EPA-460/3-76-024
October 1976
EXHAUST EMISSION
AND FUEL ECONOMY
CHARACTERIZATION
OF A MODERN
AUTOMOTIVE WANKEL
ROTARY ENGINE
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
Ann Arbor, Michigan 48105
-------
EPA-460/3-76-024
EXHAUST EMISSION
AND FUEL ECONOMY
CHARACTERIZATION OF A MODERN
AUTOMOTIVE WANKEL ROTARY ENGINE
D.E. Cole and D.J. Patterson
University of Michigan
Ann Arbor. Michigan
Grant No. R802536
EPA Project Officer: F. Peter Hutchins
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
Ann Arbor, Michigan 481 OS
October 1976
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from the
Library Services Office (MD-35) , Research Triangle Park, North Carolina
27711; or, for a fee, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by the
University of Michigan, Ann Arbor, Michigan, in fulfillment of Grant
No. R802536. The contents of this report are reproduced herein as received
from the University of Michigan. The opinions, findings, and conclusions
expressed are those of the author and not necessarily those of the Environ-
mental Protection Agency. Mention of company or product names is not
to be considered as an endorsement by the Environmental Protection Agency.
Publication No. EPA-460/3-76-024
11
-------
FOREWORD
The concern for the preservation of our environment and
dwindling petroleum resources has accelerated the search for new
powerplants by the automotive industry. Numerous powerplants,
including the gas turbine, Stirling, Diesel, and Wankel engines
are being investigated ever so carefully to determine their po-
tential to satisfy the many and complex criteria of modern
vehicles. The University of Michigan Automotive Laboratory has
been contributing to these on-going investigations for many years
with a focus on:
.measurement and control of undesirable exhaust emissions
.reduction of fuel consumption
.training engineering students to better contribute to the
solution of automotive problems.
During the past two years the Automotive Laboratory at the ,
University of Michigan has directed part of its effort to the
technical- assessment of a modern Wankel engine which is used in
the Japanese manufacturered Mazda vehicle. The following report
is a culmination of this study to characterize the exhaust emis-
sion and fuel consumption performance of the 1975 Mazda 13-B
rotary engine. The following objectives were addressed in this
experimental program'.
1. Map the fuel consumption and emission performance of
both a modern rotary engine and a conventional recip-
rocating engine of equivalent power.
2. Compare the fuel consumption and emission performance
of the selected test engines in a simulated identical
vehicle.
3. Determine the effects of the principal operating vari-
ables on both emissions and fuel consumption performance
of the rotary engine.
4. Identify features of key fundamental processes which
could cause differences in both emission and fuel con-
sumption performance with respect to the conventional
engine.
ill
-------
ABSTRACT
The objective of this experimental program was to study the
fuel consumption and emission characteristics of a modern automotive
Wankel rotary engine, compare these with a representative recipro-
cating engine and form conclusions regarding economy and emission
performance.
Fuel consumption and exhaust emission maps were developed for
a production 1975 Mazda 13-B rotary engine both with and without
thermal reactor. For comparison similar data were developed for a
3.8 liter 1975 Buick V-6 engine. Steady-state dynamometer tests
were run. In addition studies were conducted on the Mazda engine
to determine the influence of ignition timing and air-fuel ratio on
key performance factors. Engine indicator cards and spatial distri-
bution of hydrocarbons in the exhaust were also investigated.
The studies showed that the production Buick and Mazda engines
without external emission control and as installed in a hypotheti-
cal 1975 Mazda RX-4 vehicle exhibited significant differences in
fuel consumption and emission performance under steady-state opera-
tion. The fuel consumption of the Mazda was generally greater than
the Buick powered vehicle with a maximum difference of 20% at 35 mph.
Above 35 mph the difference diminished considerably and at 60 mph
the fuel consumption results were essentially the same. Likewise
the hydrocarbon emissions were found to be 6-10 times greater and
the CO emissions significantly greater. The NO emissions were
lower for the Mazda. By optimizing the mixturexratio and ignition
timing, the differences in fuel consumption, CO and NO emissions
were reduced considerably, however the hydrocarbon emissions were
still 3-5 times greater for the Mazda. Without exhaust treatment
the aldehyde emissions were found to be similar to reciprocating
gasoline engines.
Air-fuel ratio and ignition timing variations in the Mazda
influenced emissions and fuel consumption in a manner similar to
that observed in typical reciprocating engines although the effect
of ignition timing was less significant.
Using the spatial hydrocarbon distribution in the exhaust as
an indicator, it was determined that leakage of mixture past the
engine apex seals and/or incomplete flame propagation were probably
the most important factors causing the. high, level of hydrocarbon
emissions from the Mazda engine. Further evidence of seal leakage
and/or excessive heat loss was obtained from an analysis of firing
and motoring indicator diagram..
IV
-------
The Mazda engine was observed to be smooth., quiet and small
for its performance capability compared to the Buick V-6. Further,
the friction power was significantly lower than the Buick and in-
creased at a lower rate with speed increase.
This report was submitted in fulfillment of Grant No. R-
802536-01 by the University of Michigan, Walter E. Lay Automotive
Laboratory, under the sponsorship of the U.S. Environmental
Protection Agency. This report covers the period from October 1,
1974 to September 30, 1976. The experimental work was completed
as of June 1, 1976.
-------
CONTENTS
Foreword
Abstract
Figures
Tables .......... « .............................. ' ............ x .
XI 1 1
Acknowledgments • • ............ .................................
1 . Introduction .............. ......,,.,,,.,,., ............... 1
A. Background ................. ....,,.., ..... ............. 1
B. Objectives ................... . ......... , .............. 3
C. Test Program .......................................... 3
2 . Observations and Conclusions .............................. 5
3 . Test Equipment .... ........ . ............................... 9
A. Test Engines ............ ...,.,,.,,..,.,. ..... ,.....,.. 9
B . Dynamometer Installation . ........ ... .................. 12
C . Instrumentation ............ . ........... , .......... .... 12
4 . Test Procedure ............................ , . . . .......... . . 24
A. Exhaust Emission and Fuel Economy Mapping ............. 24
B. Mazda Parameter Studies ........ ....... , ............... 26
C. Aldehyde Emissions ............................. ....... 26
D. Hydrocarbon Source Analysis- - Timed and Manifold Sampl-
ing System ........ . ........... . . . ..................... 27
E. Mazda Engine Chamber Pressure Measurement ............. 27
5. Results and Discussion ................ .................... 28
A. Fuel Consumption and Emission Performance of the Mazda
13-B Engine ........................................... 28
B. Fuel Consumption and Emission Performance of the Buick
231 cu.in. V-6, Engine, no external emission control ... 51
C. Fuel Consumption and Exhaust Emission Comparison Be-
tween Mazda 13-B and Buick V-6 Engines ................ 60
D. Aldehyde Emissions from the Uncontrolled and Emission
Controlled Mazda Model 13-B Engine .................... 71
E. Influence of Air-Fuel Ratio on Specific Fuel Consump-
tion and Emissions - Mazda 13-B Engine ................ 73
F. Effect of Ignition Timing on Fuel Consumption and
Exhaust Emissions - Mazda 13-B Engine ................. 86
G. Spatial Distribution Studies of the Hydrocarbon and
Carbon Monoxide Emissions from the Mazda Engine ....... 96
H. Chamber Pressure Study ~ Mazda Engine ..... ,.,.., ...... 117
References ..,..,, ....... ............ t ..,.,..,, .......... . ..... 125
vi
-------
FIGURES
Number
1 Insulated thermal reactor used on the Mazda 13^8 engine
as viewed from the engine exhaust port contact surface .. 11
2 Schematic diagram showing operating modes of the various
air injection control valves and spark plug firing of the
Mazda 13-B emission control system 11
3 Rear quartering view of the Mazda 13-B engine installa-
tion 13
4 Side view of the Mazda 13-B engine installation 13
5 Front quartering view of the Buick V-6 engine installa-
tion 14
6 Merriam laminar flow meter and inlet surge chamber 17
7 Water cooled sample probe 17
8 Manifold sampling system used in the qualitative study
of hydrocarbon distribution in the exhaust gas 19
9 Cox timed sampling valve as installed in the exhaust
pipe of the Mazda 13-B engine 19
10 Schematic diagram of the Cox timed sampling valve 19
11 Schematic diagram of the pressure transducer as installed
in the Mazda engine with the rotor in two different posi-
tions 23
12 Wide-open throttle performance of the Mazda 13-B engine
as a function of RPM 29
13 Brake specific fuel consumption as a function of brake
horsepower, Mazda 13-B engine, with thermal reactor and
air injection 30
14 Brake specific hydrocarbon emissions as a function of
brake horsepower, Mazda 13-B engine, with thermal reactor
and air injection 32
15 Brake specific carbon monoxide emissions as a function of
brake horsepower, Mazda 13-B engine, with thermal reactor
and air injection .. 34
16 Brake specific nitrogen oxide emissions as a function of
brake horsepower, Mazda 13-B engine, with thermal reactor
and air injection 35
vn
-------
Number
17 Brake specific hydrocarbon emissions as a function of
brake horsepower, Mazda 13-3 engine, with thermal
reactor but without air injection ,,,,......,......, . . 37
18 Exhaust oxygen concentration as a function of brake
horsepower, Mazda 13-B engine, with thermal reactor
but without air injection 38
19 Brake specific carbon monoxide emissions -as a function
of brake horsepower, Mazda 13-B engine, with thermal
reactor but without air injection 40
20 Air-fuel ratio as a function of brake horsepower at
selected test conditions, Mazda 13-B engine, with thermal
reactor, but without air injection 41
21 Brake specific nitrogen oxide emissions as a function of
brake horsepower, Mazda 13-B engine, with thermal reactor
tbut without air injection 42
22 Brake specific fuel consumption as a function of brake
horsepower, Mazda 13-B engine, without thermal reactor
and air injection 43
23 Brake specific hydrocarbon emissions as a function of
brake horsepower, Mazda 13-B engine, without thermal
reactor and air injection 44
24 Brake specific carbon monoxide emissions as a function
of brake horsepower, Mazda 13-B engine, without thermal
reactor and air injection . . . 46
25 Brake specific nitrogen oxides emissions as a function of
brake horsepower, Mazda 13-B engine, without thermal
reactor and air injection 47
26 Fuel economy as a function of vehicle speed of the Mazda
13-B engine installed in a simulated Mazda RX-4 vehicle,
road load, 2.56 and 3.90:1 axle ratios 49
27 Mass hydrocarbon emissions as a function of vehicle speed
of the Mazda 13-B engine installed in a simulated Mazda
RX-4 vehicle, road load, 3.90 and 2.56:1 axle ratios ... 50
28 Mass carbon monoxide emissions as a function of vehicle
speed for the Mazda 13-B engine installed in a simulated
Mazda RX-4 vehicle, road, 3.90 and 2.56:1 axle ratios .. 52
29 Mass nitrogen oxide emissions as a function of vehicle
speed for the Mazda 13-B engine installed in a simulated
Mazda RX-4 vehicle, road load, 3.90 and 2.56:1 axle
ratios 53
30 Wide-open throttle, performance of the 1975 Buick V-6
, engine as a function of RPM ................ 54
31 Brake, specific fuel consumption as a function of brake
horsepower, 1975 Buick. V-6 with EGR and no catalyst...,. ' 56
viii
-------
Number
32 Brake specific hydrocarbon emissions as a function of
brake horsepower, 1975 Buick V-6 with. EGR and no cata-
lyst . ., . 57
33 Brake specific carbon monoxide, emissions as a function
of brake horsepower, 1975 Buick V-6 with EGR and no
catalyst . , 58
34 Brake specific nitrogen oxide emissions as a function of
brake horsepower, 1975 Buick. V-6 with EGR and no cata-
lyst 59
35 Air-fuel ratio as a function of brake horsepower for the
Buick V-6 engine 61
36 Mass emission rate of hydrocarbons, carbon monoxide, and
nitrogen oxides as a function of vehicle speed, Buick V-6
engine installed in a simulated Mazda RX-4 vehicle, road
load, 2.56 and 3.90:1 axle ratios 62
37 Road load fuel economy as a function of vehicle speed -
a comparison between the Mazda 13-B and Buick V-6 engines
installed in a simulated Mazda RX-4 vehicle 63
38 Road load hydrocarbon emissions as a function of vehicle
speed - a comparison between the Mazda 13-B and Buick V-6
engines installed in a simulated Mazda RX-4 vehicle .... 65
39 Road load carbon monoxide emissions as a function of
vehicle speed - a comparison between the Mazda 13-B and
Buick V-6 engines installed in a simulated Mazda RX-4
vehicle 66
40 Road load nitrogen oxide emissions as a function of
vehicle speed - a comparison between the Mazda 13-B and
Buick V-6 engines installed in a simulated Mazda RX-4
vehicle 67
41 Brake specific fuel consumption comparison between the
Mazda 13-B and Buick V-6 engines as a function of brake
horsepower at selected operating conditions 68
42 Motocing friction mean effective pressure and friction
horsepower as a function of RPM for the Maz'da 13-B and
Buick V-6 engines 70
43 Exhaust gas temperature as a function of brake horsepower
at selected conditions, Mazda 13-B and Buick V-6 engines. 72
44 Aldehyde emissions as a function of air-fuel ratio, the
Mazda 13-B engine at 2000 rpm, 8.7 brake horsepower 74
45 Aldehyde emissions as a function of brake horsepower at
selected conditions, Mazda 13-B engine with and without
thermal reactor -. 75
46 Brake specific fuel consumption as a function of equiva-
lence ratio, Mazda 13-B engine, MET spark ,........, 77
ix
-------
Number Page
47 Brake specific hydrocarbon emissions as a function of
equivalence ratio, Mazda 13-B engine., MET spark ........ 78
48 Brake specific carbon monoxide emissions as a function
of equivalence ratio, Mazda 13-B engine, MET spark .... 79
49 Brake specific nitrogen oxide emissions as a function of
equivalence ratio, Mazda 13-B engine, MET spark " 80
50 Brake specific fuel consumption as a function of equiva-
lence ratio, Mazda 13-B engine, stock, spark timing .... 82
51 Brake specific hydrocarbon emissions as a function of
equivalence ratio, Mazda 13-B engine, stock spark timing. 83
52 Brake specific carbon monoxide emissions as a function of
equivalence ratio, Mazda 13-B engine, stock spark timing. 84
53 Brake specific nitrogen oxide emissions as a function of
equivalence ratio, Mazda 13-B engine, stock spark timing. 85
54 Brake specific fuel consumption as a function of spark
advance, Mazda 13-B engine, 16:1 air-fuel ratio 87
55 Brake specific hydrocarbon emissions as a function of
spark advance, Mazda 13-B engine, 16:1 air-fuel ratio ... 88
56 Brake specific carbon monoxide emissions as a function of
spark advance, Mazda 13-B engine, 16:1 air-fuel ratio ... 89
57 Brake specific nitrogen oxide emissions as a function of
spark advance, Mazda 13-B engine, 16:1 air-fuel ratio ... 90
58 Brake specific fuel consumption as a function of spark
advance, Mazda 13-B engine, stock air-fuel ratio 92
59 Brake specific hydrocarbon emissions as a function of
spark advance, Mazda 13-B engine, stock air-fuel ratio .. 93
60 Brake specific carbon monoxide emissions as a function
of spark advance, Mazda 13-B engine, stock air-fuel ratio. 94
61 Brake specific nitrogen oxide emissions as a function of
spark advance, Mazda 13-B engine, stock air-fuel ratio .. 95
62 Schematic diagram of rotary engine operating cycle 97
63 Exhaust hydrocarbon concentration as a function of crank
angle relative to the start of the exhaust process, Mazda
13-B engine, 2000 rpm, 8.8 bhp. Data obtained with a Cox
timed sampling valve 100
64 Carbon monoxide concentration as a function of crank angle
relative to the start of the exhaust process, Mazda 13-B
engine, 2000 rpm, 8.8 bhp.. Data obtained with, a Cox
timed sampling valve 101
x
-------
Number
65 Exhaust hydrocarbon concentration as a function of crank
angle relative, to the start of the exhaust process,
Mazda 13-B engine, 2000 rpm, 14.2 bhp. Data obtained
with a Cox timed sampling valve , 102
66 Carbon monoxide concentration as a function of crank
angle relative to the start of the exhaust process, Mazda
13-B engine, 2000 rpm, 14.2 bhp. Data obtained with a
Cox timed sampling valve , , 103
67 Exhaust hydrocarbon concentration as a function of crank
angle relative to the start of the exhaust process, Mazda
13-B engine, 2500 rpm, 13 bhp. Data obtained with a
Cox timed sampling valve 104
68 Carbon monoxide concentration as a function of crank angle
relative to the start of the exhaust process, Mazda 13-B
engine, 2500 rpm, 13 bhp. Data obtained with a Cox
timed sampling valve 105
69 Exhaust hydrocarbon concentration as a function of crank
angle relative to the start of the exhaust process, Mazda
13-B engine, 2500 rpm, 22.4 bhp. Data obtained with a Cox
timed sampling valve 106
70 Carbon monoxide concentration as a function of crank angle
relative to the start of the exhaust process, Mazda 13-B
engine, 2500 rpm, 22.4 bhp. Data obtained with a Cox timed
sampling valve 107
71 Exhaust hydrocarbon concentration as a function of crank
angle relative to the start of the exhaust process, Mazda
13-B engine, 3000 rpm, 18.7 bhp. Data obtained with a
Cox timed sampling valve 108
72 Carbon monoxide concentration as a function of crank angle
relative to the start of the exhaust process, Mazda 13-B
engine, 3000 rpm, 18.7 bhp. Data obtained with a Cox
timed sampling valve 109
73 Exhaust hydrocarbon concentration as a function of crank
angle relative to the start of the exhaust process, Mazda
13-B engine, 3000 rpm, 34.3 bhp. Data obtained with a
Cox timed sampling valve 110
74 Carbon monoxide concentration as a function of crank angle
relative to the start of the exhaust process, Mazda 13-B
engine, 3000 rpm, 34.4 bhp. Data obtained with a Cox
timed sampling valve Ill
75 Hydrocarbon concentration as a function of exhaust sample
probe position in the exhaust pipe., Mazda 13-B engine,
2000 rpm, manifold sampling system .. 114
76 Hydrocarbon concentration as a function of exhaust sample
probe position in the exhaust pipe., Mazda 13-B engine,
2500 rpm, manifold sampling system 115
xi
-------
Number
77 Exhaust hydrocarbon concentration as a function of
sample, probe, location in the exhaust system, Mazda
13-B engine, manifold sampling system. .................. 116
78 Cylinder pressure as a function of crank angle, Mazda
13-B, at 2500 rpm, 22.4 bhp , , 118
79 Cylinder pressure as a function of volume, Mazda 13-B,
at 2500 rpm, 22.4 bhp 119
80 Cylinder pressure as a function of volume, logarithmic
plot, Mazda 13-B, at 2500 rpm, 22.4 bhp 120
81 Logarithmic plots of pressure as a function of volume,
fired Mazda 13-B engine , . ? , 122
82 Logarithmic plots of pressure as a function of volume,
motored Mazda 13-B engine 124
TABLES
Number Page
1 Test engine specifications 10
2 Gas analysis techniques 20
3 1975 Mazda RX-4 vehicle specifications and road load
data 25
4 Aldehyde emissions from a 350 cid (5735 cc) automotive
engine at 1200 rpm, 30 bhp 73
xii
-------
ACKNOWLEDGMENT S
The authors would like to gratefully acknowledge the outstand-
ing efforts of the following men who served as research assistants
or in other ways assisted: Garry Zawacki, Joseph Weissmiller,
Frank Jacobson, John White, Daniel Mehall, Anthony Brakora, Thomas
Salva and Eugene Zimmerman. The excellent support of the laboratory
technicians at the Walter E. Lay Automotive Laboratory at the
University of Michigan is also gratefully acknowledged.
