal Tech workers have extended their studies to an 
experimental configuration where the entire flame is located within the upper 
layer [10]. For this case, burning occurs entirely within a vitiated 
atmosphere. The rather surprising observation (based on earlier speculation 
in the literature) is made that very low concentrations of CO are measured for 
increasing fuel equivalence ratios up to the point where oxygen concentration 
in the vitiated environment (= 13%) is insufficient to support combustion. 
Similar observations have been made here at CFR for preliminary measurements 
in a cone calorimeter modified to allow partial vitiation of the air supply. 
These studies indicate that vitiation of the air supply is not responsible for 
high levels of CO as long as sufficient oxygen is available for complete, 
fully- involved combustion. 


In the past few years computer models have become available which allow 
the solution of the complex kinetic equations for chemical reactions and 
molecular transport in simple combustion systems [11]. Schemes involving 
reduced-kinetic mechanisms to lower the computational overhead are also under 
development [12]. Even though complete treatment is not currently possible 
(e.g., soot formation cannot yet be handled), insights concerning the behavior 
of simple products of incomplete combustion is possible. The same techniques 
should be especially useful for analyzing whether or not upper layer chemistry 
is important in compartment fires for a given equivalence ratio, residence 
time and temperature. 

Significant efforts are underway worldwide to develop models which are 
capable of predicting the behavior of turbulent combusting flows. Fires fall 
into this classification. The highly complex natures of both combustion 
(complex chemistry and molecular transport) and turbulence (wide range of 
spatial scales required to characterize the phenomena) have proven to be very 
difficult obstacles to the successful development of practically-useful 
models. Several promising avenues continue to be investigated. 


At the present time there are no extramural grants dedicated principally 
to the CO Production and Prediction priority project and the only intramural 
research funded directly by the project are the experimental turbulent mixing 
project and support for an analysis (with regard to CO formation) of 
experimental data from the Sharon, PA fire simulation and cone calorimeter 
experiments utilizing vitiated air. As evidenced by this report, current 
funding is primarily directed to providing a detailed project plan. 


Despite the fact that the total resources specifically dedicated to the 
project are very small, the actual Center effort related to CO Production and 
Prediction is quite substantial and has achieved significant progress during 
the past few years. There are two principal reasons for this apparent 

• The first is that CO formation is a problem which is central to fire 
research. Research efforts which are nominally dealing with a 
different aspect of fire have components which directly impact the CO 
Prediction priority project. 

• The second is the fact that by merely designating CO Production and 
Prediction as a priority research area and appointing a staff member 
to act as director and research coordinator has resulted in enhanced 
visibility for the project and, particularly for many of CFR's 
grantees, significant redirection of some projects toward 
consideration of problems relevant for this research topic. 

Many current in-house programs of the Center have direct impact on the CO 
Production and Prediction priority project and, in turn, are influenced by the 
existence of the project. Several examples are cited here. 

• Mr. K. D. Steckler's project on wall burning in a reduced-scale 
enclosure included sufficient measurements of gaseous species and 
flow rates to allow a crude test of the global equivalence ratio 
concept for the correlation of CO in a somewhat more realistic fire 
environment than earlier laboratory experiments. This analysis is in 
progress . 

• The simulation of the Sharon, PA townhouse fire organized by Dr. R. 

S. Levine and Mr. H. E. Nelson was designed especially to study the 
release and transport of toxic gases. This experiment provides new 
insights into the formation of CO in "real" fires. 

• The effort by Mr. R. D. Peacock to develop a data base of large-scale 
fire tests should provide new information and means for assessing the 
concentrations of CO generated by "real" fires. 

• One of the Center's cone calorimeters has been modified to allow 
burning in a simulated vitiated environment. Measurements in this 
system are being utilized to assess the importance of vitiation on CO 

• Very simple models of CO formation are being incorporated into FAST 
and Hazard I for predictive purposes. Longer-term efforts on 
chemically reacting turbulent flows are expected to provide the means 


for introducing appropriate "source terms" into fire models. The 
ability to predict CO formation will be one of their primary 
functions. Theoretical efforts by Dr. H. R. Baum and the 
experimental efforts of Dr. W. M. Pitts feed into this effort. 

Much of the Center's current experimental effort on CO Production and 
Prediction is taking place in extramural research efforts sponsored by other 
groups, or which are in support of other priority research areas. Examples 
are cited below. 

• Professor E. E. Zukoski's studies are continuing with the oversight 
of Leonard Cooper, leader of the CCFM priority project. As already 
discussed, the findings of this study along with a companion study at 
Harvard which ended several years ago are providing significant 
insights into CO formation in fires. 

• Professor R. J. Roby (in collaboration with Dr. C. L. Beyler) is 
developing a reduced- scale experimental system for investigating 
fires in enclosures for a project which is being monitored by Dr. H. 
E. Mitler. These experiments should provide additional insights into 
the appropriateness of utilizing the global equivalence ratio concept 
for correlating CO formation in fires and generate a new 
understanding as to the role of combustion outside of the enclosure 
in determining the final concentrations of CO delivered by a fire. 

• Professor R. J. Santoro is currently working on a project associated 
with the Soot Formation and Evolution priority project. As a part of 
this effort he has begun measurements designed to investigate whether 
there is a correlation between the amount of soot released from 
laminar flames and CO production. This is a fundamental area, where 
little previous work is available, which might be very important for 
understanding the mechanisms of CO formation. 

• Professor J. H. Miller of George Washington University is working on 
a fundamental study of combustion chemistry designed to improve the 
understanding of soot formation in fires. He has recently considered 
the implications of his measurements in laminar diffusion flames for 
CO formation in fires. His analysis has helped clarify why global 
kinetic expressions are not appropriate for predicting CO 
concentrations . 

• Professor G. M. Faeth of the University of Michigan, stimulated by 
his participation in the recent workshop on CO formation, plans to 
make careful measurements of CO concentrations above his buoyancy- 
driven turbulent flames. Prior to this time, CO has been treated as 
a minor product and careful measurements were not made. 

• The CFR-supported program at Factory Mutual Research Corporation has 
components which impact on the CO Formation priority project. These 


workers are considering incorporating a project aimed specifically at 
this problem. 


5.1. Workshop Findings 

The findings of the recent workshop on "Developing a Predictive 
Capability for CO Formation in Fires" provide an excellent starting point for 
a discussion of the research , required to meet the goals of the Carbon Monoxide 
Production and Prediction priority project. For convenience sake, the final 
conclusions and recommendations for the workshop [1] are reproduced here. The 
priority assigned to each area is given roughly by the order of presentation 
(the highest is first). The workshop executive summary should be consulted 
for additional background material and discussion. 

1. Experiments clearly suggest that high concentrations of CO result 
principally from burning in underventilated conditions. The degree of 
vitiation appears to be important only when oxygen concentrations are 
too low to support combustion. Experiments and theoretical efforts are 
required to understand the formation mechanisms of and to allow 
prediction of high CO concentrations. Such studies should be assigned 
the highest priority. Investigations of laminar flames and buoyancy - 
driven turbulent flames are both necessary. 

