United States
                     Environmental Protection
                     Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
                      Research and Development
EPA/600/S9-87/005 Apr. 1987
&EPA           Project Summary
                      Pollutant  Emission  Factors for
                      Gas  Stoves:  A  Literature Survey

                      Cliff I. Davidson, John E. Borrazzo, and Chris T. Hendrickson
                        Published emission factors for CO,
                      NO, NO2, NOX have been summarized
                      for gas-fired kitchen stoves. Analysis of
                      variance was then used to investigate
                      the importance of three binary factors:
                      type  of combustion, burner position,
                      and method of sampling. The emission
                      factor data were then used to estimate
                      coefficients in various multivariate re-
                      gression models. The influence of gas
                      flow  rate on emission factors was in-
                      vestigated separately. The data were
                      also used to investigate the sensitivity
                      of predicted airborne concentrations to
                      uncertainties in emission factors. Final-
                      ly, the data were used to identify critical
                      gaps in understanding emission factors.
                        This Protect Summary was developed
                      by EPA's Air and Energy Engineering
                      Research Laboratory, Research Triangle
                      Park, NC, to announce  key findings of
                      the research project that Is fully docu-
                      mented In a separate report of the same
                      title (see  Project Report  ordering In-
                      formation at back).


                      The Survey
                        Published emission factors for CO, NO,
                      N02, and NOX for gas-fired kitchen stoves
                      have  been summarized. To the extent
                      that data exist, the emission factors have
                      been  used in  a  statistical analysis to
                      investigate the most important  factors
                      influencing emissions.
                        First, the data were used to construct
                      normal probability and log probability plots
                      based on the cumulative distribution of
                      the emission factors. Correlation coef-
                      ficients were computed for both types of
                      plots. Results showed that the log prob-
                      ability plot provided a better fit to the CO
                      and NO2 data with reasonably high cor-
                      relation coefficients, suggesting that the
                      emission factors for these pollutants may
be approximated by lognormal distribu-
tions. Either normal or lognormal distri-
butions were satisfactory for NO and
NOX. For consistency, the emission fac-
tors for all four pollutants were assumed
to be  distributed  lognormally in sub-
sequent analysis.
  Analysis of variance was then used to
investigate the importance of three binary
factors in explaining the observed vari-
ations in emissions: (1) type of combustion
(poorly adjusted or well  adjusted), (2)
burner position (front or  rear),  and (3)
method of sampling (direct or indirect).
The results showed that roughly half of
the observed variance in log EF (base 10
logarithm of the emission factor) for CO
can be explained by noting if the com-
bustion is poorly adjusted.  For  N02,
roughly 30% of the variance can be ex-
plained by this factor. For NO and NOX,
the fraction of variance explained by this
factor depends on the subset of the data
chosen: fractions ranged from 0.088 to
0.56. Burner position  and method of
sampling were both  relatively  unim-
portant in explaining the observed vari-
ance for any of the four pollutants.
  The  emission factor  data were then
used to estimate coefficients  in various
multivariate regression models. The first
regression model  incorporated several
factors: type of combustion, burner posi-
tion, method of sampling, the  three two-
way interactions between these factors,
and (M-1) binary factors corresponding to
the M stoves for which  data were  avail-
able (M=27, 26, 26, and 8 for CO, NO,
N02  and NOX, respectively). Subsequent
multivariate  regression  models  were
constructed by sequentially eliminating a
factor or factors from the previous model.
Results of these tests showed that stove
differences were significant at the 95%
level in explaining the  variance in CO,

-------
  N02, and NOX emission factor. Type of
  combustion was significant for CO, NO,
  and N02. Burner position had a smaller
  but still statistically significant effect in
  explaining the variance 'in CO and N02
  emissions. Similarly, the method of sam-
  pling had a small but statistically signifi-
  cant effect for NOX emissions.
    The influence of gas flow rate on emis-
  sion factors was investigated separately.
  Statistical tests were  not  run for this
  factor due to both a lack of data and the
  presence of detailed data from only one
  study. Results of plotting all of the data
  for each pollutant on a common  graph
  showed that CO and N02 emissions vary
  considerably with gas flow rate,  while
  emissions  of NO and  NOX  are less
  variable.
    The data were used to investigate the
  sensitivity of predicted airborne concen-
  trations to uncertainties in  emission
  factors.  The solution of a one-compart-
  ment  mass balance model was used as
  the predictor  of concentration. For  the
  existing distribution of emission  factor
  data,  the current uncertainty in  CO is
  responsible for a  larger fraction of  the
  variance in predicted concentration than
  uncertainty in air exchange rate. For NO
  and NO2, however, the uncertainty in air
  exchange rate is more important than the
  emission factor uncertainties.
    Finally, the  data were used to identify
  critical gaps in understanding emission
  factors and to suggest future experiments.
  Overall, it is concluded that the influence
  of stove design, gas flow rate, and char-
  acteristics of  stove use are key factors
  which merit further study.
         C.  Davidson,  J.  Borrazzo, and C.  Hendrickson are  with  Carnegie-Mellon
           University, Pittsburgh, PA 15213.
         Jane M. Crum is the EPA Project Officer (see below).
         The complete  report, entitled "Pollutant Emission Factors for Gap Stoves: A
           Literature Survey," (Order No. PB 87-171 328/AS; Cost:  $18.95, subject
           to change) will be available only from:
                 National Technical Information Service
                 5285  Port Royal Road
                 Springfield, VA 22161
                 Telephone: 703-487-4650
         The EPA Project Officer can be contacted at:
                 Air and Energy Engineering Research Laboratory Laboratory
                 U.S. Environmental Protection Agency
                 ResearchJriangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300

EPA/600/S9-87/005
             0000329   PS

             U  S  ENVIR  PROTECTION  AGENCY
             REGION  5  LIBRARY
             230  S DEARBORN  STREET
             CHICAGO                IL    60604

-------