In addition we are indebted to the Toyo Kogyo Company of Japan
and the Buick Motor Division of General Motors Corporation for their
important contributions of test engines, technical information and
other support. In particular, we would like to acknowledge the
assistance provided by Mr. lida from Toyo Kogyo who contributed his
time and efforts to the initial development of the test facility.
Kill
-------
SECTION 1
INTRODUCTION
A. BACKGROUND
During the past ten years we have witnessed ever-increasing
interest in alternative powerplants for passenger automobiles.
Initially a major incentive for the consideration of these dif-
ferent power systems was the prospect of reducing undesirable
exhaust emissions. At the present time, however, we are concerned
with the additional and perhaps equally important consideration
of thermal efficiency. In all probability, future engines will be
judged on the basis of fuel economy and emission performance con-
sidered together in addition to other important factors such as
cost, driveability, manufacturability, etc.
Numerous powerplant concepts are being investigated, includ-
ing both external and internal combustion engines. Of the many
being considered, only one has achieved significant production
status in an automobile - the Wankel, rotary combustion engine.
The Toyo Kogyo Company of Japan is currently featuring this rotary
engine in its Madza vehicles as is the German company NSU in the
RO-80.
The rotary engine possesses numerous attractive features.
Perhaps the most important advantage is its small size which
promises to offer considerable packaging flexibility in a vehicle.
This could enhance the auto industry's ability to design a vehicle
with compact exterior dimensions while maintaining the maximum
interior volume for people and luggage. The Wankel engine is sub-
stantially lighter and simpler (at least with respect to the number
of parts). In addition it breathes well because of a greater
length (crank degrees) of induction period and because of a rela-
tively unrestricted flow path for the incoming mixture and exhaust-
ing products. In the one and two rotor versions, the induction
system is simple which reduces problems associated with maldistri-
bution of the air-fuel mixture. Also, the speed range of the
engine is broad due to the better breathing characteristics,
potentially lower mechanical friction and no limitations imposed
by valve train dynamics. Further the engine appears to have a
lower fuel octane requirement at a given compression ratio. An
important advantage in comparison to engines..with 4 or fewer cylind-
ers is it's inherent dynamic balance.
As is the case with all systems, there are areas of concern
as well. The gas-sealing network in the engine requires a complex
-------
grid which has four right angle intersections per working chamber.
Effective gas sealing is very difficult and considerable develop-
ment effort has been expended to resolve this problem and several
groups have reported success. Lubricant must be added or injected
into the working chambers with the fuel and air mixture because
the apex seals slide on the rotor .housing which is not lubricated
otherwise. Exhaust temperatures appear to be higher than those of
a conventional engine at similar operating conditions. These high-
er temperatures do aggravate the materials problem in the exhaust
system but have a positive effect on exhaust clean-up of unburned
hydrocarbons and carbon monoxide. Likewise exhaust blowdown pres-
sures are high and exhaust silencing is difficult. Fuel economy
is perhaps the most controversial aspect of the rotary engine.
A wide range of fuel consumption data has been reported.
General Motors and Curtiss-Wright (1,2) have indicated that the
fuel economy performance of their prototype engines is essential-
ly equivalent to that of reciprocating engines. Conversely the
Mazda engine has demonstrated somewhat inferior results compared
to the reciprocating engine based on EPA data (3). In part this
arises from comparing the Mazda vehicles with, small 4-cylinder
sub-compact cars with much inferior performance. Fuel economy
improvements based on EPA tests in the 1975 and 1976 model years
have been reasonably dramatic although, the results are still
inferior to the typical sub-compact car.
Another area of considerable uncertainty is cost. While the
engine is far simpler in concept than a conventional engine, it
apparently requires more sophisticated components in certain areas
of the engine. Cooling,while it is apparently not a production
problem,is more complex than with the conventional engine for
several reasons. Part of the engine combustion chamber is always
exposed to hot gases while the remaining part is in contact with
the relatively cool fuel-air mixture. This gives rise to a signi-
ficant temperature gradient across the engine. In addition the
interior of the rotor must be cooled with either oil or fuel-air
charge. Another area of uncertainty is the cold-starting perform-
ance. -Some of the early production rotary engines exhibited cold
starting difficulties but this has been resolved in later designs.
Perhaps the most controversial aspect of the engine next to
fuel consumption is emission performance. In general, the engine
is viewed as having lower NO emissiort levels than a conventional
engine. Carbon monoxide emissions are essentially a function of air
fuel ratio as with the conventional engine. The level of unburned
hydrocarbons in the uncontrolled exhaust is generally far above
that of a conventional engine. It was reported that the high
level of hydrocarbon emissions from the General Motors prototype
rotary engine was the major reason this engine was not brought to
production in 1974.
-------
B. OBJECTIVES
The incentive for the present investigation is the consider-
able uncertainty associated with both, the fuel consumption and
exhaust emission performance of the rotary engine. In addition
only relatively little quantitative data have been published on
the engine. Most manufacturers have been exceedingly reluctant
to report engine data on preproduction engines. In order to make
effective and rational decisions related to powerplants in both
the public and private sectors it is imperative to have sound
technical information. The objectives of this program were design-
ed to meet this goal and are defined as follows:
1. Map the fuel consumption and emission performance of
both a modern rotary engine and a conventional recip-
rocating engine of equivalent power.
2. Compare the fuel consumption and emission performance
• of the selected test engines in a simulated identical
vehicle.
3. Determine the effects of the principal operating variables
on both emissions and fuel consumption performance of the
rotary engine.
4. Identify features of key fundamental processes which
could cause differences in both emission and fuel con-
sumption performance with respect to the conventional
engine.
C. TEST PROGRAM
In the present study both the fuel consumption and emission
characteristics of a modern rotary engine were measured under
steady-state conditions. These data were intended to provide a
more rational assessment of the potential of the rotary engine.
The particular engine selected for this evaluation was a 1975
Mazda model 13-B which is used in the Mazda RX-4. To preserve a
realistic perspective and a basis for comparison, a conventional
spark-ignited engine, a 231 cu.in. (3786 c.c.) 1975 model Buick
V-6, was tested under similar conditions. The Buick engine was
evaluated without catalytic convertor whereas the Mazda rotary
engine was tested both with and without its production thermal
reactor.
The engines were tested over a wide range of speeds and loads
simulating steady-state operation in a 3095 Ib. (1404 Kg) vehicle.
Data points were selected to facilitate effective mapping from
road load to maximum load. Since the Mazda RX-4 uses an axle ratio
of 3.9:1, whereas Buick in most of its vehicle applications gen-
erally employs a far lower ratio (approximately 2.56:1), tests
were run to encompass operation with both axle ratios. The effect
of axle ratio on fuel consumption with both engines could be as-
certained from these data.
-------
Emissions of carbon monoxide, carbon dioxide, unburned
hydrocarbons, oxygen and the oxides of nitrogen were measured at
each test point and in addition a special series of tests were
conducted to determine the aldehyde emissions. This latter series
of experiments were run since it was believed that the rotary
engine may have had a higher level of partially oxidized hydrocar-
bons than found in the exhaust of a conventional engine. Aldehyde
measurements were made at a relatively few representative condi-
tions because of the tedious requirements of the wet chemical test
procedure used.
In addition to the basic fuel economy and emission measure-
ments , a parametric study was conducted on two of the most import-
ant operating variables, air-fuel ratio and ignition timing, to
ascertain their influence on both emissions and fuel consumption
in the rotary engine. Several representative test conditions were
chosen for this parametric study. Also to gain a better understand-
ing of key fundamental aspects of the engines, two other areas
were selected for investigation:
a. Evaluation of the thermodynamic cycle during compression,
combustion, and early expansion with emphasis on determin-
ing losses due to gas leakage. A quartz piezoelectric
pressure transducer was used to measure the pressure data.
b. Determination of sources or processes resulting in the
formation of unburned hydrocarbons. As noted earlier
the most significant emission problem is the high level
of unburned hydrocarbons. It has been speculated that
both the high surface to volume ratio in the combustion
chamber and leakage past the leading apex seal to an
.exhausting chamber are the major causes. Hydrocarbon
source information were inferred from a spatially re-
solved hydrocarbon analysis of exhaust leaving the engine.
Both a timed sampling and steady-state sampling technique
were used.
-------
SECTION II
OBSERVATIONS AND CONCLUSIONS
Numerous observations and conclusions -can be drawn from the
results of this investigation both with regard to specific exhaust
emission and fuel consumption performance of the Mazda 13-B rotary
engine and in comparison with the 1975 Buick V-6.
A. At wide-open throttle the magnitude and trends of BSFC, BMEP,
brake torque, and BHP as a function of engine speed of the Mazda
engine were generally similar to the Buick reference engine with
several exceptions.
1. The BHP of the Mazda increased almost linearly with RPM
and peaked above 5000 RPM.
2. The BSFC decreased with increasing RPM, whereas it in-
creased for the Buick engine.
B. In the variable load studies of the Mazda engine the performance
parameters of BSFC, BSHC, BSCO and BSNO emissions were determined
as a function of RPM. x
1. The BSFC both with and without air injection was similar
at a given RPM and generally decreased with increasing load.
2. The BSNO emissions with and without air injection general-
ly were observed to increase with load, but a sharp decrease occurred
near maximum load.
3. The BSHC and CO emissions were strongly related to exhaust
air injection. With the thermal reactor and air injection, BSHC and
BSCO were extremely low. Where the injected air was diverted to the
reactor cooling jacket these emissions increased sharply. At speeds
above 4000 rpm, exhaust air injection was not used arid no CO and
hydrocarbon emissions reduction occurred by exhaust reactions. In
the studies without air injection BSHC and CO emissions generally
decreased with increasing load in the speed range from 1000-3000 rpm.
Beyond this speed these emissions were reasonably constant with load
variation.
C. Without thermal reactor and air injection the Mazda 13-B
engine as installed in a hypothetical 1975 Mazda vehicle showed
the following trends at road-load conditions.
-------
1. Fuel economy increased as axle, ratio wa,s, decreased from
3.90 to 2,56:1. Maximum fuel economy occurred between 3Q and 40
miles per hour... (26 mpg with. 2.55:1 axle, 23 mpg with 3..9Q:! axle). .
2, The. BSHC emissions exhibited little variation as a function
of axle ratio and speed except at the lowest road speeds. The
minimum BSHC level, 6-7 g/mile, occurred near 60-70 miles per hour.
3. The BSCO emissions decreased almost linearly with increas-
ing speed and reached a minimum level of approximately 30 grams per
mile at 80 mph. Axle, ratio exhibited little influence on CO emis-
sions .
4. Brake specific NO emissions increased at an increasing
rate with speed and peakedxat 4 g/mile at 80 mph. The low axle
ratio of 2.56:1 produced higher BSNO emissions by about 20%.
X
D. The following observations summarize the comparison between the
performance of the Buick and Mazda engines without external emission
control as installed in a simulated Mazda RX-4 vehicle.
1. The Mazda engine powered vehicle exhibited nearly 20% poor-
er fuel economy as a maximum difference. This occurred at 35 mph.
Above 35 mph the difference between the Mazda and Buick fuel economy
diminished considerably and were essentially identical at 60 mph
with the 3.90:1 axle ratio. Our studies indicate that the major
factor causing the fuel economy deficiency of the Mazda was the
rich mixture required for the effective operation of the thermal
reactor. By optimizing both the ignition timing and mixture ratio,
it was calculated that the fuel economy for the Mazda could be
brought within 5% of the Buick powered vehicle. This result was
achieved' in the absence of any specific emission control require-
ments .
2. BSHC emissions were considerably greater, approximately
6-10 times for the Mazda. This difference appears to be a funda-
mental Mazda engine characteristic since optimization of the operat-
ing variables could only reduce this difference by a factor of 2.
3. The BSCO emissions were considerably greater for the
Mazda. However, the difference decreased with increasing speed.
The difference does not appear to be fundamental and is primarily
attributable to the richer mixtures used in the Mazda.
4. The BSNO emissions were consistently higher for the Buick
than for the Mazda even though EGR was employed on the Buick. Rich-
er mixtures and slower burning in the Mazda contributed to the dif-
ference.
E. The 3SFC was consistently higher for the Mazda than the Buick
V-6. At 2500 rpm, for example, the minimum BSFC was .48 for the
Buick and .55 for the Mazda, The difference decreased at maximum
load and at the higher speeds in the test range. The fuel economy
results from the vehicle simulation/ which were generated from the
same data- showed the same trends.
-------
F. The motoring friction horsepower and mean effective pressure
increased at a significantly higher rate, in the Buick. engine. How-
ever the lower speed friction mean effective, pressure was higher
for the Mazda engine. These, data offer a partial explanation for
the decreasing difference in fuel economy observed with increasing
speed in the vehicle simulation.
G. The exhaust temperatures at similar conditions were similar in
the Mazda and Buick engines. This is primarily attributable to
the fact that the Mazda was operating considerably richer than
the Buick. Had they been operated with equivalent mixture ratios,
the Mazda would have exhibited considerably higher exhaust tempera-
tures .
H. The aldehyde emissions of the Mazda were generally in the range
reported as typical for reciprocating engines, 30 to 200 ppm as
formaldehyde. Aldehydes increased with increasing air-fuel ratio
almost linearily and also increased with increasing load without
the thermal reactor. When the thermal reactor and air injection
were employed, the aldehydes were reduced by a factor of approxi-
mately 5.
I. Air-fuel ratio influenced the fuel consumption and emission
performance of the Mazda engine in a manner essentially similar to
that for reciprocating engines.
1. Brake specific fuel consumption was a minimum between .9
and 1.0 equivalence ratio (.lean).
2. The BSHC emissions were minimum at approximately .9 equiva-
lence ratio and increased by approximately a factor of 2 at an
equivalence ratio of 1.4.
3. The BSCO emissions were constant at less than 25 grams
per bhp hr leaner than an equivalence ratio of approximately .95
and then increased linearily with mixture enrichment.
4. The BSNO emissions were a maximum between an equivalence
ratio of .9 and 1X0. Near this maximum point, load exhibited a
considerable effect on the emission level whereas at rich mixtures,
in the range of 1.2-1.4 equivalence ratio, the effect of load was
minimal.