2. Existing experimental diagnostic techniques do not provide adequate 
capabilities to answer the questions which must be addressed when CO 
formation in fires is considered. The lack of suitable diagnostics 
represents a serious hinderance to the priority project and must be 
addressed. Inadequacies are apparent in all experiments from the 
smallest laminar flame to full-scale turbulent fires. Principal 
problems are the absence of suitable sampling techniques and the 
presence of soot in fires which acts as an interferant and, at the same 
time, must also be characterized. Existing diagnostic methods must be 
improved and more carefully applied and new methods capable of time- and 
space - resolved concentration and/or temperature measurement in soot- 
laden flows developed. Particularly important needs are for techniques 
which allow local and global equivalence ratios to be determined 
accurately . 

3. The observation that CO concentrations can be correlated in terms of the 
global fuel equivalence ratio provides the most logical starting point 


for an engineering correlation of CO production within enclosures. 
Theoretical efforts are necessary to allow the correlation in terms of 
the global equivalence ratio to be understood in terms of the local 
combustion environment within the combusting plume. Laminar flamelet 
theory provides the most logical starting point for these efforts. 
Experiments are necessary to characterize the conditions for which the 
global fuel equivalence ratio concept is valid and to verify whether or 
not it can be extended to developing full-scale fires. 

4. The understanding of the combustion behavior of upper layer gases 
emitted through a vent in an enclosure containing a fire is very poor. 
Experiments must be formulated which determine the parameters which 
control this process and correlations and/or theoretical models must be 
developed which allow the degree of CO burnout to be estimated as a 
function of appropriate fire conditions. 

5. The role of "upper-layer chemistry" must be clarified. Conditions 
(chemical and thermal) where the upper layer continues to react and for 
which it is passive (with regard to additional CO formation or removal) 
need to be identified. Full kinetic calculations and reduced mechanisms 
offer particular promise for this problem. The role of inhomogeneous 
chemistry must be experimentally investigated and incorporated into 
kinetic models. 

6. Scaling effects must be investigated. It is recognized that full-scale 
testing is necessarily limited and that experiments on smaller scales 
must be devised. This is only possible if scaling effects are 

5,2. Discussion of Research Requirements 

The members of the workshop have done an outstanding job of identifying 
the principal requirements of a coherent research program designed to address 
the need for the development of CO predictive capabilities. The workshop was 
composed of members who reflected both fundamental and engineering approaches 
to the problem and the recommended research programs are a balanced 
representation of both research strategies. Note that the actual 
recommendations are not based on this criterion. 

There are a number of less specific requirements which must be addressed 
during the proposed priority project. For instance, up to now there has been 
no systematic analysis (beyond the necessarily brief review by Mulholland [7] 


which only considered three sets of data) of existing full-scale fire test 
data to ascertain what can be learned concerning the formation of CO. Such an 
analysis is not straight-forward. Large-scale fire tests are generally not 
designed to investigate the generation and movement of toxic gases. Gas 
sampling positions in different tests are highly variable and the accuracy of 
the reported measurements is often suspect and uncharacterized. The 
development of a large-scale fire test data base should prove very useful. It 
is clear that the CO Production and Prediction project should provide input as 
future large-scale tests are being planned in an attempt to improve the impact 
of these tests on the understanding of CO formation. 

The six areas listed in the last section provide a general outline of the 
research which is necessary to meet the goals of the priority project. In 
order to be successful the five-year plan must effectively incorporate efforts 
in each of these areas. In addition, an effective management strategy must be 
employed to ensure that effective communication and feedback exists between 
the various (and seemingly disparate) parts of the project and that potential 
users of the results (i.e., fire modelers) are kept apprised of progress and 
provide feedback concerning their requirements. 

In order to outline a program dealing with a such a complex research 
question as CO formation and prediction, which involves a large number of 
researchers and disciplines, it is necessary to make certain assumptions 
concerning the resources and time available. Other assumptions must be made 
also. These are discussed in the Section 6. 

In Sections 7-9 a program is described which effectively addresses each 
of the six research areas suggested by the workshop members. A table of more 
specific components is also provided. In some cases these components are 


currently part of the Center's projects. In other cases, new research efforts 
are required and examples of researchers and/or organizations having suitable 
expertise are identified for informational purposes. Cost estimates are 
included in Section 8. A description of the management structure to be 
employed appears in Section 10. 

6. Project Assumptions 

Several assumptions have been utilized in the development of the research 
plan which is presented in Sections 7-9. The most important are discussed 
here . 

6.1. Research Approach 

In order to have an effective and broad project on CO Prediction it is 
necessary to pursue both fundamental and engineering approaches to the 
problem. Even though no clear demarcation exists between the two approaches, 
it is generally possible to characterize various research efforts as being one 
or the other. 

The current state of knowledge suggests that the engineering approach is 
more developed and offers the most viable near-term means for the development 
of correlations which can be incorporated into existing CFR models of fire 
behavior. Fundamental research is required to provide an understanding of the 
physical and chemical processes underlying the engineering correlations and to 
allow these findings to be extrapolated to cases where experiments have not 
been performed. A more fundamental understanding will become crucial as 
efforts are made to incorporate realistic chemistry and physics into the 
"source terms" of fire models. 


6.2. Duration of Project 

It is assumed that the priority project will last for a five-year period. 
As in the majority of research projects, this estimate of the period required 
to complete the project is based on the expectation that no major "surprises" 
occur during the project and that the assumptions made concerning the current 
state - of -knowledge will be shown to be correct. 

6.3 Project Goals 

Clearly the general goal of improving the understanding of and predictive 
capability for the formation of CO in fires is far too general to serve as a 
legitimate measure of project success. More specific goals (milestones) must 
be formulated. The following milestones in terms of the engineering and 
fundamental components of the project represent realistic goals for the five- 
year program. 

1. By the end of the third year of the project an engineering 
correlation for CO concentrations in terms of the global 
equivalence ratio will be available for incorporation into fire 
computer models. Further work during the final two years of the 
project will be required to fine tune this correlation and 
determine the appropriate conditions for which its use is valid. 

By the end of the five-year period it will be possible to 
incorporate the effects of the combustion behavior of products 
exiting an enclosure on CO formation. 

2. By the end of the five-year period the fundamental portion of the 
project will have identified the important parameters controlling CO 
formation in fires. A general understanding of the reasons for the 
success of the global equivalence ratio concept in terms of the 
combustion behavior within the plume will have been attained. 
Modelling of CO formation for very simple systems (e.g., a single gas 
burner located in an enclosure) will be possible. Treatment of more 
complex configurations and fuels will not be possible in the absence 
of significant advances in other Center priority projects. 


6,4, Availability of Project Resources 

The availability of resources is a key component in the successful 
conduct of the priority project. In order to prepare a realistic project 
plan, it is necessary to assume that certain resources will be provided. As 
described in Section 4 and enumerated in Section 8, significant Center 
resources are currently directed at the project despite the fact that most of 
these efforts are not currently targeted specifically on the CO formation 
problem. As the research project becomes more highly focused and developed, 
it is assumed that a rough doubling of the current funding commitment will be 
needed. These additional funds will be used to augment existing research 
programs and initiate new efforts which are not currently part of the Center 

6.5, Assumptions Concerning Current Understanding 

In this section some of the assumptions which are being made concerning 
appropriate research approaches are discussed. This topic is included here 
since the success of the project in attaining its goals will be strongly tied 
to the appropriateness of these assumptions. 

A major assumption of this work is that the global equivalence ratio 
concept offers a means for correlating CO production in real fires. Some 
confidence is gained by the success of this approach in well-controlled 
laboratory experiments. However, it must be remembered that no scaling 
studies have been performed and that upper -layer temperatures in the 
laboratory experiments are not generally as high as attained in actual fires. 
The global equivalence ratio concept could fail in either of these situations. 
If this occurred, a new approach for an engineering correlation would need to 


be adopted and the time table for the development of an useful correlation 
would certainly slip. 

A more subtle problem is implied by the inclusion of area 2 in the 
recommended research areas listed in Section 5.1. It is assumed that it will 
be possible to verify the global equivalence ratio concept for large-scale 
fire tests. Such a verification requires accurate gas species measurement 
within the upper layers of large-scale fires. However, it is unclear if the 
necessary species measurements can be performed in such an environment 
utilizing existing experimental diagnostics. The uncertainty involved here is 
the major reason members of the workshop recommended that diagnostic 
development be assigned such a high priority in a project on CO formation [1]. 
For the purposes of planning it must be assumed that experimental measurements 
which are accurate enough to verify or disprove the global equivalence ratio 
concept for full-scale fires will be available within the projected project 
time frame. 

The research effort on the fundamental side requires assumptions which 
are even less certain than for the engineering side. This is usually the case 
for research projects aimed at both fundamental understanding and engineering 
solutions. The assumption that suitable diagnostics will be available is more 
crucial. In order to improve the fundamental understanding of CO formation, 
time- and space -resolved measurements within combustion regions of laminar and 
turbulent buoyancy- driven flames will be required. Here the lack of suitable 
diagnostics could prove debilitating to the project. 

Models have been developed for treating combusting, buoyancy- driven 
turbulent flows. The most widely employed is the laminar flamelet model [13] 
incorporated into suitable turbulence models. This approach has been used for 


a large number of combustion systems. It must be remembered that Professor 
Bilger of the University of Sydney has questioned the basis for the laminar 
flamelet concept and has suggested that reactions may actually be distributed 
in broad regions of the plume [14]. The most widely utilized turbulence model 
for predicting turbulent behavior is the k-e model. No attempts have been 
made to use the k-e model and laminar flamelet concept for predicting the high 
levels of CO generated by fires in enclosures. It will be assumed that the 
laminar flamelet concept in conjunction with the k-e model for turbulent 
mixing provides a viable approach for modeling CO formation. This hypothesis 
must be carefully tested. If it turns out to be incorrect, the development of 
models for treating CO formation in simple configurations could be delayed 
significantly . 


A coherent research plan is crucial if both the fundamental and 
engineering approaches to the CO Production and Prediction priority project 
are to make substantial progress. In some cases, necessary research is in 
progress or is planned, and it is only necessary to proceed along an existing 
path. In other cases, new programs must be nurtured either here in the Center 
or by the extramural research program. 

The areas of research listed in Section 5.1 provide a good starting point 
for the discussion. Additional research areas will also be mentioned. For 
those cases where existing Center programs are in place, the relevant programs 
will be mentioned. In some cases existing research efforts have importance 
for two or more of the research areas. Other subjects are not currently part 


of the Center program. Examples of researchers currently working in related 
areas are provided. 

7.1. Required Research Components 

7.1.1 Fundamental and Engineering Investigations of Underventilated Flames 

Previous research findings strongly suggest that high CO concentrations 
occur principally for underventilated burning. Most past combustion research 
has not considered this problem. A few laminar flame studies are reported in 
the literature and the Center has ongoing efforts on enclosure fires which are 
relevant (in particular, the grant projects of Professors Zukoski and Roby and 
the wall burning study of Mr. K. D. Steckler) . 

The work at Cal Tech has made very good progress in the past and 
continues to do so. Professor Zukoski completely understands the importance 
of his work to the CO Prediction priority project and has been responsive to 
its needs. One limitation of the work is that it has not considered fuel 
effects. Investigations of fuel effects are proposed for the next three-year 
cycle . 

Professor Roby plans to investigate burning in a reduced- scale enclosure 


for a variety of fuels. The measurements generated by this grant program will 
be directly applicable to the CO Production and Prediction priority project 
and should complement the work of Professor Zukoski. This program should be 
closely monitored to ensure compatibility with both the CCFM and CO Prediction 
priority projects. 

Currently there is no Center program on underventilated burning of simple 
laminar flames even though some preliminary experiments have been reported in 
the literature. Given the high importance assigned to this research area, a 


new effort should be initiated in this area with a general goal of identifying 
the fundamental chemistry and physics responsible for the formation of high 
levels of CO during underventilated combustion. A large number of fuels 
should be utilized in order to characterize the importance of fuel effects on 
CO formation. This study will not only provide an improved understanding of 
any engineering correlations which are developed, but will also be absolutely 
essential as more fundamental models for fire source terms are developed. 

Many researchers have the necessary background and technical expertise to 
initiate a program on underventilated burning. The necessary expertise is 
located here in the Smoke Dynamics Research Group and at a number of 
universities. Examples of research groups having such capabilities include 
workers who are currently supported by the CFR extramural program such as 
Professors Faeth, Santoro, or Miller, and others such as researchers at 
Princeton, Stanford, or the University of California campuses of Berkeley, San 
Diego, Irvine, or Los Angeles. 

7.1.2. Development of Flame Diagnostics 

At first glance it might appear as if the development of accurate 
measurement techniques for flame gases is outside the scope of this priority 
project, but the members of the workshop rated the development of new 
diagnostics for chemical species as being critical to success. Such 
development is viewed as being particularly relevant when measurements in 
flame regions with high soot concentrations must be made. Current methods for 
species measurements are subject to a variety of uncertainties, and no 
systematic investigations of the effects of flames on probe sampling of gases 
of interest are available. This is true despite the fact that some members of 


the research community have argued that many past measurements of CO in flame 
regions have had large systematic errors. 

The accurate measurement of flame gases in regions of high temperature is 
important to both engineering and fundamental approaches. One of the basic 
questions which must be answered in developing correlations for CO formation 
is the effect of additional combustion outside of the room of origin on the 
final concentrations of CO delivered by a fire to remote locations. This 
question can only be answered satisfactorily if accurate measurements of flame 
gas concentrations are available both within and removed from the room of fire 
origin. Measurements must be made for both large-scale and reduced- scale fire 
tests. Since the upper layer temperatures in actual fires can reach 
temperatures of approximately 1300 K, accurate gas sampling for temperatures 
at least this high must be possible. 