J. In general the effect of ignition timing variations on fuel
consumption and emissions appeared to be considerably less than
expected with reciprocating engines.
1. The BSFC increased with, spark, retardation. For example
at 2000 rpm, 26.3 bhp, the BSFC decreased from .57 to .67 with 30°
retard from MBT.
2. The BSHC emissions decreased with spark, retard with lean
mixtures but exhibited little change when the "stock" richer
mixtures were used.
-------
3. Ignition timing did not influence. BSCO emission
4. Brake, specific NO emissions increased slightly with-
ignition retardation. "*
K. The hydrocarbon formation processes within the combustion
chamber of the Mazda engine are. judged to be non-homogeneous based
upon studies of the exhaust gas. The exhaust from the region of
the chamber nearest the apex seal was particularly high in hydro-
carbons in comparison to the exhaust from the more central portion
of the chamber. The variation between the maximum and minimum was
approximately 3 to 1.
These data suggest that the primary sources of unburned
hydrocarbons in the Mazda exhaust were from leakage of unburned
fuel past the apex seal and incomplete burning in the trailing apex
seal region of the chamber. The carbon monoxide emissions demon-
strated little spatial variation in the exhaust gas suggesting that
the mixture ratio was reasonably homogeneous in the combution
chamber.
Additionally the relatively high 02 content in the exhaust of
the Mazda engine even with rich fuel-air mixtures and no air injec-
tion suggests that a significant fraction of the fuel-air mixture
is exhausted without burning.
L. From the evaluation of the pressure records of the Mazda engine
and by comparing them with data from the literature, several obser-
vations can be made.
1. Firing engine - The polytropic exponents for both compres-
sion and expansion were lower than data reported for reciprocating
engines, suggesting that the heat transfer and/or leakage of the
gases near top dead center was greater in the rotary engine.
2. Motoring engine - The motoring engine exhibited higher
polytropic exponents than the firing engine. Compared to a recip-
rocating engine they were lower during compression and higher during
expansion. Again this suggests greater leakage and heat losses in
the rotary engine.
M. Qualitative Observations
The Mazda engine was remarkably smooth and quiet in the test
cell, whereas the Buick V-6 was considerably "rougher", particular-
ly at low speeds. Both engines demonstrated excellent durability
and no maintenance other than normal servicing was required.
-------
SECTION III
TEST EQUIPMENT
A. TEST ENGINES
Pertinent data for the 1975 Mazda Model 13-B rotary engine
and Buick V-6 test engines are shown in Table I. The Mazda engine
was furnished by the Toyo Kogyo Company of Japan and the Buick
engine by the Buick Division of General Motors Corporation. The
Mazda engine was evaluated both with and without its emission con-
trol system whereas the Buick was tested without a catalytic
reactor but with all other emission control devices including EGR.
As received, both engines were tailored to meet the 1975 emission
standards of 1.5, 15 and 3 g/miles of hydrocarbons, carbon monoxide
and oxides of nitrogen respectively.
Mazda Emission Control System
The Mazda 13-B engine was equipped with the standard 1975 U.S.
version emission control system consisting principally of an insu-
lated thermal reactor, Fig. 1, an air injection pump and an elec-
tronic control system. The standard Mazda vehicle emission control
box was replaced with a manually operated control by which it was
possible to control the air pump flow valves and spark plug firing
in a manner similar to the vehicle system. The operating sequence
of the air pump control valves and trailing and leading spark plugs
is shown in Fig. 2. In the Mazda system both load and speed
sensitive air control valves in addition to control of spark plug
firing were employed. These factors together with mixture ratio
were carefully tailored to ensure vehicle compliance with the 1975
Federal emission standards. The "A" air control valve is speed
sensitive and in the off position directs air to the exhaust gas
in the thermal reactor. When this valve is actuated (on) air is
diverted to the reactor cooling jacket. In the present studies
the valve was operated manually. The "C" valve is load sensitive
and permits functioning of a vacuum operated air injection cut-out
switch. At just "off idle" load the "C" valve is turned on. Air
injection is still allowed but now when the throttle is opened,
the vacuum operated air-injection "cut-out" valve stops air
delivery to the exhaust and directs it to the reactor cooling
jacket. Both the "C" valve and load sensitive "cut-out" valve
were operated automatically in this test program. A detailed
explanation of the emission control system is included in ref. 4.
No EGR is used to control NO .
-------
Model
TABLE 1
TEST ENGINE SPECIFICATIONS
Mazda Rotary
1975-13-B
Displacement in -(.cc)
80
(1308)
Rated BHP at RPM
Rated Torque at RPM
ft. Ib (Mm)
Compression Ratio
Carburetion
Ignition
Ignition Timing
Emission Control
Valve or Port
Timing
160 cid equivalent
4-cycle displacement
110 @ 6000
117 (159) <§ 3500
9.2:1
Buick V-6
1975 V-6
231 (3777)
110 @ 4000
175 (237) @ 2000
8.9:1
Downdraft, 2-stage Rochester Model 2GC
4-barrel
Kettering system-
2-spark plugs/housing
.Delco HEI system
0°BTC leading plug, 15° 12° ETC @ 600 rpm
ATC trailing plug, @ 500 rpm
Thermal reactor with air Catalytic reactor,
injection plus engine EGR, engine modifi-
modification cations, carb. pre-
heat
Intake opens 32°ATC
Intake closes 50°ABC
Exhaust opens 75°BBC
Exhaust closes 38°ATC
Intake opens 17°3TC
Intake closes 73°ABC
Exhaust opens 68°BBC
Exhaust closes 29°ATC
Volume displaced by 1 rotor face in 1 rotor revolution multiplied
by the number of rotors.
10
-------
Figure 1, Insulated thermal reactor used on the Mazda 13-B engine
as viewed from the. engine exhaust port contact surface.
WOT
LOAD
A Voiveon, C Voive on,Vacuurn..'
Cut-Out oruLeading Plug, No Air
•
Nwoive on xx>xSNx
$x\\\\\\\\VCSS>W
-------
B. DYNAMOMETER INSTALLATION
The Mazda 13-B and Buick V-6 engines were both installed in
Room 243 at the University of Michigan Walter E. Lay Automotive
Laboratory, and connected to a 200 HP Model Westinghouse electric
dynamometer. Since it. was neces.sary to test the Mazda engine at
higher speeds than permitted in the dynamometer cell, this engine
was connected to the dynamometer through a manual shift Mazda 4-
speed transmission which was run in third gear for the higher
speed tests. The installation is shown in Figs. 3 and 4. Because
of a slight inefficiency of the transmission, the dynamometer load
was adjusted to compensate for the .additional drag in the transmis-
sion. Toyo Kogyo provided mechanical efficiency data for this
correction. The Buick engine, which was connected directly to the
dynamometer, is shown in Fig. 5. Both installations were perform-
ed according to established practice.
1. Engine Cooling
Water was used as a coolant in an open, tower-type cooling
system. A thermostatically-controlled flow of city water was used
to maintain the coolant inlet temperature in a range specified by
the manufacturer.
2. Oil System
Since the Madza engine uses the "lubricant to cool the rotor
it was necessary to supplement the normal cooling available from
the engine itself. A special cross flow water-oil heat exchanger
was installed in the dynamometer cell for this purpose. The proper
control of the lubricant was maintained with the engine's own
internal thermistat rather than employing separate temperature
controls outside of the engine. No external oil cooling was re-
quired for the Buick V-6.
3. Exhaust System
The standard laboratory exhaust system was used to remove
the engine exhaust. In all tests either the standard major engine
system components (thermal reactor and muffler) or a simpler system
with a restriction inserted (to permit back pressure adjustment
to that of the "stock" system) was used. The Mazda engine was run
both with and without the thermal reactor installed. No catalytic
converter was used on the Buick.
C. INSTRUMENTATION
1. Fuel Measurement
The fuel rate was measured with a Model 1971 Testron volume-
tric burette associated with a timer and electronic counter. The
timer was triggered by electric contacts when the fuel in the
burette was displaced by electrolyte. The air flow measurements
12
-------
Fiure .
Rear quarter
installation
ing view o£ the Ma 3d a 13-3 engine
Figure 4, Side view of the Mazda 13-3 engine installation
-------
Figure 5.
Front quarterin
installation .
view of the 3uick V-6 eng.in
-------
were made concurrently with the fuel measurements. In this way,
any abnormalities in fuel or air consumption could be detected
and noted when evaluating the data. Several fuel readings were
taken during each run and averaged. The run was rejected if the
fuel flow varied by more than 5%.
Since the fuel volume was measured, it wa,s necessary to
determine the fuel density at the time and location of measurement.
This was done before each test by measuring the API (American Pe-
troleum Institute) number of a fresh sample of the indolene test
fuel with an API hydrometer. Fuel density was corrected to the
burette temperature.
i
2. Engine Air Measurement
Carburetor air flow was measured by a calibrated Meriam,
Model 50 MC2-4SF, 400 CFM laminar flow element in conjunction with
a Meriam, Model 34FB2, micromanometer, Fig. 6. To minimize flow
pulsations through the laminar element, the flow meter was mounted
on the inlet of a large surge tank. The surge tank was equipped
with a "blowout" membrane to prevent damage to the tank and flow
meter in case of "back-fire".
From the calibration curve of the flow element, the differen-
tial pressure across the flow element was converted to volumetric
flow rate. In turn the flow rate was corrected for temperature,
viscosity, pressure, and humidity to obtain a dry air-mass flow
rate. Before and after each test, the humidity was determined from
wet and dry bulb temperature readings.
3. Temperature Measurement
Because of the possible effect of temperature on engine
operation and exhaust emissions (both inside of the engine and in
the exhaust system), to ensure that the engines were stabilized
prior to data acquisition and to permit calculation of certain
performance factors, various temperature data were acquired. These
included the exhaust, oil, water, fuel and ambient temperatures.
All temperature measurements were made using thermocouples and
various potentiometer read-out devices.
A Brown dual-range, continuous indicating potentiometer was
used to read the copper-constantan (0-250°F, -18-121°C) and chro-
mel-alumel (0-1800°F, -18-982°C) thermocouples. The temperatures
at the exhaust port and at downstream sample probes were measured
using bayonet-type chromel-alumel thermocouples. These thermo-
couples were unshielded, and therefore the actual gas temperatures
were higher than those indicated because of the radiation losses
from the thermocouple to the cooler exhaust pipe surfaces.
4. Exhaust Emission Measurement
a. Exhaust Gas Sampling
15
-------
Water-Cooled Probes — For all of the baseline and parametric
variation tests a 1/4 in. (.,64 cm) water-cooled, stainless-steel
sample probe was used to collect the hot exhaust gas, Fig. 7. The
water-jacketed probe was not cooled at the sample inlet but at the
exterior surface of the exhaust pipe. The probe was designed to
cool the exhaust sample and minimize the possibility of chemical
reactions in the sample line. The "cooled" sample was then ducted,
at room temperature, to the analytical instruments. While these
probes could be positioned to sample anywhere across the diameter
of the pipe, early tests indicated no significant concentration
gradient across the diameter of the exhaust pipe downstream of
the reactor in the case of the Mazda engine or in the sample
section of the Buick's exhaust. As a result, the.probes were ad-
justed to sample from the center of the exhaust stream.
Teflon tubing (.1/4 in, .64 cm) 'was used for all sample lines
involved with hydrocarbon measurement and nylon tubing (1/4 in,
.64 cm) for the remainder of the sampling circuit. Previous work
with exhaust sampling bags has indicated that many common tubing
materials, including nylon, absorb hydrocarbons readily (5).
Because the reactor output hydrocarbon concentrations were often
very low and the inside of tubing-comprises a relatively large
surface area, Teflon tubing was used for any lines involved in
hydrocarbon measurement.
To prevent water vapor and particulate contamination of the
analytical system, an ice bath and particulate filter made by
Gelman Instrument Company were installed in the sampling circuits
ahead of the analyzers.
Uncooled Probes (Hydrocarbon Distribution Studies) -- In the
steady-state sampling studies to determine spatially resolved
hydrocarbon emissions in the exhaust a series of 10 sample probes
located 3 in (7.6 cm) apart along a length of exhaust pipe were
used, Fig. 8. The first probe .was located 1.5 in (_3.8 cm) from
the exterior face of the exhaust port. These probes were all un-
cooled and employed a single inlet hole in a plane parallel to the
axis of the exhaust pipe. Exhaust was sampled continuously. It
was impossible to cool these probes because of cost. However, in
prior studies it had been found that quenching of the sample had
relatively little influence on the higher concentration emission
data.
Flow from these probe.s was ducted to a manifold in which a
series.of valves enabled introduction of sample from a given probe
to the main sample line. Only the exhaust from one rotor was
sampled in this phase of the study. Specially contructed indivi-
dual exhaust pipes of 1.5 in (3.8 cm) inside diameter were fabri-
cated. The length of the pipe fitted with sample probes was de-
signed to accommodate, in the sample section, at least the total
mass of exhaust gas displaced by one rotor face at the highest
load tested and at the temperature and pressure of the exhaust.
16
-------
riani laminar flow meter and inlet surge chainbe;
ScilTiOiS OJTODfi-
17
-------
Sampling Valve (Hydrocarbon Distribution Studies) — The
key device used in the study to determine the time resolved com-
position of the exhaust leaving the exhaust port was a Cox type
6 timed sampling valve, Fig. 9. The valve was located 2.5 in.
(6.4 cm) from the exterior face of the drive end (toward dynamo-
meter) exhaust port. A schematic diagram of the Cox valve is
shown in Fig. 10. Basically the device consists of a 15/64 in.
(.6 cm) diameter valve head with a stem which extends into and
is a permeable core in an electromagnetic coil. An adjustable
spring maintains the valve in contact with its seat. At a speci-
fied time in the test cycle a large capacitor is discharged
through' the coil and the magnetic field established attempts to
eject the valve. The valve lifts approximately .012 in. (.03 cm)
from its seat and during this interval the sample is obtained.
Duration of valve lift was found to be less than 10 msec. Because
of low valve lift and relatively short duration of opening only a
small fraction of sample was extracted from the exhaust. The
sample flow rate through the valve was less than .5 cfh (3.9E-.06
m^/sec). The timed capacitor discharge was derived from a special-
ly constructed device available at the University of Michigan
Automotive Laboratory. The valve could be opened at a known and
controllable crank angle position. Since the unit was designed
for 4-cycle reciprocating engines, a pulse was generated and there-
fore the valve opened every 720 degrees of crankshaft rotation.
This meant that exhaust was sampled from every other rotor chamber
rather than the same one. Thus the exhaust was sampled from
chamber 1,3,2,1,4,2,...i.e. every other cycle. The sample obtain-
represented the integral of the exhaust from all three chambers.
Therefore differences between the performance of individual
chambers could not be detected.
b. Emission Analyzers
The various exhaust emission analysis techniques are sum-
marized in Table 2.
NDIR Carbon Monoxide, Carbon Dioxide — Carbon monoxide and
carbon dioxide concentrations were measured with NDIR analyzers.
This technique measures the differential absorption of infrared
energy between a fixed reference gas and the 'exhaust sample gas.
Because of the high concentration of CO in the rotary engines un-
controlled exhaust (0-10%) and even higher levels of CO- in all
cases (0-15%), short sample cells were used. Since with lean
mixture operation and/or after-exhaust treatment, concentrations
of CO can be very low, it was necessary to supplement the short
path Beckman instruments with a long path Horiba, Model A1A-21 CO
analyzer with a range of 0-1%. Before each test, the NDIR anal-
yzers were calibrated with dry N2 as a zero gas and several known
span gases.
FID Hydrocarbons — A Beckman 109A unheated flame ionization
detector (FID) was used to measure total hydrocarbons in the ex-
haust. This instrument is typical of those used by EPA and others
for modal hydrocarbon measurement.
18
-------
"»v ••
? 7 HI
Manifold sampling system used
of hydrocarbon distribution in
ie qualitative
exhaust eras.
9, Cox timed sarapling valve as installed in the exhaus-
pipe of the Mazda 13~B engine.
Valve
Electromagnetic
Coils
Return
Spring
!j.gura 10, Scheaiatic diagrara of the Cox timed sampling valve.
-------
TABLE 2
GAS ANALYSIS TECHNIQUES
Exhaust
Specie Technique
Carbon monoxide NDIR
Carbon monoxide
Carbon dioxide
Nitric oxide
Hydrocarbon
Oxygen
Aldehydes
NDIR
NDIR1
Chemilumines-
cent
FID2
Amperometric
DNPH3
Manufacturer
Beckman Inst. Model 315A
Horiba Inst. Model AlA-21
Beckman Inst. Model 315A
Thermoelectron Inst.
Model 10A
Beckman Inst. Model 109A
Beckman Inst. Model 715
Wet chemical and Bausch
&.Lomb Spectronic 20
Spectrophotometer
Range
0-10%
0-1%
0-15%
0-10,000
ppm
0-5000+
ppm
0-5% or
2-25%
1. NDIR-nondispersive infrared
2. FID-flame ionization detector
3. DNPH-dinitrophenylhydrazone wet chemical method.
A separate hydrocarbon sampling circuit, complete with con-
densate trap, pump, and particulate filter was branched off the
main sample line. Sample flow rate to the detector was mainted
between 4.7 E-05-4.8 E-05 m-^/sec with the aid of a Fischer and
Porter rotameter and needle type, flow control valve. Since the
accuracy of the FID is greatly affected by the sample bypass flow,
a Fischer and Porter 7.0 SCFH, rotameter was installed in the by-
pass line to insure the sample flow and calibration gas flow
through the burner were equal.
While the instrument was used in conventional form, the oper-
ating parameters were optimized to minimize error due to 0- inter-
ference and to provide equal response to the various hydrocarbons
in the exhaust.
FID fuel.
A mixture of 40% H2 and 60% He was used as the
For the Cox timed sampling valve studies, the sample system
was modified to minimize its volume to a minimum. This facilitat-
ed instrument response with the very low flow rates observed with
the Cox valve.
Oxygen -- The oxygen concentration in the exhaust was measur-
ed with a Beckman Model 715 amperometric O2 analyzer. Since the
20
-------
composition of the dry exhaust gas is not altered by passage
through the NDIR analyzers, the oxygen transducer was placed at
the outlet of the CO- analyzer. Since this instrument has a linear
response to oxygen concentration (actually partial pressure of 0-) >
it was calibrated with pure N_ as the zero gas and atmospheric
oxygen as the 21% 0- span gas.
Several gases can cause 0- interference, including SO-, C1-,
H-S and NO. Since most of these compounds are. not present in
measurable concentrations, only NO could cause significant inter-
ference. However, because the NO concentrations ranged from 50-
2500 ppm, its interference was also negligible.
The 0- meter provides temperature compensation for gas temp-
eratures between 32-110°F (0-43°C). After passing through the ice
bath, the sample temperature is within this range.
Chemiluminescent NO — The NO and NO- emissions were measur-
ed with a Thermoelectron Model 10A Chemiluminescent NO and N02
Analyzer. This instrument is capable to measuring either of the
two nitrogen oxides or a combined concentration. All NO data is
reported as NO. A discussion of the basic operation of this ana-
lyzer follows.
Sample is flowed to a cylinderical plug flow reactor which is
maintained at a low pressure, typically five to seven torr. Simu-
ltaneously, ozone manufactured in a built-in ozonator is introduc- •
ed into the reactor and well mixed with the sample in the region
near the detector. A reaction between NO and 0, occurs which pro-
duces some electronically excited NO- molecules. The decay to the
ground state of these excited molecules emits light (photons) in
the wave length region 0.6-3 u. This light is detected by a photo-
multiplier. The analyzer measures NO only, not N0_. Because of
the operating principle, the response is theoretically linear with
NO concentration. Response is limited only by the volume of the
sample lines and flow rate. By converting any exhaust NO- to NO,
a value of total nitrogen oxides, NO can be attained. Tne con-
version technique is based upon a surface reaction in which NO-
decomposes to NO. To accomplish this, the sample gas, before
entering the Chemiluminescent detector, is flowed through a stain-
less steel tube which is resistance heated to 1200°F (649°C). This
unit was periodically calibrated using appropriate span gases of
NO in nitrogen.
Aldehyde Measurement - Mazda Engine — Total exhaust aldehyde
emissions were determined colorimetrically using the DNPH (dinitro-
pheylhydrazone) wet chemical method, (6). In this procedure, a
small portion of the exhaust was bubbled through 2,4-dinitrophenyl-
hydrazine solution. This gas volume was measured with a wet test
meter. The precipitate formed was recovered, dissolved and treat-
ed with various chemicals until a color producing ion was formed.
The optical density of this mixture was measured by a Bausch and
Lomb Spectronic 20 Spectrometer calibrated on a formaldehyde
21
-------
solution. Since the. optical density is proportional to the amount
of carbonyls in the sample and the principal carbonyl in the exhaust
is formaldehyde, the DNPH wet chemical test is a good measurement
of the aldehyde concentration in the exhaust.
The aldehyde sample was taken approximately 20 in (50.8 cm)
downstream Strom the thermal reactor. Heater tape wrapped around
the sampling line maintained the sample temperature between 220-
240°F (.104-116°C) to prevent dissolution of all aldehydes in condens-
ed water.
5. Pressure Measurement
Pressure was measured with a Kistler Model 601 piezoelectric
transducer and Model 568 charge amplifier. The transducer was cal-
ibrated using a dead weight tester. It was installed in the trail-
ing spark plug hole in place of the auxiliary spark plug. Thus
during the firing sequences only the leading plug was fired. Figure
11 shows a sketch of the pressure transducer as installed in the
chamber and the rotor in two positions with respect to the location
of the transducer.
22
-------
CRANK ANGLE ~ 115° (or-245°)
Intake
Exhaust
Pressure Transducer
Leading
Spark Plug
Expansion Gases
CRANK ANGLE-120° (or -240°)
Intake
Exhaust
Intake Mixture
Pressure Transducer
Leading
-*Spark Plug
Figure 11- Schematic diagram of the pressure transducer as
installed in the Mazda engine with the rotor in
two different positions.
23
-------
SECTION IV
TEST PROCEDURE
The testing program was divided into the following segments.
.Emission and fuel consumption map of the Mazda engine both
with and without external emission control (thermal reactor).
.Emissions and fuel consumption map of the Buick V-6 engine
used for comparative purposes.
.Parametric studies on the Mazda engine to determine the
influence of F/A ratio and ignition timing on emissions and
fuel consumption at selected test conditions.
.Evaluation of exhaust aldehyde composition at selected test
conditions.
.Determination of pressure records.
.Timed sampling and manifold sampling studies to determine
the sources of hydrocarbon emissions.
A. EXHAUST EMISSION AND FUEL ECONOMY MAPPING
The 1975 Mazda RX-4 vehicle was selected for definition of the
engine performance requirements. Pertinent vehicle data is given
in Table 3. The road load data for this vehicle, which was fur-
nished by Toyo Kogyo, is also presented in Table 3. Two different
axle ratios,3.90:1 and 2.56:1 were used to match the engine to the
vehicle. Mazda normally uses the higher ratio in the RX-4 and
Buick generally employs a much lower ratio in its heavier vehicles.
Both ratios were selected to span the range typically used in most
vehicles.
The engines were "broken in" per each manufacturers recom-
mendations. In addition key operating parameters were checked
to ensure that they were within a range specified by the manu-
facturer.
The basic wide-open throttle performance of both engines was
determined according to SAE procedure J245 with the exception that
the stock exhaust system was used on the Mazda. Since the operator
was located in the test cell the dynamometer speeds were limited
to 3500 rpm maximum. However, the Mazda engine was run at a con-
siderably higher speed, 5000 rpm, without exceeding 3500 dynamo-
meter rpm through the use of transmission speed reduction. Since
24
-------
the Mazda engine-transmission combination was used, all dynamo-
meter loads were adjusted to compensate for transmission efficiency.
Pertinent transmission data for the Mazda are also shown in Table
3.
TABLE 3
1975 MAZDA RX-4 VEHICLE SPECIFICATIONS
AND ROAD LOAD DATA
Curb Weight:
Overall Weight:
Tire Size and Radius:
Transmission Ratios and
(.Efficiency) :
2795 Ib. (2168 kg)
3095 Ib. (1404 kg)
195/70SR13, 11.4 in (0.291 m)
1st - 3.683
2nd - 2.263
3rd - 1.397 (.95)
4th - 1.0 (.98)
Vehicle,
MPIi
10
20
30
40
50
60
70
80
Speed
Km/hr.
16.1
32.2
48.3
64.4
80.5
96.5
112.6
128.7
Engine RPM(§
3.90:1
572
1144
1716
2288
2860
3432
4004
4576
Axle Ratio
2.56:1
375
751
1126
1502
1877
2253
2628
3003
Horsepower
HP
1.72
3.86
6.85
11.11
17.05
25.11
35.71
49.26
All Mazda operation above 3500 rpm was in third gear. The
Buick engine maximum speed was limited to 3500 rpm since it is
normally used at a far lower maximum speed than the Mazda. In
general sufficient numbers of points were selected to facilitate
development of reasonably smooth fuel consumption and emission
performance curves. A particularly large number of points were
used to map NO data with the Buick engine which was equipped with
an SGR controlxvalve. Significant inflections of the NO data
were observed for small load changes in the test range.
All emission and fuel consumption data were reduced and plot-
ted with the aid of a Hewlett-packard 9830A programmable calcula-
tor-computer. Emission measurements are reported on a mass basis
(grams) which necessitated the development of a program to convert
concentration data to a mass base. This task was complicated when
air injection was used because of exhaust dilution with the added
25
-------
air. A carbon mass balance technique between the engine inlet
and exhaust was used to generate a correction for the dilution
factor.
B. MAZDA PARAMETER STUDIES
Fuel/Air Ratio
The fuel/air ratio was varied with the aid of a specially
modified carburetor and air aspirator system furnished by Toyo
Xogyo. This system enabled the pressure on the fuel in the carbu-
retor float bowl to be varied from a partial vacuum to a positive
pressure and this caused the fuel rate to be varied. Since it was
expected that the effect of fuel/air ratio would follow similar
trends at all speeds and loads, only three operating conditions
were studied; 2000 rpm, 8.7 and 26.3 bhp, and 3000 rpm, 18.7 bhp.
Two series of variable fuel/air ratio investigations were perform-
ed:
1. with standard ignition timing
2. MET* timing
Ignition Timing
Ignition timing was varied by rotating the distributor and
observing the spark advance or retard with a timing light direct-
ed at a flywheel mounted scale. This test series was conducted
with both the stock carburetor mixture ratio and with a controlled
air/fuel ratio of 16/1. The latter test was run with 16/1 air-fuel
ratio rather than the typical reference setting of LET (leanest
air-fuel ratio for best torque) because the 16/1 ratio could be
easily set, was expected to be near the point of maximum thermal
efficiency, and a reasonable operating point should a catalyst
be used in the future.
The ignition timing studies were also conducted at 2000 rpm,
8.7 and 26.3 bhp and 3000 rpm, 18.7 bhp.
C. ALDEHYDE EMISSIONS
Since the DNPH technique for aldehyde measurement is extreme-
ly tedious, data were acquired at selected test conditions. One
series was conducted at 2000 rpm, 8.7 bhp with variable air-fuel
ratio and MET ignition timing. The other tests were run at 2000
rpm, 8.7 and 27 bhp and 3000 rpm, 18.5 bhp both with and without
the thermal reactor installed. Details of the sampling procedure
are presented in the earlier instrumentation section and in Refs.
6 and 7.
*MBT - minimum spark advance for best torque
26
-------
D. HYDROCARBON SOURCE ANALYSTS - TIMED AND MANIFOLD SAMPLING
SYSTEM
Time Resolved Sampling
This phase of the study was conducted with the Cox fast-
response sampling valve. Time resolved samples of the exhaust
were collected at 2000 rpm, 8.7 and 14.2 bhp, 2500 rpm, 13 and
22.4 bhp, 3000 rpm, 18.7 and 34.3 bhp from the specially construct-
ed exhaust pipe. The time of sample acquisition was coordinated
with crankshaft position which.in turn was related to rotor posi-
tion. All data expressed on a concentration basis.
Manifold Sampling
In this segment of the investigation data was acquired at the
same conditions as in the foregoing series. It should be noted
that this data was far more qualitative than that obtained with
the Cox valve.
E. MAZDA ENGINE CHAMBER PRESSURE MEASUREMENT
Pressure data were acquired under both motoring and firing
conditions at 2000 'and 2500 rpm. The firing data were taken at
part load. Motoring data were taken at both part and wide-open
throttle. All data were recorded on a Tectronix dual-beam oscil-
loscope. These results in turn were replotted, processed and
analyzed with the aid of the Hewlett-Packard Model 9830A calculator.
27
-------
SECTION V
RESULTS AND DISCUSSION
A. FUEL CONSUMPTION AND EMISSION PERFORMANCE OF THE MAZDA 13-B
ENGINE
1. Base Performance Data - Wide-Open Throttle
Initially a series of baseline wide-open throttle performance
tests were conducted per SAE procedure J245. The corrected brake
horsepower, brake mean effective pressure, brake torque, and brake
specific fuel consumption are plotted as a function of RPM in Fig.
12. Generally the shapes of these curves are reasonably similar
to those for reciprocating engines with the exception of the brake
specific fuel consumption, BSFC. In this case, BSFC decreased
significantly with increasing speed. This behavoir is probably
attributable to the fact that the leakage of the fresh fuel and air
mixture from a burning chamber to the exhausting chamber is essen-
tially constant as a function of time at a given mean effective
pressure. As the speed is increased, this leakage decreased as a
percentage of the total mass of fuel and air provided to the engine,
thereby having a lesser impact at the higher speeds than at the
lower speeds. In addition, the friction horsepower increased at a
lower rate than generally observed in a typical reciprocating spark-
ignited engine, Refs. 8,9 and Fig. 42. The brake torque and brake
mean effective pressure, EMEP, curves exhibited normal behavior al-
though they peaked at a higher RPM than expected for a low-speed
reciprocating engine. The brake horsepower, BHP, increased almost
linearly with speed and attained a maximum beyond the test range.
The trends evident in these wide-open throttle performance curves
appear to be consistent with the excellent breathing characteristics
and the lower friction horsepower at high engine speeds which are
characteristic of rotary engines.
2. Fuel Consumption and Emission Performance Maps with Thermal
Reactor, Air Pump and Exhaust System Installed
The BSFC map at the various test speeds is plotted as a
function of BHP in Fig. 13. It should be. noted that BHP was
selected as the abscissa coordinate rather than BMEP for two
important reasons. First, this helped to spread the data to pre-
vent a significant "stack up" of the various curves. Secondly,
with the brake horsepower information it is far easier to match
the engine to the. total vehicle.. The BSFC curves have reasonably
28
-------
130
120
110
'5 100
Q.
CL 90
LJ
m 80
70
60
.8
1975 Mazda I3'B
~ .7
Q.
x:
.Q
.6
01
CD
.5
NO
100
90
80
70
60 I
GO
50
40
30
20'
10
h- BSFC
1
I
120
100 £
80
60
40 g
<
20 m
1000 2000 3000 4000 5000
RPM
Figure 12. Wide-open throttle performance of the Mazda 13-3
engine as a function of RPM.
29
-------
1.5
1.4
1.3
1.2
5 1.1
f 1.0
CP
o 0.9
u_
c/)
00 0.8
0.7
0.6
0.5
1975 Mozdo 13-B
Thermal Reactor,
Air Inj.
RPM
+ 1000
O 1500
A 2000
X 2500
D 3000
V 3500
a 4000
O 4500
O 5000
I
I
I
I I I I
10 20 30 40
50 60
BMP
70 80 90 100 110 120
Figure 13.
Brake specific fuel consumption as a function of
brake horsepower, Mazda 13<-B engine, with thermal
reactor and air injection.
-------
typical shapes, decreasing relatively quickly as BHP increased
because friction horsepower was rapidly decreasing as a fraction
of indicated power. For the lower speed runs the. minimum BSFC
point was attained at less than peak load which, is typical of most
spark-ignited engines. At moderate, and high speeds, the wide-open
throttle BSFC was less than at slightly lower loads which is not
typical. This can be attributed to the relatively lean mixtures
furnished by the Mazda carburetor at peak loads.
The exhaust emission characteristics of the Mazda 13-3 engine
with stock air injection with controls, and thermal reactor are
plotted in Figs. 14 through 16. Specific emission rates are pre-
sented as grams of pollutant per brake horsepower hour and plotted
as a function of BHP.
Figure 14 is a plot of brake specific hydrocarbon emissions,
BSHC, as a function of BHP. At each speed, the lower load ranges
were characterized by generally low levels of hydrocarbon emissions.
As the load was increased to the point where the vacuum-operated
air valve (C-valve) cut out the air injection to the reactor, the
hydrocarbon emissions increased significantly. The C-valve did not
operate at the 1000 rpm point but the air injection rate was not
sufficient to achieve a sustaining reaction at higher loads. In
addition both spark plugs were firing at 1000 rpm. For assistance
in understanding the actuation of the air injection to the thermal
reactor, the reader is referred to Fig. 2 in Section III. At 4500
and 5000 rpm, which are well beyond the normal driving range except
for a transient acceleration or deceleration mode, the hydrocarbon
emissions were relatively flat or constant at approximately 10 gram/
hp-hr. At these speeds the A valve was "on" diverting the air
from the exhaust ports to the reactor cooling jacket. Also both
spark plugs were operating in each housing reducing the hydrocarbon
input to the reactor. The tests at 4000 rpm were conducted with
the A valve "off". This speed, as shown in Fig. 2, is normally
the switching speed for the valve.
The brake specific carbon monoxide, BSCO, emissions are plot-
ted in Fig. 15, and show a similar shape to the hydrocarbon data.