In attempting to develop more fundamental approaches for understanding CO 
formation the requirements for gas sampling become even more severe. One of 
the principal goals of the priority project is to understand the success of 
the global equivalence ratio concept in terms of the behavior of the turbulent 
plume. Such an understanding requires that gas concentrations be available 
with high spatial and temporal resolution. It is safe to say that at the 
present time no suitable technique for such measurements has been demonstrated 
for sooting turbulent flames. 

The lack of suitable diagnostics may turn out to be the largest roadblock 
on the pathway to an improved understanding of CO formation. For this reason, 
it is recommended that a component of the priority project be dedicated to the 
development and characterization of suitable diagnostics. The author has 
identified many researchers who have an interest in this problem, but none who 


have a strong record of demonstrated ability in this area. Within the Center 
Drs . G. W. Mulholland and T. Kashiwagi have expressed an interest in 
developing new diagnostics, and Dr. K. C. Smyth and Professor J. H. Miller 
have developed specialized techniques for very weakly sooting flames. Workers 
at Factory Mutual Research Corporation seem to have taken the most care in 
developing probes which are properly "quenched" , but careful testing of their 
probes has not been performed. Professor R. J. Santoro of Penn State has 
apparently made some progress in the development of a new probe which is not 
susceptible to soot clogging. This research should be monitored very 
carefully. Professor W. Grosshandler of Washington State University 
(currently at NSF) has also indicated that probes having the desired 
characteristics (at least for time-averaged measurements) can be developed 

[ 15 ]. 

It is believed by the author that this area is one of the most crucial 
and, at the same time, uncertain research topics of the priority project. It 
is clear that some progress must be made if significant advances on the 
priority project are to be possible, but it is also clear that there are not 
sufficient resources available to have a wide-ranging program. It is 
suggested that current projects which have diagnostic development components 
(such as the projects directed by Dr. T. Kashiwagi, Dr. K. C. Smyth, and 
Professor R. J. Santoro, and the Factory Mutual Research Corporation work) be 
monitored very carefully and that these investigators be encouraged to work 
along these lines. One project devoted entirely to measurements of flame 
gases should be supported by CFR. This research project may turn out to 
intramural or extramural . 


7.1.3. Use of The Global Equivalence Ratio Concept 

Research findings suggest that the correlation of CO concentrations in 
terms of the global equivalence ratio in the upper layer provides the most 
viable and best- developed approach for incorporating CO formation into 
existing Center fire models. Before this can done with confidence several 
experiments must be performed. Some of these experiments have been deemed 
important enough to be included as separate research areas. Thus research 
areas 4-6 in Section 5.1 provide information which is required before the 
concept can be applied with confidence to full-scale fires. 

Other projects which impact on this area are discussed here. In area 3 
of Section 5.1 it is pointed out that theoretical efforts are required to 
allow the success of the global equivalence ratio in correlating CO formation 
to be understood in terms of the behavior of the turbulent plume. The most 
logical starting point for such a study is to utilize the laminar flamelet 
model with the k-e turbulence model and attempt to perform a calculation which 
mimics the experimental arrangement of Professor Zukoski and his coworkers. 

As a first approximation the behavior of the combusting buoyant plume 
below the upper layer can be treated as a fully-ventilated, free combusting 
plume. The calculation can be used to solve the problem for the flow until a 
downstream position is reached where the interface between the upper and lower 
layers is located. This is a standard calculation which has been previously 
tested. The results of this calculation can then serve as the starting 
conditions for a reacting plume flowing into surroundings with the same 
composition of the upper layer. For a first attempt, the upper layer will be 
assumed to be infinite and recirculation effects and quenching ignored. The 
correctness of these assumptions can be tested by comparison of the calculated 


results with the experimental results when a hood is present. This extended 
calculation should provide an indication of whether or not such a modeling 
strategy offers a useful approach for understanding CO formation in fires. 

In the above discussion it has been implicitly assumed that gas species 
and temperature will have the same dependence on the local equivalence ratio 
within the flame as found for fully-ventilated fires. This can not be 
confirmed without experimental tests. The experiments described above for 
laminar burning with underventilation should allow this hypothesis to be 
checked. If the hypothesis fails, the observed dependence of gas species on 
the local equivalence ratio can be incorporated into the model. 

There are a number of laboratories which are capable of performing the 
computational study outlined here. Professors J. P. Gore and G. M. Faeth 
(current CFR grantees) come to mind immediately. At the present time, the 
Center is not one of these. However, during the next fiscal year the Center 
is planning to procure a sophisticated computer program package named 
"Phoenics" from CHAM of North America. This package is designed to allow k-e 
modeling of combusting flows. The problem outlined here would provide an 
excellent opportunity to exercise this code for a problem of great interest to 
the Center. Assuming the availability of the code, this part of the program 
should likely be performed within the Center. 

It is clear that more fundamental models of chemically reacting turbulent 
flows will ultimately be required to model accurately the formation of CO as 
well as many additional "source terms" for fire models. The Center priority 
project on Turbulent Combustion is performing long-range research in this 
area. The experimental program on turbulent mixing is currently being 
supported by the CO Prediction priority project. This project will eventually 


contribute to modeling of fire "source terms" as well as an understanding of 
CO formation. Like many of the Center's programs, the turbulent mixing 
project is interdisciplinary, and it has potential impact on a number of the 
priority projects. It should be maintained as part of the long-term efforts 
on modeling of plume behavior, chemically reacting turbulent flow, and CO 

7.1.4. Effect of Combustion of Upper-Layer Gases Outside of the Enclosure 

The gases and soot particles which form the upper layer of enclosure 
fires are often found to be highly vitiated and fuel rich. The fuel 
stoichiometry is generally rich enough to support combustion when these gases 
are mixed with air containing sufficient 0 2 . When these gases exit the 
enclosure through vents, additional combustion is frequently observed at 
locations where these gases contact higher concentrations of oxygen. 

Large number of uncertainties exist concerning this process. The 
conditions required for such burning are not characterized, and the degree to 
which such burning can remove the partially-pyrolized fuel has not been 
determined. It is generally believed that combustion of this typfe results in 
less CO being delivered at positions away from the fire, but one can imagine 
conditions of partial combustion where higher CO levels would be produced. 
Clearly, the behavior of such burning is one of the major areas of uncertainty 
concerning the formation of CO in fires. 

Even though this type of behavior has significant implications for many 
fire research applications (e.g., fire safety and fire spread), there has been 
no systematic investigation of the problem. Currently, Professor Roby is 
setting up an experiment in which he plans to make observations concerning 


such burning. Hopefully, this project will provide reliable data and allow 
some preliminary engineering correlations (insights?) to be developed. 

With regard to full-scale fire testing within the Center, if careful 
measurements of gas concentrations (particularly CO) can be made at the exit 
and downstream of the burning region outside of the enclosure, it should be 
possible to characterize the effectiveness of such burning in removing CO. 
Ultimately, a more fundamental approach will be required to deal with this 
aspect of fire behavior. However, in the absence of observational data and 
engineering correlations, this problem seems beyond the scope of the current 
priority project. 

On the basis of the above discussion it is recommended that Professor 
Roby be encouraged to pursue his current studies on the burning of partially- 
pyrolized gases exiting his reduced- scale enclosure and that this work be 
monitored very carefully by Center personnel. If sufficient progress occurs, 
a project focused on this particular fire area should be initiated either here 
or in an extramurually- funded program if a suitable proposal is received for 

In the meantime, consideration should be given to "piggybacking" on 
Center full-scale fire tests to make measurements of CO both within the upper 
layer of the fire enclosure and downstream of the position where burning has 
occurred. Such measurements will certainly provide information concerning the 
effectiveness of this type of burning in removing (or creating) CO. 

7.1.5. Stability of Upper Layers at High Temperatures 

The studies at Cal Tech have demonstrated that the concentrations of CO 
and other flame gases are stable for temperatures as high as 900 K [16]. This 


is encouraging since it suggests that it should be possible to ignore the 
effects of upper layer chemistry for temperatures below this level. However, 
temperatures in the upper layers of real fires have been shown to attain 
levels as high as 1300 K. For these conditions it is conceivable that gas 
phase reactions occurring on a tens-of-seconds time scale could modify the CO 
concentration and invalidate the global equivalence ratio concept. It is also 
possible that chemical reactions can occur on the surfaces of soot particles 
at lower temperatures . 

In order to develop a fundamental understanding of CO formation in fires 
and identify the ranges of conditions over which the global equivalence ratio 
concept is valid, it is necessary to investigate both homogeneous and 
inhomogeneous chemistry in the upper layer. 

Two different approaches should be adopted. Computer codes are available 
(such as those based on the Sandia Combustion Research Facility ChemKin 
package [11]) which allow quite complicated sets of chemical kinetic equations 
to be handled. Clearly, it will be impossible to treat all of the gas-phase 
chemical species in the upper layer using a full -kinetic treatment. However, 
it is expected that only species in relatively high concentrations will be 
important. This should allow the number of species involved and the total 
number of chemical reactions to be limited to realistic values. 

A useful approximation will be to treat the upper layer as a well -stirred 
reactor at a given temperature. Initial concentrations of gas phase species 
(representative of the fire environment) can be input and the time history of 
the concentrations monitored. If the concentrations are calculated to be 
invariant for the time scales of interest (usually on the order of tens of 
seconds), the upper layer can be considered as being chemically passive. 


However, if strongly varying concentrations of CO are predicted, the upper 
layer must be considered as chemically active and the global equivalence ratio 
concept would be expected to fail in this case. 

The use of full chemical -kinetic codes has proliferated during the past 
few years and the expertise for their use can be found at a large number of 
institutions. Professors P. R. Westmoreland at the University of 
Massachusetts and M. D. Smooke at Yale are two well-known workers in this 
area. There have also been recent attempts to mount these codes on CFR 
computers, and some progress has been made. However, much work remains to be 
done within the Center before meaningful, research-grade calculations can be 
attempted. The CO Prediction priority project offers an opportunity to 
incorporate this important capability into in-house Center projects. 

Heterogeneous chemistry is much more difficult to model than homogeneous 
gas -phase chemistry. An experimental approach should be adopted for this part 
of the upper- layer stability problem. Furnace experiments can be devised in 
which soot from actual combustion process (or a suitable surrogate) are placed 
in a heated flow tube. Appropriate mixtures of CO, C0 2 , H 2 , 0 2 , H 2 0, and 
other flame gases are allowed to flow through the heated tube containing the 
small solid particles and the chemical composition of the mixture at the end 
of the flow tube is monitored by an appropriate analytical technique such as 
gas chromatography or mass spectrometry. 

As a first step, these experiments should be performed at temperatures 
low enough to ensure the absence of significant gas -phase reactions. If no 
important heterogeneous reactions are identified, it will be necessary to go 
to higher temperature where competition between homogeneous and inhomogeneous 
reactions is possible. In order to identify the role of inhomogeneous 


reactions it will be necessary to make measurements in the presence and 
absence of the solid phase particle. Full or partial modeling of the 
appropriate gas phase kinetic equations might also be useful here. 

There have been some measurements of the role of solid particles and 
catalytic surfaces on the conversion of CO to C0 2 (and vice-versa). However, 
there have been no studies concerned with the chemical stability of upper 
layers in fires. This is a new direction for fire research. The expertise 
for such a project exists within the Center, but a redirection of research 
effort and the development of a new apparatus would be required. Professor W. 
Grosshandler of Washington State University (currently a rotator at NSF) has 
expressed an interest in this problem and proposed to have a student work at 
the Center in conjunction with him (gratis, one day a week) on a simple flow 
reactor during the upcoming summer. This proposed summer project offers a 
viable means at a relatively low cost to either initiate a longer-term grant 
project in this area or to develop the necessary expertise within the Center. 

7.1.6. Scaling Studies 

This is the last of the major research areas listed by the workshop 
members. It reflects the fact that only limited numbers of full-scale fire 
tests will be possible for characterizing CO formation. It is almost certain 
that it will be necessary to utilize measurements in smaller-scale fires to 
provide the number of experiments required to cover the range of possible 
conditions necessary for useful engineering correlations and to characterize 
experimental variability. 

In order to insure that the results of smaller-scale investigations are 
meaningful for full-scale fires, it will be necessary to investigate the 


scaling relationships between the smaller-scale and appropriately chosen 
large-scale experiments. 

The most likely scenario for such a test is to compare the findings from 
the smaller-scale experiments which are currently being set up at VPISU with 
one or more (certainly a very few) full-scale experiments performed within the 
Center. It is possible that the full-scale test can be "piggy-backed" on a 
planned large-scale fire test, but it seems more likely that a specially- 
planned test(s) will be required. In any case, attempts should be made to 
compare the effectiveness of the global equivalence ratio concept in 
predicting CO concentrations in the upper layers on each scale as well as the 
effectiveness of burning outside of the enclosure in removing CO. Before such 
a scaling test can be run, the proper scaling parameters must be determined 
for fire properties such as fuel loading, ventilation parameter, and fire heat 
release rate. Fortunately, there are a number of researchers in the Center, 
including Dr. J. G. Quintiere, who are experts in this research area. With 
these scaling laws it will be possible to construct a large-scale fire test 
specifically designed to test the scaling properties relevant to CO formation. 


The discussion thus far has identified the major research areas where 
work is required in order to improve the understanding of CO in fires and 
develop methodologies for its prediction. In this section, recommended 
specific components of the project are listed in tabular form (Table 1). Each 
is identified as being of an engineering, fundamental or dual nature. A short 
statement of purpose (goals) for each component is provided. Work currently 
supported by the Center (intramural or extramural) is listed. Either current 


researchers working on the problem or examples of researchers having the 
required expertise to work on the component are compiled. An estimate of the 
time required to achieve the goals of the component is provided along with an 
indication of timing for work on the component relative to the five-year 
period of the CO Production and Prediction priority project. 