Initially at the lower speeds and lighter loads the emission level
was low followed by a rapid increase in BSCO at the higher load
conditions due to elimination of the thermal reactor air injection.
At 4500 and 5000 rpm the lighter load CO emissions were moderately
high, approximately 50 grams/hp-hr, followed by a spike to approxi-
mately 200 grams/hp-hr. Thereafter the CO emissions were reduced
rapidly to approximately 100 grams/hp-hr through the remainder of
the load range. The spike was caused by a region of rich mixture
at moderate load. Air was not being injected at 4500 and 5000 rpm.
At 5000 rpm, for example, the air-fuel ratio was 13:1 at 71 bhp.
At wide-open throttle, the production carburetor provided an air-
fuel ratio of 14.5:1.
31
-------
u>
K)
RPM
1000
O 1500
A 2000
X 2500
D 3000
V 3500
D 4000
O 4500
5000
1975 Mazda 13'B
Thermal Reactor,
Air Inj.
0
110 120
Figure 14.
Brake specific hydrocarbon emissions as a function of
brake horsepower, Mazda 13-B engine, with thermal
reactor and air injection.
-------
Figure. 16 is. a plot of the. brake spe.ci.fic NO emissions,
BSNO , as a function of BHP, In general the behavior was relative-
ly erratic as is the. case with, most gasoline engines. At 1000
through 350Q rpm, the emissions increased to a maximum followed by
a decrease at higher loads. The peak was associated with relative-
ly lean mid-range mixture ratios. At 4500 and 5000 rpm, the NO
emission levels followed a consistent pattern. At light loads x
the firing of both spark plugs caused the NO to be higher than
light loads at the lower speeds. With increasing load NO increas-
ed, followed by a dip associated with the enrichment described
earlier. It should be recognized that the thermal reactor has
virtually no effect on the NO concentration. It is an oxidation
device for the control of CO and unburnsd hydrocarbons.
33
-------
300
u>
I975 Mozdo I3-B
Thermal Reactor,
Air Ini.
+ 1000
O 1500
A 2000
X 2500
D 3000
V 3500
D 4000
O 4500
5000
40 50
100 110 120
Figure 15.
Brake specific carbon monoxide emissions as a function
of brake horsepower, Mazda 13-B engine, with thermal
reactor and air injection.
-------
U)
en
10
9
8
7
6
O>
3
2
1
0
0
RPM
+ 1000
O 1500
A 2000
X 2500
D 3000
V 3500
D 4000
O 4500
a 5000
1975 Mazda 13-
Thermal Reactor,
Air Inj.
10 20 30 40
_L
50
60
BMP
70 80 90 100 110 120
Figure 16.
Brake specific nitrogen oxide emissions as a function
of brake horsepower, Mazda 13-B engine, with thermal
reactor and air injection.
-------
3. Exhaust Emission and Fuel Consumption Performance Without
Reactor Air Injection
In this series of tests the air injection was cut off to the
thermal reactor and used totally as a coolant in the jacket around
the reactor. The BSFC performance of the engine with and without
the air injection-was essentially the same, as would be expected,
therefore these data are not re-plotted.
The hydrocarbon emissions are shown in Fig. 17. At 1000 and
1500 rpm the data were erratic. At moderate loads the emission
level increased rapidly to a maximum and then exhibited a moder-
ate decrease at the highest loads tested. The variation appeared •
to be consistent with variations in the engine air-fuel ratio.
No explanation is apparent for the disproportionally low minimum
at the light loads since the NO emissions were also relatively
low. Hydrocarbon performance may have been influenced by a leak
through the air control valve to the thermal reactor at these
points. The remainder of the hydrocarbon data followed a reason-
ably consistent pattern. As speed and load were increased, the
specific hydrocarbon emission level trended downward. The 4500
rpm point exhibited a slight variation from this behavior with a
moderate peaking of the data in the mid-load range. The hydro-
carbon results are closely related to air-fuel ratio and signifi-
cantly to speed and load. As noted earlier at a given mean effec-
tive pressure the leakage rate of unburned fuel from the combustion
to exhausting chamber is thought to be essentially constant in
mass time per unit. With increasing mixture flow the leakage con-
tribution to the total decreased on a fractional basis. The trend
of the data strongly supports the view that leakage past the apex
seal is a predominent factor in the formation of exhaust hydro-
carbon emissions. In addition, at the higher loads at a given
speed a more favorable condition exists for the after or thermal
reaction of the unburned fuel and air-mixture in the exhaust. Of
course 'this assumes that sufficient 0- is present.
The 02 content of the exhaust provides an indication of leak-
age and/or misfire. Figure 18 is a plot of exhaust 0- concentra-
tion as a function of 3HP at several representative engine speeds.
With the rich mixtures normally used in the Mazda engine, one
would expect the 0- content of the exhaust to be low if combustion
is reasonably complete. The significant 02 content present is a
strong indication that a - substantial fraction of the mixture has
not burned and either reflects leakage to the exhausting chamber,
incomplete flame propagation or unburned mixture trapped and then
released from the chamber crevices. It is our belief that the
leakage factor is the most important of the three. Unfortunately,
it is not possible to separate the influence of the various hypo-
thesized sources with conventional emission sampling and analysis
techniques. Gas chromotographic studies could provide consider-
able insight as to the source however .
36
-------
30
1975 Mozda 13-B
Thermol .Reoctor
No Air Inj.
20
U)
I
Q.
O>
o
X
-------
3.0
2.0
U)
CO
z:
LJ
O
>-
X
o
1.0
0
1975 Mazdo I3~B
Thermal Reactor
No Air Inj.
RPM
O 1500
X 2500
V 3500
_L
0 10
Figure 18,
20
30
40
BMP
50
60
70
80
Exhaust oxygen concentration as a function of brake
horsepower, Mazda 13-B engine, with thermal reactor
but without air injection.
-------
The BSCO emission data are shown in Fig. 19. In general the
data follow a similar pattern to that of the hydrocarbon emissions.
The light load data at 1000 and 1500 rpm was relatively low, again
perhaps caused by slight air leakage to the thermal reactor. The
remainder of the data appear to be consistent with the trend in
mixture ratio.
For comparative purposes, plots of air-fuel ratio as a
function of BHP are. shown in Fig. 20, at several representative
speeds. The performance of the Mazda carburetor is reasonably
typical of conventional engine carburetors, although the mid-
range mixture ratios are considerably richer than normally used.
The thermal reactor required careful tailoring and generally rich
mixtures to operate effectively.
Figure 21 is a plot of the BS NO as a function of BHP and
shows that the NO emissions with or without air injection were
similar. Certainly this was expected since the major operating
variables were the same. The variation exhibited seems consistent
with our experience with other engine data in which variations
in atmospheric conditions, mixture ratio, and spark timing from
test to test, could readily cause the observed differences.
4. Exhaust Emission and Fuel Consumption Performance Without
Thermal Reactor and No Exhaust Port Air Injection
In this series of tests the thermal reactor was replaced with
a conventional exhaust manifold. The air-injection air pump was
operated but no air was injected into the exhaust stream. The data
from this test series have been used in the comparison with the
Buick 231 CID V-6 engine. As anticipated, fuel consumption and
emission results are similar to the previously discussed tests with
thermal reactor but without air injection. Several minor anomolies
were observed which may have been related to changes in carburetion,
thermal oxidation in the engine and/or exhaust system, and effects
due to day-day variations in atmospheric conditions.
The BSFC curve is shown in Fig. 22. In general the data shows
that the specific fuel consumption decreased with increasing load
and speed which was the case for the engine operating with the
thermal reactor with and without the air injection. The only dif-
ferences of significance were the disappearance of the slight peaks
observed in the earlier data at 3000 and 3500 rpm in the mid-load
ranges. This reduction in the mid-load range can be attributed to
a slightly leaner mixture relative to the prior data. At 4000 rpm
the BSFC curve is of similar shape but the moderate load data is
slightly higher than observed in Fig. 13.
The BSHC emission performance of the engine as a function of
BHP is shown in Fig. 23. These curves are almost identical to the
data with the thermal reactor and no air injection. In the study
39
-------
300
250
1975 Mazda 13-B
Thermal Reactor
No Air Inj.
200
O
jr
a.
_a
O
O
CO
CD
150
100
50
0
i
1
1
RPM
4- 1000
O :1500
A 2000
X 2500
D 3000
V 3500
a 4000
O 4500
1
1
0
10 20 30 40
50 60
BMP
70 80 90 100 110 120
Figure 19.
Brake specific carbon monoxide emissions as a function
of brake horsepower, Mazda 13-B engine, with thermal
reactor but without air injection.
-------
15
14
I975 Mazda I3-B
Thermal Reactor
No Air In).
LJ
ID
12
11
10
RPM
O 1500
X 2500
V 3500
1
0 10
20 30 40 50
BHP
60
70
80
Figure 20.
Air-fuel ratio as a function of brake horsepower at
selected test conditions, Mazda 13-B engine, with
thermal reactor, but without air injection.
-------
13
1975 Mazda I3~B
Thermal Reactor
No Air Inj.
.X
10
N)
O.
_C
_a
\
CP
CO
CD
RPM.
+ 1000
O 1500
A 2000
X 2500
D 3000
V 3500
D 4000
O 4500
110 120
Figure 21.
Brake specific nitrogen oxide emissions as a function
of brake horsepower, Mazda 13-B engine, with thermal
reactor but without air injection.
-------
OJ
1.5
1.4
1.3
1.2
2 11
I 1-0
o
0.9
0.7
0.6
0.5
0.4
0
1975 Mazda 13 ~B
w/o Thermal Reactor 8
Air Inj.
RPM
+ 1000
O 1500
A 2000
X 2500
D 3000
V 3500
D 4000
O 4500
O 5000
4000
I
1
I
10 20 30 40
50 60
BHP
70 80 90 100 1 10 120
Figure 22. Brake specific fuel consumption as a function of brake
horsepower, Mazda 13-B engine, without thermal reactor
and air injection.
-------
o
X
CD
70
60
50
40
30
20
10
1975 Mazda 13 ~B
w/o Thermal Reactor
Air Inj.
0
RPM
+ 1000
O I500
A 2000
X 2500
0 3000
V 3500
D 4000
O 4500
O 5000
1
1
1
0 10 20 30 40
50 60 70
BMP
80 90 100 110 120
Figure 23.
Brake specific hydrocarbon emissions as a function of
brake horsepower, Mazda 13-B engine, without thermal
reactor and air injection.
-------
with the thermal reactor the light load hydrocarbon emissions at
1000 rpm were low and followed no distinct trend. As noted this
was believed to be due to leakage of air past the air control
valve. Hydrocarbon emissions generally decreased with increasing
speed and load. An exception to this trend was observed at the
highest load setting at most test speeds because of mixture enrich-
ment. Another anomally compared to the data for the engine with
thermal reactor and no air injection is the substantially higher
hydrocarbon emissions at 4500 rpm maximum load. The mixture ratio
was essentially the same as that observed in the prior test series.
Therefore the increase was probably due to a slight misfire. Con-
sidering the reasonable scatter expected in engine emission data,
it appears that the results are remarkedly repeatable.
The BSCO emissions are shown in Fig. 24. The anomalies
observed for the hydrocarbon emissions at the highest loads at
both 1000 and 500 rpm are present. Again this can be attributed
to a slight inadvertent air injection in the studies with the
thermal reactor. The general trend shows clearly the predominant
influence of mixture ratio on CO emissions. Again the data is
reasonably consistent with the results shown earlier.
The brake specific NO emissions are shown in Fig. 25. The
trend of the data is consiltent with the measurements obtained
from the engine with thermal reactor and both with and without air
injection. However, significant variations in the vertical position
of most of the curves with respect to earlier data were present.
At 1000 rpm, the light-load emissions are less than 1 gram/bhp-hr
followed by a rapid increase to a maximum at approximately 5.5
which in turn is followed by a rapid decrease to less than 1 at
maximum load. The peak is considerably higher than observed earli-
er. At 1500 rpm the shape and magnitude of emissions are similar.
At 2000-3500 rpm the shape is consistent, showing that BSNO is a
maximum in the mid-load range. The peak emissions were consistent-
ly lower than those observed in the study with the thermal reactor.
The 4000 rpm data is similar in shape to the lower RPM results,
and similar in magnitude to the earlier data. At 4500 RPM the
curve is characterized by two peaks at light and moderate load,
and the curve is generally lower than observed with the thermal
reactor. At 5000 rpm the characteristic shape is similar to the
4500 rpm results. The trends at a given speed were consistent with
variations in air-fuel ratio and appear to be influenced by igni-
tion timing as well. In this engine the timing is set in a re-
latively narrow band which suggests that at both higher speeds
and lighter loads the effective timing may be retarded from the
MET setting, whereas M3T is approximated more closely at lower
speeds and at heavier loads. The effect of ignition timing on
all emissions will be discussed in a later section. The shifts
in magnitude of brake specific NO emissions from data observed
earlier is in most cases signficant. This can be attributed to
several factors. Normally the repeatability of NO emissions is
X
45
-------
_c
I
Q.
.£>
O
o
CO
CD
700
650
500
550
500
450
400
350
300
250
200
150
100
50
0
1975 Mazda 13'B
w/o Thermal Reaction 81
Air Inj.
RPM
+ 1000
O 1500
A 2000
X 2500
D 3000
V 350O
D 40OO
O 4500
O 50OO
1
1
1
1
1
1
1
1
1
0 10 20 30 40
Figure 24.
50 60
BMP
70 80 90 100 110 120
Brake specific carbon monoxide emissions as a function
of brake horsepower, Mazda 13-B engine, without thermal
reactor and air injection.
-------
11
10
9
8
o
1 4
0
RPM
+ 1000
O 1500
A 2000
X 2500
D 3000
V 3500
D 4000
O 4500
O 5000
D
1975 Mozdo 13-B
w/o Thermal Reaction 8
Air Inj.
X
I
I
0 10 20 30 40
50 60
BMP
70 80 90 100 110 120
Figure 25.
Brake specific nitrogen oxides emissions as a function
of brake horsepower, Mazda 13-B engine, without thermal
reactor and air injection.
-------
not good because NO is extremely sensitive to small variations
in operating -parameters. For example, a variation in air-fuel
ratio of as much as 1/2 ratio was observed between the test with
thermal reactor to the present test. In addition, changes in
atmospheric conditions such, as inlet temperature, and humidity
appear to have a reasonably significant effect. Another poten-
tial factor causing the deviation is related to the basic charac-
teristics of the data. The NO emissions tend to rise quickly
with increasing load followed By a rapid decrease. A slight shift
in the load setting could cause a considerable variation in
magnitude of NO emissions. For example, at 1000 rpm if the mid-
range load had Been set slightly lower than in the test, the
•magnitude of the NO peak could have been attenuated significantly.
j£
5. Fuel Consumption and Emission Performance of the Mazda 13-B
Engine without Thermal Reactor in a Simulated Mazda RX-4 Vehicle
To provide a more graphic basis for comparison, the fuel
consumption and emission performance of the rotary engine as instal-
led in the simulated Mazda RX-4 vehicle were determined under road
load conditions. These data are plotted in Fig. 26 through 29 as
a function of vehicle speed. Two axle ratios, 2.56 and 3.9:1, were
selected as the .final drive ratio. These data in no way reflect
the type of performance that could be expected on the EPA emission
cycle, rather they reflect the projected performance of this engine
in a simulated vehicle at a series of steady-state test conditions.
The fuel economy in miles per gallon is plotted in Fig. 26.
The shape of the curves and their relative position are similar to
those expected for reciprocating engine powered vehicles. The use
of the higher axle ratio results in a considerably lower fuel e-
conomy, approximately 3 miles per gallon, across the range of the
test. With the lower axle ratio the engine is operating at a high-
er load factor and lower RPM for a given vehicle speed. At the
higher load factor the engine thermal efficiency is significantly
greater because of the reduced throttling losses, less exhaust
residual dilution and other more minor factors. In addition, the
lower RPM results in less engine friction (higher mechanical ef-
ficiency) . The shapes of these curves are similar. The fuel e-
conomy is relatively constant through the mid-range, suggesting
that at the lower speeds the engine efficiency is increasing at
approximately the same rate as the road load horsepower require-
ment as vehicle speed increases. At the higher speeds the road
load horsepower, which contains an important term proportional to
velocity cubed (aerodynamic resistance), increases rapidly and be-
comes the dominant factor resulting in a significant reduction in
fuel economy.
The hydrocarbon emissions in grams per mile are plotted as a
function of miles per hour in Fig. 27. There is no significant
difference exhibited with the two axle ratios. The level of emis-
sions would be considered high by comparison with typical recipro-
cating engines. The trend of the data suggests that mass hydro-
carbon emissions decrease with speed. This characteristic can
48
-------
o
CL
I
>-
Ld
35
30
o 25
o
o
Ld
X-
X
1975 Mozdo 13-B Engine
Simulated Mazda RX-4 Vehicle
Road Load
2.56 Axle Ratio
20
3.90 Axle Ratio
15
0
10
20
30 40 50
VEHICLE SPEED (MPH)
60
70
Figure 26.