In a sense, this table provides a general summary of the five-year 
research plan for meeting the goals of the CO Production and Prediction 
priority project. It provides an indication of the complexity involved and 
the wide range of questions which must be effectively addressed before a 
useful understanding and a predictive capability for CO formation is 
available. Note that each of the major areas of research recommended by the 
workshop participants is addressed by one or more of the research components 
included in the table. 


Table 1 lists the research components which have been identified as being 
required to make significant advances in the understanding and prediction of 
CO formation in fires. In this section estimates are provided for the annual 
funding levels required to support each of these components . 

As has already been noted, significant research efforts of relevance to 
this project are being supported by CFR even though the funding is not 
allocated directly to the CO Formation Priority Project. Current expenditures 
for each of the eighteen project components have been estimated. 

Table 2 lists each research component, whether CFR or grantees (assumes 
the submission and funding of appropriate unsolicited proposals) are projected 
to perform the work, an estimate of FY89 CFR expenditures on the component 


© o u 
& >. $ 
° £ t 

© 3 


g ° -s 

3 © 
a o > 




o © 
a n 

03 a 

pi s 

•g Sr 

© .-3 

H r— I 

f3 iH 

o £ 

X E— i 

5 03 
o i 


~H *H 
M +3 
© G 

S ? 

© M 

n © 

J: c 

O 3 

o m p 
o © > 


n £ X 

<M O *J 

43 •— I 

o 3 

P -P 

i £ 
g J 

4-3 3 © ® ""3 

W 43 > © © 

© *o © ® n 

n © X *-< o G 

•rl *J *J © O *H 

43 ~H 
O 43 

© G 
M © 

5 £ 
© © 
O G 
H 3 


43 -H M 

W 3 O 

© 6 1 43 

43 © © © 

33 O 

T3 O G W 
C *-3 O © 
O 00 O 00 

c I 

M 43 n 

43 JC g 43 

•g d * g 

63 <D _j ^ 
43 M 43 

a © © o 
o n ^ o 
»-3 3 3 w 











**H 43 O O 
43 0 43 G © 

<3 W © M 

© 3 © © 

00 O © -H 



© 43 © 43 



£ * 
© u 
*o © 


O -4 


3 43 
43 © 

w ^3 

43 © 43 

© «-3 

n 3 *o 

o Ja © 

*Q *3 43 

© 3 © 

J *3 n 

I £5 

Q © "O 

^3 43 43 

© a 

> © c 

•*3 O *H 

& o o 

© o L> 

W -H 3 
43 W 

O © O 

U M 3 

3 a 
43 00 G 
O -n © 

TJ a C 

2 I 5 


I 0) O 
O G -*3 
© O 43 
JZ -H U) 
43 4 3 3 

T3 5) ■§ 

G -H O 
0 43 a 


-3 © 43 

© > G 
43 G © 
G -H «-3 
© 3 

E *-3 »Q 

-3 © 13 

n o 3 

© -3 43 

£ 1 

C > 

G **3 


G n 

-h o 

X 43 













































m in 

i i 

co «— i 

2 © 

H T3 
O C 

CO © 

m X 
© in 
44 in 

O o 













^ T 3 
o C 
o © 

Q 5 

&. O 

£ 5 

















































O 44 

« o *o 

© 4J X 

m © 

© -H 4-5 

X E w 

4-5 © m © 

o C in o So 

© © © Qj © 

14 00 *-4 E »-4 

© O O O 

jC 14 -4 O 14 

U © 40 © 

O 4 (0 ft 

O © © © C4 

H £ D- 00 D 

40 U O X 

•»4 © 40 

T 3 a © 

C a-D T3 

O 3 -h C 



co O 

4 O © © -H 

.c © m o o © 

O M © r -4 © 14 

© m 
X « 
5 oo 

a © 

C4 T3 T3 

3 -h C 

© O 3 co O 

© —4 4 44 © 

40 c 3 C 

o M a © © 

£ 0 n o u 

O M © *-4 © 

© CO 

f S. 

5 0® 

© C — l 

© CJ T3 

40 14 >4 40 


3 C 40 -«4 in 

^4 -h 73 © 

© E © M 

> © l M **4 

W -4 M CU 44 

40 3 40 

© in in 44 
•H © © O 
14 © 40 

a E © 

O © -4 

U o >4 © 
UU -H > 
a. 44 © 

© 44 r4 T) 
O © © 

© r-4 in > 

73 00 © CO M 

3 C O © © 

■—4 -H (0 O 10 

f — 1 »4 --4 © © 



in © 

E •— * C *— 1 

g s 

o O 
44 U £ 

f-4 4 O © 4 O 14 © 

O © *H «-4 

> © © 

So 3 

X 73 

O C 

X tO 


© c 

Pa © 

3 op 

O r-4 4J O C © 

40 E 
co o 

© 14 

1-4 44 

O 40 14 CO 

o © o © 

44 14 44 00 




© 44 U 
00 O 40 

-r4 co in 

40 10 -j C 

co 40 E O 
© o © -^ 
> © X 40 
C 44 O © 

■s § 



E 44 00 CJ 

-4 0 0 

14 14 O 

© c © o 

p, O 40 

K -4 © C 

U3 40 £ O 

X U 

© © 44 


C © co 
M Pa X 
3 Pa 3 
03 3 O 

© CO 

CU 4-> 

a 3 
3 O 

4-> O © 

C O o 

g 14 1 ° 
a o h 

O 44 *— 4 

•—< 3 

© © 44 

> CO 

© © C 

Q £ -h 


Incorporation of CO Engineering Incorporate findings of CO New start CFR staff Years 

formation in fire priority project into CFR fire 

models models 

w G 

<0 63 

£ d 
*3 63 

63 ffi 
> > 
ffi ffi 

»— I ,c 


to a> 



c • 

o M 

T5 Q 

OT .C 
© +3 

•C -H 

•— J P 

-Q © 
o q & 
® k ® 


O © 

© TJ 

® 3 

P 3 
^ O 
M .C 

P 3 
P ^ P 
w P 3 
ffi © 2 

P P 
G 04 U 
® © O 

ffl © 

•O P 

O p 

*D ffi - 

. i is 

D W 

■§ * 
O 43 


U CT) 
ffi CO 
T3 >4 
G d 
3 „ 

O i*S 
ffi m 

M >> 

<0 « 
=C O 

© 43 


W O 

P d 
^ O 
M M 
O d 

O * 
^ P 
3 G 

S a 

M © ffi 
ffi XI C P 
*— I ^ O M 
P 00 o 

2 60 P 10 
d ffi © 

aa o o ffi 

© < p 
W ffi 60 ffi 
ffi C p G 
44 P T3 

£5 S.S 


ffi S4 

TJ © 
G C 
3 p 

3 • 

e © 

® •o 

G © 

5 3 

60 G 

O © *4 W 


00 G 44 U 

d P P 44 
O © d 
P > ffi G 

> 3 G 
ffi O' O O 
Q ffi o U 

c c 

O P 




E 1 6 

G P 


> <3 

P O 
P 3 
> P 

G >4 
p O 
^ 44 


10 Experimental Investiga- Unkown 10k 90k 90k 90k - - Professor Santoro’s work on soot CO interactions fits here, 

tion of the effects of This component will also require a completely new start, 

heterogeneous chemistry 
on CO concentrations 




H 4-3 
<D W 

(X 0) 


^ -4 
00 0 
00 M 

T3 -H 
G 40 

(0 in 

G <0 
© o 

O W 

© o 

*4 M 

O p, 


M 40 

O G 
a © 
G o 

-r4 0 

G — < 
— t 3 

40 <44 


0 0G 

4-3 U-. 


© 0 
o -C 
in 4-3 



O m 
0 G 
> 3 
0 44 

£ 43 © 

C/3 © © 
-C X 

in o 

1 i 

s ■? 

in o 

48 g 

O 0 
-H TJ 
M G 

04 3 

© M M *-4 

O 0 

G O 


£ O ® 

© in 

£< -to 

tX 3 
3 O 

X 5^ O 0 

© in -4 

a. -»-> o 
a 3 c 

3 0© 

40 4-3 0 

© © 14 

TJ E -h 
M 44 
44 O 
O 44 0 


40 o <o 

CU o 

© m 

E *4 i 
U O r-4 
O 44 -4 
*-4 3 in 

© © 44 40 

> in in 

© © C © 

Q XI -4 40 


(both intramural and extramural), and projected levels of funding for the five 
years of the priority research plan. Since these estimates are by necessity 
crude, no allowance for inflation has been incorporated (i.e., the numbers are 
in 1989 dollars) . The last column of Table 2 contains comments concerning 
some of the assumptions which have been utilized in making the cost estimates 
and projections. 

Table 3 provides a summary of the current spending and projected budgets 
for the priority research project. Note that the budget increases over the 
first three years to a level roughly double that of current expenditures and 
then falls over the remaining two years of the effort. 
















The discussion and tables included in Sections 7 and 8 provide a 
comprehensive research project plan for meeting the objectives of the CO 
Production and Prediction priority project. In this section the reasons 
behind some of the choices of research components and funding levels are 
considered . 

The two principal goals for the priority project are given in Section 6.3 
(page 12). Two of the components of the project (numbers 4 and 5) deal 


specifically with these goals. Note that both are intended to be carried out 
by CFR staff. 

The first goal is develop an engineering correlation for CO 
concentrations based on the global equivalence ratio concept within three 
years. Research component 4 provides substantial resources for the first 
three years to allow CFR researchers to utilize the current understanding of 
and ongoing research results on CO formation to develop the correlation 
necessary for inclusion into CFR fire models. 

A number of the other research components are designed to feed 
information into this effort. Most important are numbers 2, 9, 10, 14, 15 and 
16. Since results from these studies must be available before component 4 can 
achieve goal 1, the time frames for these supporting efforts are heavily 
weighted toward the first three years of the project (see Table 1). 

For planning purposes it has been assumed that a suitable engineering 
correlation will become available during the third year of the project. 
Substantial funding is included for the incorporation of the correlation 
(component number 17) into whichever fire model(s) is deemed appropriate. 

In developing project goal 1 (see page 12) it was recognized that even 
though it should be possible to formulate an engineering correlation for CO 
concentration in terms of the global equivalence ratio within the first three 
years of the project, certain aspects of the problem can not be completed 
within this time frame. The burning behavior of gases exiting the enclosure 
is sure to be one of the remaining uncertainties. Components 11 and 12 are to 
address this problem. These components are funded during years 2-4. It is 
anticipated that a sufficient understanding of the controlling parameters will 
be available by year 5 to allow crude estimates of the effects of such burning 


on the CO concentration correlation. Funding is provided for component 4 in 
years 4 and 5 to incorporate these findings into the correlation. Any fine 
tuning of the correlation which is necessary as the result of new experimental 
findings will also be done during this period. 

The second goal of the priority project deals with fundamental aspects of 
the problem. Success in this area has been defined as the identification of 
the important parameters controlling CO formation in fires, a general 
understanding of the success of the global equivalence ratio in correlating CO 
production in enclosures, and the development of a model for CO formation for 
a simple, idealized enclosure fire burning a gaseous fuel. As was true for 
the first goal discussed above, several of the research components deal with 
this aspect of the problem. 

Research components 1, 2, 6, 7, 8, 9, 10, and 13 are required to generate 
the research findings to meet the fundamental project goal. Components 1, 2, 
7, 8, 9, and 10 will aid in the identification of the important physical 
processes responsible for the formation of high levels of CO during 
ventilation- limited burning. Note that 2, 9, and 10 are also necessary to 
validate the engineering correlation developed to meet the first goal of the 
proj ect . 

In order to have confidence in the engineering correlation which is 
generated, it is necessary to have some understanding of the physical basis 
for the success of the correlation. The fundamental studies mentioned in the 
last paragraph will provide the insights required to develop this 
understanding. At the same time, component 13 deals specifically with the 
question of whether or not the high levels of CO formed during vitiated 
burning can be understood simply in terms of turbulent combustion processes. 


Turbulent combustion modeling is to be employed for this purpose. Note that 
the experimental results of component 1 are required as input for this 
research. For this reason, the timing of component 13 lags behind 1 by a 
year . 

The final goal of the fundamental studies is to develop a model capable 
of predicting the time development and concentrations of CO for a simple 
prototypical fire in an enclosure. Research component 5 deals specifically 
with this topic. Development of this model is slated for years 3-5 of the 
project. This component is near the end of the project period because the 
successful development of the model will require the results of other 
components . 

At the present time, it is not possible to state what form this model 
will assume. It may be found that the k-e modeling approach developed during 
component 13 can be extended to predict the time dependence of a developing 
fire in an enclosure. Alternatively, this approach may prove to be 
inappropriate and more realistic modeling of chemically reacting turbulent 
flows may be necessary. Research component 6 deals with this approach. In 
either case, it should be possible to make a choice of direction by the end of 
year three. Some initial analysis is likely to be required. Resources 
allocated for component 13 for years 4 and 5 are to be used for the actual 
development of the chosen model. 

The majority of research components included in Tables 1 and 2 generate 
research findings of direct relevance to the priority project. Components 3 
and 18 do not fall into this classification. Component 18 is designed to 
provide overall management for the wide-ranging priority project. The 


rational for including this as a specific component is described in the 
following section. 

Component 3 provides funding for the development of flame diagnostics. 
Table 2 shows that it should be funded at a level of 150k for the first two 
years of the priority project. This relatively high level of funding reflects 
the crucial need for these new diagnostics which has already been discussed. 
The members of the workshop placed a very high priority on this research area. 
The effort should be funded only for the first two years since developments in 
later years would be too late to be of value to the priority project. 