Fuel economy as a function of vehicle speed of the
Mazda 13-B engine installed in a simulated Mazda RX-4
vehicle, road load, 2.56 and 3.90:1 axle ratios.
80
-------
en
o
16.0
14.0
to
| 12.0
o»
CO
o
CO
CO
LJ
CO
CO
10.0
8.0
O
cr 6.0
o
o
Q 4.0
2.0
0
Mazda 13-B
Road Load
+ + 3.90 Axle
O O 2.56 Axle
1
0 10 20 30 40 50 60 70
VEHICLE SPEED (MPH)
80 90 100 110
Figure 27.
Mass hydrocarbon emissions as a function of vehicle
speed of the Mazda 13-B engine installed in a simulated
Mazda RX-4 vehicle, road load, 3.90 and 2.56:1 axle
ratios.
-------
probably be attributed to a significant decrease in the percentage
of mixture that is leaked to the exhausting chamber volume with an
increase in engine air flow. An additional factor causing this
decrease is leaning of the air-fuel ratio in the mid-operating
range.
The mass emission rate of CO is shown in Fig. 28. Here also
little difference is evident between the two axle ratios. The
data trends downward significantly with increasing vehicle speed al-
though with the 3.90:1 axle mass emission rate increased near maximum
speed. The similarity of the two curves is indicative of the fact
that the carburetted air-fuel ratio used with the two axle ratios
is essentially the same.
Figure 29 shows the mass emission level of NO as a function
of vehicle speed. Clearly the level of NO emissions increases
with increasing speed because of a significant increase in engine
load factor. The 2.56 axle ratio exhibits a greater mass NO emis-
sion rate than the 3.90 axle because it also causes the engine to
operate at a greater load factor than with the higher axle ratio.
The curve shapes for the NO data are reasonably typical of that
for reciprocating gasoline Ingines installed in vehicles.
B. FUEL CONSUMPTION AND EMISSION PERFORMANCE OF THE BUICK 231 CU..
in. V-6 ENGINE, NO EXTERNAL EMISSION CONTROL
The Buick 231 cu. in. V-6 engine was tested to obtain a piston'
engine comparison to the Madza 13-B rotary engine. The engine was
tested without emission control to facilitate comparison to the
rotary engine without emission control. The Buick engine is repre-
sentative of a modern medium-sized reciprocating engine, utilizing
a relatively efficient open combustion chamber, and has demonstrated
attractive EPA fuel economy performance in vehicle installations.
In addition, it is evident that with the increasing concern for
energy conservation as well as reduction of emissions, there will
be a general scaling down in the displacement of American automo-
tive engines. The V-6 promises to be a very important configura-
tion in American passenger cars during the next ten year period.
1. Base Performance Data - Wide Open Throttle
Initially the Buick engine was tested per SAE procedure J245
to determine its wide-open throttle performance characteristics.
The engine was not tested with the standard exhaust system, but
with an exhaust restriction installed to provide the back pressure
of a production system. The WOT performance data are plotted in
Fig. 30. The BHP, BMEP, and brake torque data are all corrected
per SAE procedure to standard conditions. The torque and mean
effective pressure curves are relatively flat with a maximum occur-
ring between 2000 and 2500 rpm. The BHP increases at a decreasing
rate beyond 2500 rpm which is typical. The engine was not tested
to the BHP peak because this region is not generally important in
51
-------
U1
160
140
120
E
01 100
i
CO
i 80
CO
CO
o
o
CO
60
40
20
0
Mazda 13-B
Road Load
+ + 3.90 Axle
O O 2.56 Axle
I
I
I
I
0 10 20 30 40 50 60 70
VEHICLE SPEED (MPH)
80 90 100 110
Figure 28.
Mass carbon monoxide emissions as a function of vehicle
speed for the Mazda 13-B engine installed in a simulat-
ed Mazda RX-4 vehicle, road load, 3.90 and 2.56:1 axle
ratios.
-------
8.0
6.0
o>
E
i
CO
s40
CO
LJ
x
O
co or.
to 2.0
0
Mazda 13-B
Road Load
+ + 3.90 Axle
O O 2.56 Axle
0 10 20
Figure 29.
30 40 50 60 70 80
VEHICLE SPEED (MPH)
90 100 110
Mass nitrogen oxide emissions as a function of vehicle
speed for the Mazda 13-B engine installed in a simulat-
ed Mazda RX-4 vehicle, road load, 3.90 and 2.56:1 axle
ratios.
-------
165
150
135
120
UJ
5
00 105
90
1975 Buick23lcidV-6
120
100
80
QL
60 OQ
40
20
1.0-
0.9-
_ 0.8
w.
x:
O.
S
CO
CD
0.6
0.5
0.4
Brake Torque
•BSFC
1000
2000
RPM
3000
180
150
LJ
90
60
30
a:
ao
4000
Figure 30
Wide-open throttle performance of the 1975 Buick V-6
engine as a function of RPM.
54
-------
the typical driving cycle and because of safety consideration in
the laboratory. The specific fuel consumption increased almost
linearily with RPM. This increase is associated primarily with
the nearly cubic increase in friction horsepower with increasing
engine speed.
2. Fuel Consumption and Exhaust Emission Maps of the Buick V-6
Engine with EGR and without Catalyst
Figure 31 is a plot of the BSFC as a function of BHP in the
range from 1000 to 3500 rpm in 500 rpm increments. These data
are perceived to be typical of modern reciprocating engine be-
havior also. Specific fuel consumption was high at low-load
factors because friction horsepower is a significant fraction of
indicated horsepower. At moderate loads BSFC was reduced to
approximately 0,5 lb/hp-hr. and then increased sharply at wide-
open throttle because of mixture enrichment. This enrichment was
particularly large at high loads since the part-load mixtures were
generally leaner than stoichiometric in most of the operating
ranges. Generally the curves were much smoother than those of the
Mazda.
Brake specific hydrocarbon emissions are shown in Fig. 32.
All the data were reasonably consistent. At the lowest speeds,
1000-2000 rpm, and light loads, residual dilution and perhaps re-
duced combustion efficiency caused a reasonably high level of
hydrocarbon emissions. Idle mixture enrichment and spark retard
were also factors in these higher hydrocarbon mass emissions. As
the load was increased, BSHC emissions were reduced since BHP was
increasing at a greater rate than air flow. At and near wide-open
throttle where mixture enrichment was used for maximum power and
internal engine cooling the BSHC emissions increased sharply.
The BSCO emissions shown in Fig. 33 are typical of recipro-
cating engines in which reasonably lean mixtures are used through
most of the driving range. At high loads mixture enrichment caused
a dramatic increase in CO level.
The BSNO emissions are shown in Fig. 34 as a function of BHP.
These data proved to be very complex and difficult to acquire
because of the rapid change in NO emissions with load change at
each speed due to EGR valve action. Consequently, a large number
of data points were used to generate the precise shape of these
curves. In general, the NO emissions increased with increasing
loads. However, they were attenuated as the exhaust gas recircu-
lation was increased. Near maximum load NO levels were reduced
sharply because of mixture enrichment. Thexirregularity in these
curves is reasonably typical of modern engines equipped with vacu-
modulated EGR, and suggest the potential of better EGR control
schemes. Ignition timing is an additional factor affecting NO
levels. However, it was believed that the Buick timing was pro*-
grammed to follow closely the MBT settings except at very light
55
-------
en
OY
1.5
1.4
1.3
1.2
1.1
1.0
.0
o 0.9
CO
00 0.8
0.7
0.6
0.5
0.4
1975 Buick 231 cid V-6
EGR, no cotalyst
1
1
1
RPM
+ 1000
O 1500
A 2000
X 2500
0 3000
V 3500
1
1
1
0
10 20 30 40
50 60
BMP
70 80 90 100 110 120
Figure 31.
Brake specific fuel consumption as a function of brake
horsepower, 1975 Buick V-6 with EGR and no catalyst.
-------
en
RPM
1000
1500
A 2000
X 2500
3000
1975 Buick 231 cid V-6
EGR.no catalyst
10 20 30 40 50 60 70 80 90 100 110 120
Figure 32. Brake specific hydrocarbon emissions as a function of
brake horsepower, 1975 Buick V-6 with EGR and no
catalyst.
-------
250
200-
150-
tn
CO
i
Q.
O
o
CO
CD
1975 Buick 231 cid V-6
EGR.no catalyst
RPM
+ 1000
O 1500
A 2000
X 2500
D 3000
3500
100
0 10 20
Figure 33.
30 40 50 60 70 80 90 100 110 120
BMP
Brake specific carbon monoxide emissions as a function
of brake horsepower, 1975 Buick V-6 with EGR and no
catalyst.
-------
16
1975 Buick231 cid V-6
EGR, no catalyst
12
Ul
I
Q.
_0
O>
X
O
z:
CO
QQ
8
0
L
1
1
1
1
1
RPM
+ ' 1000
O 1500
A 2000
X 2500
D 3000
V 3500
1
1
0
10 20 30 40 50
60 70
BMP
80 90 100 110 120
Figure 34.
Brake specific nitrogen oxide emissions as a function
of brake horsepower, 1975 Buick V-6 with EGR and no
catalyst.
-------
loads where significant re.tarda.tion was. used.
Figure 35 shows a plot of air'-fuel ratio as a function of
horsepower at 1750 and 3000. rpm. The mixture, is relatively close
to stoichiometric at low loads leaner than stoichiometric at mod-
erate loads and is enriched considerably at the higher load oper-
ating conditions.
Using the emission maps, the Buick engine was installed in a
hypothetical vehicle, Mazda RX-4, with two different axle ratios,
2.56 and 3.9:1 to assess the simulated steady-state mass rate of
emissions. The various emissions in grams/mile are plotted as a
function of vehicle speed at road load conditions in Fig. 36. The
CO emissions increase moderately with increasing speed and then
are reduced. The CO emissions for the higher axle ratio are great-
er than for the lower 2.56:1 axle ratio. The hydrocarbon emissions
are reasonably similar in shape to the CO curves. However, above
35 mph the high axle ratio exhibits a lower hydrocarbon emission
level than with the lower axle ratio.' Axle ratio has apparently
little effect on the NO emissions which increased at an increas-
ing rate with speed.
C. FUEL CONSUMPTION AND EXHAUST EMISSION COMPARISON BETWEEN
MAZDA 13-B AND BUICK V-6 ENGINES
The fuel consumption and exhaust emission performance of the
test engines are plotted as a function of vehicle speed as install-
ed in the simulated Mazda RX-4 automobile in Figs. 37 through 40,
Figure 37 is a plot of fuel economy in miles per gallon as a func-
tion of speed for both engines with two different axle ratios.
With the 2.56 axle ratio the Mazda engine exhibits a maximum of
approximately 20% lower fuel economy and this occurs at 35 mph.
With the 3.90=1 axle ratio the maximum difference is approximate-
ly 15% which also occurs at 35 mph. As vehicle speed is increased
beyond 35 mph, the difference in fuel consumption between the two
vehicles decreases considerably. This can be attributed primarily
to the significantly lower friction at high speeds with the Mazda
engine. It must be recognized that a large part of the observed
difference is associated with the type of emission control techni-
que selected for the engines. The Mazda is operating with far
richer fuel-air ratios than those for best thermal efficiency
whereas the Buick V-6 is operating very close to its peak efficien-
cy air-fuel ratio. The Mazda carburetor is calibrated rich because
the thermal reactor used in the present Mazda design requires a
rich mixture to develop a self-sustaining oxidizing reaction for
control of the unburned hydrocarbons and CO. Another important
consideration relates to the vehicle package itself. In this
particular vehicle simulation,.the weight was the same for both
the rotary and V-6 engines. The weight of an actual vehicle with
a rotary engine would be less and thus the vehicle could be made
smaller and lighter for a given interior package volume especially
if secondary weight savings were made. These factors suggest that
60
-------
18
1975 Buick23lcid V-6
EGR, no catalyst
1750 rpm
-Q 3000 rpm
16
<
g
H-
o:
UJ
14
a
10
i
l
l
l
l
i
l
l
l
0
10 20 30 40
50 60
BMP
70 80 90 100 110 120
Figure 35. Air-fuel ratio as a function of brake horsepower for
the 1975 Buick V-6 engine.
-------
8
1975 Buick23lcid V-6
EGR, no catolyst
B * A 2.56 oxle
Q O A 3.90 axle
/
10
Figure 36,
20
30 40 50 60
VEHICLE SPEED (MPH)
70
80
90
Mass emission rate of hydrocarbons, carbon monoxide,
and nitrogen oxides as a function of vehicle speed,
Buick V-6 engine installed in a simulated Mazda RX-4
vehicle, road load, 2.56 and 3.90:1 axle ratios.
-------
o
O
O
LL)
_)
UJ
30-
25
20
10
0
0
Figure 37.
Road Load Fuel Economy
O—O 2.56 Axle- Mazda 13-B
+—4-3.90 Axle-Mazda 13-B
£—& 2.56 Axle- Buick V-6
O—O 3.90Axle-BuickV-6
I
10 20 30 40 50 60
VEHICLE SPEED (MPH)
70
80
Road load fuel economy as a function of vehicle speed-
a comparison between the Mazda 13-B and Buick V-6
engines installed in a simulated Mazda RX-4 vehicle.
-------
the deviation in fuel economy performance, would be reduced with
optimization of both, operating parameters and vehicle package.
The hydrocarbon emissions in grams per mile are plotted for
both vehicles in Fig. 38. In this figure a major deficiency of
the present Mazda engine design is clearly demonstrated. The
hydrocarbon emissions of the Mazda are approximately eight times
those of the Buick powered vehicle. The major factors causing
this major difference, are:
.Significantly richer air-fuel ratio employed by the Mazda
.Major loss of unburned mixture either due to leakage or
poor combustion
The compact combustion chamber of the Buick has a minimum quench
area and its open design promotes efficient flame propagation
throughout the total volume. In addition, the engine is well seal-
ed from the ambient with the poppet type exhaust valves and multi-
ring gas sealing around the pistons.
Figure 39 shows the CO emissions in grams per mile as a func-
tion of vehicle speed. The difference in CO emissions between
engines is also dramatic although it is not of the same magnitude
as the-hydrocarbon emissions. The deviation, between engines
expressed in this figure is primarily associated with the far
richer mixture ratio employed in the Mazda. At an equal fuel-air
ratio the mass CO emissions would be essentially the same. On
the other hand the hydrocarbon emissions while also related to air-
fuel ratio are in addition associated with fundamental -processes
not closely related to operating variables but rather to basic
design characteristics.
The mass emission rate of NO as a function of vehicle speed
for the two engines is shown in F?g. 40. In this case the emis-
sion performance relationship between the two engines has revers-
ed and the Buick exhibits a higher level of NO emissions than
the Mazda. This greater level in the Buick isxin part due to the
leaner mixtures employed although it must be recognized that the
Buick V-6 was equipped with EGR whereas the Mazda engine was not.
The fact that the data for this comparison vvere taken from both
engines without their normal exhaust after-treatment devices is of
little consequence since neither the thermal reactor or presently
used designs of catalytic reactors affect NO emissions.
J\
Representative brake specific fuel consumption data from the
Mazda 13-B and Buick V-6 engines are plotted as a function of BHP
in Fig. 41 at selected engine speeds. As is true of the miles per
gallon plot the Buick exhibits considerably better fuel consump-
tion. The brake specific fuel consumption is approximately 20%
lower throughout most of the test range although there are regions
where it is similar particularly at high engine speeds and near
maximum load at all speeds. It should be observed that if the
Mazda engine had been operated with as lean carburetion as the
Buick engine, the difference in BSFC and for that matter MPG
64
-------
a\
JU
E
\
co
CO
co
16.0
12.0
O
co 10.0
CO
LJ
•2.
a
CO
a:
o
o
cr
a
80
6.0
4.0
2.0
+
\
Rood Load Hydrocarbon Emissions
O-—O 2.56 Axle-Mazda 13-B
+ + 3.90Axle-Mazda13-B
2.56 Axle-Buick V-6
3.90Axle-BuickV-6
\
\
\
\
\
-1-
0 10
Figure 38.
30
40 50 60 70
VEHICLE SPEED (MPH)
80 90 100 110
Road load hydrocarbon emissions as a function of
vehicle speed - a comparison between the Mazda 13-B
and Buick V-6 engines installed in a simulated Mazda
RX-4 vehicle.
-------
CTl
Qt
E
x
CO
2
O
160
140
120
100
80
UJ
o 60
o
40
20
4-
\
\
Rood Load CO Emissions
O O 2.56 Axle-Mazda 13-B
+ + 3.90Axle-Mozdo13-B
A A 2.56Axle-BuickV-6
3.90Axle-BuickV-6
\
\ V
\+\
\\
/
0
0 10
Figure 39.