The recommended level of funding is unlikely to lead to major 
breakthroughs in flame diagnostic technology, such breakthroughs would require 
much higher annual funding over a multi-year period. The recommended funding 
is intended to provide the incentive for careful assessment and testing of 
existing technologies with the goal of providing the most accurate 
experimental measurements possible. Existing technologies which might be 
explored have been discussed in Section 7.1.2. 


The complexity associated with CO formation and prediction in fires 
precludes an effective project by one or even a few individuals. As mentioned 
in Section 3, the current understanding of CO formation is very poor despite 
the availability of many uncoordinated investigations. The amount of 
expertise required to make meaningful progress and the complexity of the 
problem requires that a large number of research problems be addressed and 
that investigators have a wide range of capabilities. 


The reality of extremely limited resources necessitates the evaluation of 
ways to achieve meaningful progress on the project without utilizing 
unrealistic amounts of new resources. Fortunately, a review of Center 
programs revealed (see Sections 4, 8, and 9) that a large number of research 
efforts are underway which have direct impact on the CO project. The talent 
represented by this research spans a number of areas and represents a cross- 
section of the best fire researchers in the country. On this basis, a 
research plan has been formulated which builds on these existing efforts. By 
utilizing these existing research components (with some redirection) along 
with selected new research efforts, it has been possible to develop a research 
plan which brings significant resources to bear on the problem of CO 
prediction within a reasonable budget. It seems likely that significant 
progress will be made if this research plan is implemented. 

One of the most difficult aspects of the proposed research effort will be 
the coordination and direction of the various components. A great deal of 
effort will be required on the part of the project manager to stay apprised of 
progress on the various efforts. Minor changes in program directions which 
are required to meet the goals of this priority project will require 
negotiation with those responsible for the various components. The importance 
of project management is emphasized by including it as one the components in 
the research plan (Component 18, Tables 1 and 2). 

The organization of the workshop provided an opportunity for the project 
manager to not only become better acquainted with many of the relevant efforts 
underway in the Center, but to also develop a better rapport with the 
researchers involved. This familiarity will prove invaluable as this very 
ambitious project develops. The review of the literature during the past year 


has formed the foundation for the project proposed here and at the same time 
has provided the tools necessary to monitor the wide range of research which 
has been proposed. 


[1] Pitts, W. M. Executive summary for the workshop on developing a 
predictive capability .for CO formation in fires. National Institute of 
Standards and Technology NISTIR 89-4094; 1989 May. 68 p. 

[2] Harland, W. A.; Anderson, R. A. Causes of death in fires. Proceedings, 
Smoke and Toxic Gases from Burning Plastics; 15/1-15/19; Jan. 6-7, 1982; 
London, England. 

[3] Harwood, B.; Hall, J. R. What kills in fires: smoke inhalation or burns? 
Fire J. 83: 29-34; 1989 May/June. 

[4] Wilson, W. J. Large scale fire tests. J. Fire Flam. 7: 112-124; 1976. 

[5] Levine, R. S.; Nelson, H. E. Full-scale simulation of Sharon, PA fatal 

fire: data and analysis. To appear as a National Institute of Standards 

and Technology Internal Report. 

[6] Rasbash, D. J.; Stark, G. W. V. The generation of carbon monoxide by 
fires in compartments. Fire Research Station F.R. Note No. 614; 1966 Feb. 
38 p. 

[7] Mulholland, G. W. Position paper regarding CO yield. Letter report to 
Richard G. Gann, Chief, Fire Measurement and Research Division, Center 
for Fire Research; 1988, June 16; 8 p. 

[8] Beyler, C. L. Major species production by diffusion flames in a two- layer 
compartment fire environment. Fire Safety J. 10: 47-56; 1986. 

[9] Zukoski, E. E.; Toner, S. J.; Morehart, J. H.; Kubota, T. Combustion 
processes in two-layered configurations. Fire Safety Science. 

Proceedings of the Second International Symposium. New York, NY: 
Hemisphere; 1989. 295-304. 

[10] Morehart, J. H. ; Zukoski, E. E. Chemical species produced by natural gas 
fires near the extinction limit. Part of Final Report for National 
Institute of Standards and Technology Grant No. 60NANB600638 ; 1988 
October . 


[11] Kee, R. J.; Miller, J. A.; Jefferson, T. H. CHEMKIN : a general- 
purpose, problem- independent , transportable, Fortran chemical kinetics 
code package. Sandia National Laboratories Report SAND80-8003; 1980 
March; 203 p. 

[12] Peters, N. ; Kee, R. J. The computation of stretched laminar methane-air 
diffusion flames using a reduced four-step mechanism. Combust. Flame 
68: 17-29; 1987. 

[13] Peters, N. Laminar flamelet concepts in turbulent combustion. Twenty- 
first Symposium (International) on Combustion. Pittsburgh, PA: The 
Combustion Institute; 1988. 1231-1250. 

[14] Bilger, R. W. The structure of turbulent nonpremixed flames. Twenty- 
second Symposium (International) on Combustion; Pittsburgh, PA: The 
Combustion Institute; 1989. 475-488. 

[15] Grosshandler , W. Personal communication. 

[16] Toner, S. J.; Zukoski, E. E.; Kubota, T. Entrainment, chemistry, and 
structure of fire plumes. National Bureau of Standards Government 
Contract Report NBS-GCR-87-528 ; 1989 April; 200 p. 


M14A IHEV. 2-tO) 



2. Performing Organ. Report No. 

3. Publication Date 



October 1989 

SHEET (See instructions) 



Long-Range Plan for a Research Project on Carbon 
Monoxide Production and Prediction 


William M. Pitts 

$. PERFORMING ORGANIZATION (If joint or other thon NBS. see instructions) 



7. Contract/Grant No. 

I. Type of Report & Period Covered 



| | Document describes a computer protram; SF-185, FIPS Software Summary, is attached. 

11. ABSTRACT (A 200-word or less factual summary of most significant information. If document includes a significant 
bibliography or literature survey, mention it here) 

This document presents a five-year plan for the Center for Fire Research (CFR) 
Priority Research Project on Carbon Monoxide Prediction. Sections of the report 
provide a justification for the priority project, assess the current state of 
knowledge, summarize current relevant CFR research efforts, discuss specific 
research needs, list major assumptions utilized in formulating the research 
program, outline a research plan designed to meet the goals of the project and 
address the specific research needs, provide a rough timetable and budget, and 
present a discussion of the project philosophy and management. 

12. KEY WORDS (Si* to twelve entries; alphabetical order; capitalize only proper names; and separate key words by semicolons) 

carbon monoxide, combustion, fire gases, fire hazard, fire prediction, toxic gases 

13. availability 

I y| Unlimited 

| | For Official Distribution. Do Not Release to NTIS 

| | Order From Superintendent of Documents, U.S. Government Printing Office. Washington, D.C. 


| X| Order From National Technical Information Service (NTIS), Springfield, VA. 22161 

14. NO. OF 



15. Price 


UiCOMWDC •04»-W»0