I oi fy^-o-tr^—r^
20 30 40 50 60 70
VEHICLE SPEED (MPH)
80 90 100 110
Road load carbon monoxide emissions as a function
of vehicle speed - a comparison between the Mazda
13-B and Buick V-6 engines installed in a simulated
Mazda RX-4 vehicle.
-------
8.0
6.0
to
E
o>
\
CO
2 4.0
CO
CO
UJ
x
O
CO
CO
Rood Lood NOX Emissions
O O 2.56 Axle - Mazda 13"B
3.90 Axle - Mozdo I3~B
2.56 Axle-BuickV-6
3.90 Axle- Buick V-6
2.0
20
30
40 50 60 70
VEHICLE SPEED (MPH)
80
90 100 110
Figure 40 - Road load nitrogen oxide emissions as a function of
vehicle speed - a comparison between the Mazda 13-B
and Buick V-6 engines installed in a simulated Mazda
RX-4 vehicle.
-------
CO
1.5
x:
a
JC.
.0
O
Lu
DO
1.0
0.5
^
0
O—O Mazda 13'B
Buick V-6
— J* C^4500rpm
- ISOOrpm
±
j_
_L
^SOO^pm
10
20 30 40
50 60
BMP
70 80 90 100 110
Figure 41.
Brake specific fuel consumption comparison between the
Mazda 13-B and Buick V-6 engines as a function of brake
horsepower at selected operating conditions.
-------
would have been considerably reduced, For example, consider
operation at 2QOQ rpm. In the Buick. engine with, the development
of 20 hp, the air-fuel ratio was 15.2:1 and in the Mazda engine
at approximately the same load, the measured air-fuel ratio was
13.5:1. If the mixture ratio had been optimized for thermal
efficiency in the Mazda, the BSFC could have been reduced from
approximately 0.63 to 0.56, a 13% improvement. Thus air-fuel
ratio optimization alone could bring the Buick V-6 and Mazda
engine performance considerably closer together. Unfortunately,
as noted before, the rich mixtures used in the Mazda are required
because of the selection of the thermal reactor emission control
device. If, in addition, to air-fuel ratio optimization the
spark had been advanced closer to its optimum (.MET) as was the
apparent case for the Buick V-6, the fuel consumption differences
would have been even less. Thus it can be concluded that while
one observes a significant deficiency of the rotary engine in
comparison to the reciprocating engine, this defficiency is not
totally associated with fundamental factors but rather because
it has not been optimized with respect to operating parameters
because of the choice of the thermal reactor as a control system.
This consideration is particularly important when estimating the
long-term potential of the rotary engine.
One of _the most interesting comparisons between the Mazda 13-B
and Buick V-6 engines is evident in the plot of friction horsepower
and friction mean effective pressure as a function of engine speed
in Fig. 42. These data were all taken at wide-open throttle in a
motoring engine per accepted procedure. The friction horsepower
curves show that the rotary possesses a significant advantage over
the V-6 under these conditions. This advantage is attenuated
somewhat in the plot of friction mean effective pressure which
rates the friction performance on a per unit displacement basis.
In any event the rotary appears to have a favorable relationship
to the reciprocating engine at higher speeds and of significance
is the fact that the rate of increase in both the FMEP and FHP is
considerably lower for the rotary engine. It must be remembered
that when one views the brake thermal efficiency of the total
engine this reflects both the efficiencies of the various thermal
processes (indicated thermal efficiency) and, in addition, the
mechanical efficiency of the mechanism. Thus while the combustion
chamber surface to volume ratio of the rotary is not as well suit-
ed to attaining an ultimate indicated thermal efficiency, it does
possess attractive mechanical efficiency which can at least parti-
ally compensate for indicated performance deficiencies. Further,
these data lend credence to the argument that the rotary engine
is well suited to high-speed operation. It must be remembered,
however, that the data presented here as is the case with most
reported friction data was obtained from a motoring engine. Thus
it is a measurement under an unrealistic operating condition.
For example, in the rotary engine, the gas pressure behind the
sealing grid, which assists contact of the seals with various
engine surfaces, is considerably lower than in the firing engine.
The same effect is present in the reciprocating engine to a
69
-------
40
in
O.
u3 30
^ 20
cc
U.
10
Buick V-6
Mazda 13-B
40
30
20
or
10
0
Buick V-6
Mazda 13-B
0
Figure 42.
1000 2000 3000
RPM
4000
5000
Motoring friction mean effective pressure and friction
horsepower as a function of RPM for the Mazda 13-B
and Buick V-6 engines.
70
-------
degree. It is only possible to infer the. relative friction
performance from this data rather than to say that is is a true
measure of firing engine friction.
Another interesting aspect of the Mazda performance in com-
parison with the Buick is the relationship between their re-
spective exhaust gas temperatures. While this comparison is re-
latively qualitative because of the nature of exhaust temperature
measurement (.limited radiation shielding and variation in sensor
location with respect to the engine exhaust port) it is useful
for assessing the potential for "clean-up" due to thermal reaction
in the exhaust. Exhaust temperature is plotted at selected speeds
as a function of brake horsepower in Fig. 43. In general these
results showed that the exhaust temperature of the Mazda is simi-
lar to the Buick. engine. Exhaust temperature data in the litera-
ture, Ref. 10 , suggests that the Mazda' exhaust temperature should
be considerably higher at a given speed and load than the recipro-
cating engine. This behavior may be true here as well if the
engines had been compared at an equivalent set of operating condi-
tions. But, as noted earlier, the Mazda engine is operating with
a considerably richer mixtures resulting in a significant reduc-
tion in exhaust temperature over most of the operating range. It
must be noted however that temperature is only one factor govern-
ing the ability to oxidize the exhaust gas. This is particularly
true with a thermal reactor. Mazda believes a sacrifice in
exhaust temperature is more than offset by enrichment of the mix-
ture.
D. ALDEHYDE EMISSIONS FROM THE UNCONTROLLED AND EMISSION
CONTROLLED MAZDA MODEL 13-B ENGINE
Aldehydes, particularly formaldhyde and acrolein have been
shown to be important constituents in spark-ignited engine exhaust,
(6,7,n). They are apparently direct contributors to problems
associated with photochemical smog and exhaust odor.
Aldehydes which are an intermediate product of hydrocarbon
combustion are generally associated with low temperature combustion
reactions. Wall-quenching or quenching by excessive exhaust dilu-
tion are possible ways in which partially oxidized hydrocarbon
molecules may arise. In addition low temperature reactions may
occur during expansion, blowdown, and exhaust which may either
decrease or increase the aldehydes emissions.
Little published data exists on aldehyde emissions from inter-
nal combustion engines. This is in part because of the complexity
of making such measurements rapidly. The wet chemical techniques
used are complex and time consuming.
71
-------
I
LJ
or
I
cr
LJ
CL
2
LU
CO
<
X
X
Id
1700
1600
1500
1400
1300
1200
1100
1000
900
/y
Buick V-6
&—A 1500 RPM
V—V 2500
O C> 3500
Mazda 13-B
O-O 1500
D—0 2500
H 1- 3500
O—O 4500
0
Figure 43.,
10 20 30 40 50 60 70 80 90 100 110 120
BMP
Exhaust gas temperature as a function of brake_horse-
power at selected conditions, Mazda 13-B and Buick V-6
engines.
72
-------
Several series of experiments were conducted on the Mazda
engine to determine the aldehyde concentration using the DNPH
method. In the first series the aldehydes were determined at
2000 rpm, 8.7 bhp for the uncontrolled engine at several air-fuel
ratios. These data are plotted in Fig. 44. The aldehyde concen-
tration was observed to increase almost linearly with air-fuel
ratio. This trend is generally consistent with behavior exhibited
by spark-ignited reciprocating engines. Clearly in the rich
mixture operating region, typically used in the Mazda engine for
effective operation of thermal reactor, the aldehydes are sub-
stantially lower than with operation at lean mixture ratios.
A second series of tests was conducted with and without the
thermal reactor at stock engine operating conditions. The test
points selected were 2000 rpm, 8.7 bhp and 26.3 bhp and 3000 rpm,
18.6 bhp. Figure 45 shows clearly that the thermal reactor
effectively reduces the aldehyde concentration by a factor of more
than 4. The reactor is particularly effective at higher speeds
and higher loads. The effect of load is shown in the data and
illustrates that the aldehydes tend to increase with load, although
additional factors in this relationship would be changes in mixture
ratio and ignition timing.
For comparative purposes typical aldehyde data for a 350 cid
(,5735 cc). spark ignited reciprocating engine are presented in
Table 4. Note that the data were, not obtained from the Buick
reference engine nor at similar operating conditions to the Mazda.
Therefore only a highly qualitative comparison is justified.
TABLE 4
ALDEHYDE EMISSIONS FROM A 350 cid (5735 cc) AUTOMOTIVE
ENGINE AT 1200 RPM, 30 BHP (Ref. 11)
Air-Fuel Ratio Aldehydes - ppm CH.,, 0
12 120
14 130
16 170
18 200
The trends of aldehyde emissions as a function of air-fuel ratio
are similar for both engines.
E. INFLUENCE OF AIR-FUEL RATIO ON SPECIFIC FUEL CONSUMPTION AND
EMISSIONS - MAZDA 13-B ENGINE
Three representative test conditions were selected for the
parametric study of air-fuel ratio. These were 2000 rpm, 3.7,
and 26.3 bhp and 3000 rpm, 18.7 bhp. Two series of tests were
73
-------
180
160
140
O 120
100
CO
o 80
CD
CC
O
60
40
20
0
11
1975 Mazda 13-B
2000 RPM
8.7 BMP
12
13 14 15
AIR FUEL RATIO -A/F
16
17
18
Figure 44.
Aldehyde emissions as a function of air-fuel ratio,
the Mazda 13-B engine at 2000 rpm, 8.7 brake horse-
power.
74
-------
70
60
o
£50
Q.
Q.
CO
40-
O
£ 30
<
o
20
10
1975 Mazda 13-B
2000 RPM
Without
Thermal
Reactor
D 3000 RPM
With Thermal Reactor
2000 RPM
D
3000 RPM
10
20
25
30
Figure 45. Aldehyde emissions as a function of brake horsepower
at selected conditions, Mazda 13-B engine with and
without thermal reactor.
75
-------
conducted, the. one with MET ignition timing and the other with
stock ignition timing. The. trends observed were in general
similar to those observed for conventional reciprocating spark-
ignited engines (,12,13J_. For each series the individual dependent
variables of brake specific fuel consumption and emissions are
plotted as a function of equivalence, ratio* for the three test
conditions.
1. MET Timing
Brake specific fuel consumption as a function of equivalence
ratio is plotted in Fig. 46. The minimum occurs slightly lean of
stoichiometric and is significantly lower for the 2000 rpm, 26.3
bhp test point. The result at 26.3 bhp can be attributed to a
higher load factor as well as more efficient combustion resulting
from lower exhaust residual dilution and faster burning, because
of reduced heat transfer per unit mass of the working fluid. At
8.7 bhp, 2000 rpm the residual dilution is probably the most
significant factor in the higher BSFC results.
Brake specific hydrocarbon emissions are plotted in Fig. 47.
In general the curve shapes are similar to one another and to the
BSFC curves. However, several interesting points are evident. At
2000 rpm, 8.7 bhp the minimum occurs at approximately .95 equiva-
lence ratio and the upturn in BSHC at leaner mixtures is gradual.
At 2000 rpm, 26.3 hp, the minimum occurs near an equivalence ratio
of .85, however, the data is not definitive in this region. The
3000 rpm, 18.7 bhp curve was the most erratic.. It demonstrated
a significant dip in hydrocarbon emissions at the leanest mixtures,
and no minimum point was evident in the range tested. The signi-
ficantly lower emissions at lean mixtures compared to the 2000 rpm
conditions may have been due to some exhaust reaction in the
exhaust system without the thermal reactor. This reaction was
observed in earlier work on a Curtiss-Wright rotary engine (10).
The BSCO emissions, shown in Fig. 48, were similar to one
another at the three speed and load conditions and exhibited
behavior similar to that expected for reciprocating spark-ignited
engines. Air-fuel ratio is well recognized as the dominant factor
in CO emission levels.
Figure 49 is a plot os BSNO emissions. The NO emissions
are maximum slightly on the lean side of stoichiometric. The
important effect of engine load on this constituent is shown by
the comparison between the 26.3 and 3.7 bhp data at 2000 rpm. The
maximum at the heavier load setting is approximately five times
*Eqmvalence ratio or relative fuel-air ratio is the ratio of the
actual fuel-air ratio to the stoichiometric fuel-air ratio or the
ratio of the stoichiometric air-fuel ratio to the actual air-fuel
ratio.
76'
-------
1.3
1.2
I.I
1.0
i
Q.
-C Q
•-'
o
.8
C/)
°o .7
1975 Mozdo 13-B
.8 .9
Figure 46.
I
1
MBT Spark
2000 rpm
8.7 bhp
3000 rpm
18.7 bhp
2000 rpm
26.3 bhp
.0 I.I 1.2 1.3
EQUIVALENCE RATIO
1.4
1.5
1.6
Brake specific fuel consumption as a function of equiva-
lence ratio, Mazda 13-B engine, MBT spark.
-------
CO
45
40
35
f 30
CL
JC.
JD
™ 25
u
S 20
CD
15
10
0
.8
1975 Mozdo 13-B
.9
Figure 47.
1
1
MBT Spark
.0 I.I 1.2 1.3
EQUIVALENCE RATIO
3000 rpm
18.7 bhp
2000rpm
8.7 bhp
2000 rpm
26.3 bhp
1.4
1.5
1.6
Brake specific hydrocarbon emissions as a function of
equivalence ratio, Mazda 13-B engine, MBT spark.
-------
-C
.0
CP
O
O
700
650
600
550
500
450
400
350
300
250
200
150
100
50
0
8
1975 Mazda I3'B
MBT Spark
3000 rpm
18.7 bhp
.9
1.0 I.I 1.2 1.3
EQUIVALENCE RATIO
2000 rpm
8.7 bhp
A
2000 rpm
26.3 bhp
1.5
1.6
Figure 48.
Brake specific carbon monoxide emissions as a function
of equivalence ratio, Mazda 13-B engine, MBT spark.
-------
co
o
MBT Spark
1975 Mazda 13-B
2000 rpm
26.3 bhp
3000 rpm
18.7 bhp
2000 rpm
8.7 bhp
I.I 1.2 1.3
EQUIVALENCE RATIO
Figure 49.
Brake specific nitrogen oxide emissions as a function
of equivalence ratio,- Mazda 13-B engine, MBT spark.
-------
that at the lower load. This difference, however, essentially
disappears at mixture ratios greater than 1.1. The 3000 rpm NO
emissions near stoichiometric mixture ratio are greater than the
light load 2000 rpm emissions probably because of higher combustion
temperatures. Decreased residence time at these higher tempera-
tures is less significant than the temperature effect.
2. Stock Ignition Timing
The effects of air-fuel ratio on specific fuel consumption
and exhaust emissions for the stock ignition timing are shown in
Figs. 50 through 53. In this test series the effective ignition
timing was retarded with respect to the stock, setting at both
leaner and richer mixtures than those furnished by the stock
carburetor. The data trends are similar to the results with MET
timing. The BSFC data, Fig. 50, are almost identical in shape but
are universally displaced to higher values as expected.
The BSHC emissions shown in Fig. 51, also follow a similar
trend to those with MET spark. However at richer mixtures the
BSHC was generally slightly higher than at MET. In addition, at
leaner mixtures Cequivalence ratio of approximately .9) the emis-
sions appeared to decrease significantly. The primary factor in
this attentuation was the effective, spark retard from M3T due to
slower burning coupled with, increased thermal reaction.
The BSCO emissions are shown in Fig. 52 with stock timing.
The results are essentially identical to those with MET timing
and show the very clear relationship between CO and the equiva-
lence ratio. No evidence is present indicating a CO thermal
reaction with lean mixtures as the CO emissions at this condition
are already low. In addition, CO does not react as readily as
unburned hydrocarbons in the thermal reactor. Brake specific NO
emissions are shown in Fig. 53. Both,the trend and magnitude of
the data are similar to that observed with MBT timing. It was
anticipated that the effective retard would result in a moderate
reduction in NO from the data at MBT but this did not occur.
X
In summary, it is evident that the air-fuel ratio influence
on emissions and fuel consumption of the rotary engine is similar
to that expected for reciprocating gasoline engines. One exception
may be the apparent significant thermal reaction of hydrocarbons
at lean mixtures at some conditions, particularly with slight
retardation of the timing from MBT. This can be attributed to
the normally higher exhaust temperatures (at similar operating
conditions) in the rotary engine in comparison to those observed
in a reciprocating engine.
31
-------
CD
tsj
.c
I
Q.
x:
.O
\
.o
O
U.
(/)
GO
1.3
1.2
I.I
1.0
.9
.8
.7
.6
.5
Stock Spark
1975 Mazda I3~B
1
.8 .9
Figure 50.
1.0 I.I 1.2 1.3
EQUIVALENCE RATIO
D 2000 rpm
8.7 bhp
3000 rpm
18.7 bhp
2000 rpm
26.3 bhp
1.4
1.5
1.6
Brake specific fuel consumption as a function of equiva-
lence ratio, Mazda 13-B engine, stock spark timing.
-------
00
u>
Q.
.c
cn
OQ
45
40
35
30
25
20
15
10
0
.8
Stock Spark
1975 Mazda 13 ~B
2000 rpm
26.3 bhp
.9
1
2000 rpm
8.7 bhp
3000 rpm
18.7 bhp
1.0 I.I 1.2 1.3
EQUIVALENCE RATIO
1.4
1.5
Figure 51. Brake specific hydrocarbon emissions as a function of
equivalence ratio, Mazda 13-B engine, stock spark timing
-------
GJ
.£=•
700
650
600
550
500
450
o>
o 350
m 300
250
200
150
100
50
0
8
Stock Spark
1975 Mazda 13-B
2000 rpm
8.7 bhp
2000 rpm
26.3 bhp
3000 rpm
18.7 bhp
.9
.0 I.I 1.2 1.3
EQUIVALENCE RATIO
1.4
1.5
1.6
Figure 52.
Brake specific carbon monoxide emissions as a function
of equivalence ratio, Mazda 13-B engine, stock spark
timing.
-------
OD
U1
10
8
.c
-O
\
3 6
X
§ 5
CO
CD
0
.8
1975 Mazda 13'B
2000 rpm
26.3 bhp
Stock Spark
3000 rpm
I8.7bhp
2000 rpm
8.7 bhp
.9
1.0 I.I 1.2 1.3
EQUIVALENCE RATIO
1.5
1.6
Figure 53. Brake specific nitrogen oxide emissions as a function
of equivalence ratio, Mazda 13-B engine, stock spark
timing.
-------
F. EFFECT OF IGNITION TIMING ON FUEL CONSUMPTION AND EXHAUST
EMISSIONS - MAZDA 13-B ENGINE.
Two series of tests were conducted to ascertain the effect of
ignition timing on fuel consumption and emission characteristics.
The first series was conducted with a fixed air-fuel ratio, 16:1,
which was relatively close to the LBT* setting for the Mazda engine.
The second series was conducted with the stock mixture ratio as
delivered by the production carburetor. Three test conditions were
selected (identical to those used for the air-fuel ratio parametric
tests), 2000 rpm with 8.7 and 26.3 bhp and 3000 rpm, 18.7 bhp.
1. Air-Fuel Ratio of 16:1
The data for the first series of tests at 16:1 air-fuel ratio
are shown in Figs. 54-57. These illustrate brake specific fuel con-
sumption, hydrocarbon, carbon monoxide and nitrogen oxides emissions
as a function of ignition timing. The BSFC data, Fig. 54, shows
that £he thermal efficiency was relatively insensitive to ignition
timing near the MET spark timing. At the 2000 rpm, light load and
3000 rpm condition the BSFC increased greatly with significant
spark retardation. In general the trend of the results is consist-
ent with similar data for reciprocating engines although the Mazda
appears to be less sensitive to ignition advance.
The BSHC emissions, Fig. 55, decreased with spark retard at
all conditions. The major hydrocarbon reduction observed at 3000
rpm with spark retard was probably caused by enhanced thermal re-
action both, within the engine and in the exhaust collector due to
an increase in gas temperature caused by spark-retard. A slight
increase in misfiring with spark advance may also have been pre-
sent. To a limited extent these factors appear to be present at
other conditions. Additional factors probably include changes in
instantaneous combustion chamber surface/volume ratio which influ-
ence wall quenching and changes in combustion chamber pressure which
affect the leakage contribution to total hydrocarbon emissions.
Brake specific CO emissions in Fig. 56 indicate little varia-
tion with ignition timing. The lean air/fuel ratio used produced
the generally low level.
The BSNO emissions in Fig. 57 are interesting in comparison
to typical data observed in a spark-ignited reciprocating engine
which generally exhibits decreased NO emission with spark retard.
The rotary engine data showed that there was relatively little
influence of spark timing on NO emissions. However a small in-
crease was observed with spark Retard. This behavior is probably
caused by factors that increase maximum cycle temperature or reduce
the effectiveness of the NO decomposition reaction rather than
residence time, factors. Slightly delayed (from optimum for best
*Leanest mixture for best torque.
86
-------
oo
1.3
1.2
1.1
1.0
0.9
0.8
o
0.7
QQ
0.6
0.5
0.4
-20
1975 Mazda 13-B
-10
16 = 1 Air-Fuel Ratio
Spark Advance
J_
J_
2000 rpm 8.7 bhp
3000 rpm 18.7 bhp
2000rpm 26.3 bhp
-A A
0 10 20
SPARK ADVANCE (Deg. BTDC)
30
40
50
Figure 54. Brake specific fuel consumption as a function of
spark advance, Mazda 13-B engine, 16:1 air-fuel ratio.
-------
CD
CO
45
40
35
30
1 25
ex ^°
3 20
o
m 15
10
1975 Mazda 13-B
16=1 Air-Fuel Ratio
JMBT Spark Advance
_ 2000 rpm
26.3bhp
0
-20
2000 rpm 8.7 bhp
3000 rpm 18.7 bhp
-10
0 10 20 30
SPARK ADVANCE (Deg. BTDC)
40
50
Figure 55. Brake specific hydrocarbon emissions as a function of
spark advance, Mazda 13-B engine, 16:1 air-fuel ratio.
-------
700
600
1975 Mazda 13-B
16=1 Air-Fuel Ratio
Spark Advance
500
CO
I
a.
.c
_a
X,
en
O
O
CO
CD
400
300
200
100
-20
2000rpm 8.7 bhp
t—2000rpm 26.3 bhp
3000rpm 18.7 bhp
0 10 20 30
SPARK ADVANCE (Deg. BTDC)
40
50
Figure 56. Brake specific carbon monoxide emissions as a function
of spark advance, Mazda 13-B engine, 16:1 air-fuel ratio.
-------
o
11
10
9
8
5 7
i
Q.
I 6
O>
x 5
o
z
4
GQ ^
3
2
I
0
-20
1975 Mazda 13-B
16:1 Air-Fuel Ratio
JMBT Spark Advance
2000 rpm 26.3 bhp
+
18.7 bhp
—0
2000 rpm 8.7 bhp
-10
0 10 20
SPARK ADVANCE (Deg. BTDC)
30
40
50
Figure 57. Brake specific nitrogen oxide emissions as a function
of spark advance, Mazda 13-B engine, 16:1 air-fuel ratio.
-------
BSFC). high, temperature burning is suggested rather than additional
burning late, in the expansion or exhaust processes. An additional
factor in the slight NO increase is the increase, in engine air and
fuel flow required to maintain a constant horsepower with spark
retard. However, without a detailed study this analysis is specu-
lative. The relative position of the curves follow the expected
trend based on the BSFC results, The relatively poor combustion
at 2000 rpm light load is characterized by a reasonably low overall
level of NO emissions whereas the higher load 2000 rpm point is
considerably more efficient resulting in higher levels of NO
emissions. It was anticipated that spark retardation would signi-
ficantly reduce the NO emissions and these observations to the
contrary suggest spark retard to be an ineffective control strategy
for NO .
x
2. Stock Air-Fuel Ratio
The second series of tests was conducted at the same speeds
and loads but with the stock carburetion. The BSFC results are
shown in Fig. 58, and are similar to those, of the leaner mixture
studies. However, as expected the curves were displaced to a
greater BSFC level because of the richer mixtures.
The hydrocarbon emission data in Fig. 59 was also displaced
vertically from the data at a leaner mixture. In addition there
was an apparent diminished after-reaction of hydrocarbons with
spark retard that was particularly evident at the 3000 rpm test
condition with 16:1 air-fuel ratio. This arose because the mixture
was too rich to permit reaction in the exhausting chamber and ex-
haust collector (.limited 0.., availability) . Hence spark retardation
exhibited relatively little influence on the hydrocarbon emission
level.
Brake specific CO emissions are shown in Fig. 60 for the stock
carburetion. These data clearly demonstrate that there is essential-
ly no influence of spark timing on this pollutant. The major factor
was the higher CO arising from richer combustion as is evidence by
the vertical displacement of the curves in comparison to the data
at 16:1 air-fuel ratio.
The brake specific NO emissions are shown in Fig. 61. The
results appear to be consistent with established theory in that
they are generally displaced lower than for the data with 16:1
air-fuel ratio. The 2000 rpm, 8.7 hp data were not included because
they were in error and were impossible to recheck. The general
increase of the NO emissions with spark retardation is generally
unexpected as notea earlier. In any event the influence of spark
retardation appears to be far less signification than observed in
a reciprocating engine in which, spark retard effectively reduces
N0x.
91
-------
ID
1.3
1.2
1.1
-r 1.0
_c
I
Q.
u
Li-
en
GO
0.9
0.8
0.7
0.6
0.5
0.4
1975 Mozdo 13-B
_L
J_
_L
Stock Air-Fuel Rotio
JMBT Spark Advonce
2000 rpm 8.7 bhp
-20 -10
Figure 58.
3000 rpm 18.7 bhp
2000 rpm 26.3 bhp
A
0 10 20
SPARK ADVANCE ( Deg. BTDC )
30
40
50
Brake specific fuel consumption as a function of spark
advance, Mazda 13-B engine, stock air-fuel ratio.
-------
MD
OJ
40
35
30
.c
I
Q.
Jd
o 20
i
GO
15
10
5
01—
-20
1975 Mazda I3"B
Stock Air-Fuel Rotio
j MBT Spark Advance
2000 rpm 8.7 bhp
_L
_L
±
300O rpm 18.7 bhp
2000 rpm 26.3 bhp
-10
Figure 59.
0 10 20
SPARK ADVANCE (Deg. BTDC)
30
40
50
Brake specific hydrocarbon emissions as a function of
spark advance, Mazda 13-B engine, stock air-fuel ratio.
-------
VD
700
600
500
400
CL
.c
en
o 300
o
C/)
GO
200
100
O1—
-20
1975 Mozdo 13-B
-Q-
1
-10
Figure 60.
-o-
H~
Stock Air-Fuel Ratio
JMBT Spark Advance
t
-0-t
JL
2000 rpm 8.7 bhp
3000rpm 18.7 bhp
-0 ——0
2000 rpm 26.3 bhp
-A A
±
0 10 20 30
SPARK ADVANCE ( Deg. BTDC)
40
50
Brake specific carbon monoxide emissions as a function
of spark advance, Mazda 13-B engine, stock air-fuel
ratio.
-------
11
10
9
8
7
1975 Mazda I3'B
Stock Air-Fuel Ratio
|MBT Spark Advance
Q.
£1
.0
v. :
CP
X
O
2
m 3
2
1
0
_L
2000 rpm 26.3 bhp
3000 rpm 18.7 bhp
-20
-10
Figure 61
0 10 20
SPARK ADVANCE (Deg. BTDC)
30
40
50
Brake specific nitrogen oxide emissions as a function
of spark advance, Mazda 13-B engine, stock air-fuel
ratio.
-------
It is perhaps appropriate, here, to ma,ke. several general obser-
vations related to the. effect of ignition timing on combustion and
therefore fuel consumption and emission performance of the Mazda
engine. The. measured performance parameters were, generally found
to be far less sensitive, to variation in timing than in reciprocat-
ing engines. The. major reason behind this behavior is the shape
of chamber volume during the combustion phase of the cycle.
The mixture burns relatively slow (low turbulence) until the
combustion chamber approaches its minimum volume, TDC, position.
At this point the volume of the leading portion of the chamber
begins to increase rapidly and a commensurate shrinkage occurs
in the trailing to leading section. This increased mixture
motion greatly accelerates the burn rate. A relatively wide
variation in ignition timing during the "slow burn or low
turbulence" portion of the combustion cycle, (before TDC) would
therefore be expected to have little influence on the total
combustion process and consequently a limited effect on fuel con-
sumption and emissions.
G. SPATIAL DISTRIBUTION STUDIES OF THE HYDROCARBON AND CARBON
MONOXIDE EMISSIONS FROM THE MAZDA ENGINE
These tests were conducted to gain some inference as to the
internal engine sources of hydrocarbon emission in the rotary
engine. The unburned hydrocarbons were the focus of this funda-
mental emission source study because they have been the most dif-
ficult emission to control from this engine.
The exhaust hydrocarbon emissions of the rotary engine are a
result of several processes which are also important in the con-
ventional 4-stroke engine (14,15). These common processes, and how
they apply particularly to this engine are outlined below:
Wall Quenching - An important emission source results from the
well known but little understood wall quenching phenomena. As the
propagating flame front approaches the relatively cool chamber wall,
it is extinguished or quenched a finite distance from the wall.
This so-called "quench zone" retains raw fuel-air mixture, partially
burned and/or cracked fuel, carbon monoxide, and other less noxious
products. During the exhaust process some of these "quench zone"
products are scavenged from the chamber walls and released to the
atmosphere, and some remain to be partially burned on the succeed-
ing cycle.
This emission source is particularly significant in the rotary
engine for two basic reasons.
.This engine has a relatively high surface/volume ratio in the
combustion chamber and thus, a large, quench zone volume. This
contribution may be partially compensated by higher wall
temperatures resulting from the localization of combustion in
one housing quadrant. The schematic of the engine combustion
cycle shown in Fig. 62 illustrates the geometric factors in-
volved.
96
-------
IGNITION
EXPANSION
EXHAUST
Figure 62. Schematic diagram of rotary engine operating cycle.
.The trailing apex seal of a given rotating combustion
chamber "scrapes" the quench layer off the rotor housing
and into the peripheral exhaust port.
Crevice Effect or Gap Quenching - The second source arises
from another "quench" process. In this case, if two relatively
cool surfaces enclosing a fuel-air mixture are brought progressive-
ly closer together, a certain separation distance (quench distance)
will be reached beyond which a flame will not propagate. The mass
contained in this quench region or "dead volume" consists of a
high-percentage of fuel-air mixture and, therefore, has an extreme-
ly high concentration of unburned hydrocarbons. Where significant
crevices exist unburned hydrocarbon emissions are large.
The present configuration of the rotary engine has substantial
"dead volume" including:
.The volume radially beyond the side gas seal between the
rotor and side housing.
.The leading and trailing portions of the rotor surface at
those phases of the cycle where they approach close to the
rotor housing.
Blowby of Unburned Charge - Leakage of unburned mixture to the
atmosphere can also affect the emissions significantly. Even a
small percentage loss of unburned mixture can have a large effect
on the overall concentration of hydrocarbons in the exhaust.
A conventional four-stroke cycle automotive engine blowby loss
is to the crankcase and only a very small portion of unburned mixture
if any, is lost past the closed exhaust valve and during the valve
overlap period. The rotary engine blowby leakage loss to the crank-
case is zero with side intake ports since all side gas seal leakage
is automatically vented to the intake ports. The leakage of raw
97
-------
mixture past the leading apex seal, hpweyer, passes directly
through, the. exhaust port into the. exhaust system. .Further, if a
peripheral inlet port configuration were, to be. used, unburned
mixture could be. lost during the. unavoidably higher port overlap
period. Mixture, leakage, is probably the single most significant
source of hydrocarbon emissions in the rotary engine depending on
the design and conditions of the sealing system. High exhaust
oxygen concentration even with rich mixture combustion and no air
injection is evidence of a leakage problem.
Incomplete Combustion - Incomplete combustion may be a
source of hydrocarbon emissions, particulary with very rich or
lean fuel-air mixtures and/or poor mixing of the fuel and air.
This problem is intensified where exhaust residual is high as it
is at light-load conditions. This factor may not be of primary
importance because the rotary engine'operates reasonably well on
lean mixtures and has intense combustion chamber turbulence in the
leading portion of the chamber. Turbulence promotes complete com-
bustion. However, some evidence of incomplete combustion in the
more quiescent trailing section of the chamber has been provided
by high-speed movies.
Unfortunately it is impossible to "ride" with engine rotor and
quantify the true source of the hydrocarbon emissions. Instrumenta-
tion ar.cl other hardware complexities have prevented use of a timed
sampling valve in the engine combustion chamber to obtain spatially
resolved emission data. However, evidence of the hydrocarbon forma-
tion ir. the combustion chamber can be obtained from the exhaust
since it was believed that relatively little mixing occurs in the
engine after combustion and during expansion and exhaust. The
spatial distribution present during the hydrocarbon formation pro-
cesses may be qualitatively present in the exhaust as it leaves
the engine. Hence a program was developed to determine the spatial-
ly resolved hydrocarbon content in the exhaust. For this a special-
ly constructed exhaust pipe was built. Two different methods were
tried.
.Cox timed sampling valve
.Series of steady-state sampling probes
Unfortunately neither of these techniques permit an exact evaluation
of the hydrocarbon sources since several of the sources can affect
the same region of the engine chamber. For example it is impossible
to discriminate between hydrocarbons leaking past the leading apex
seal of one chamber into the trailing portion from incomplete flame
propagation in this region. However, knowledge of the region of
the chamber most responsible for the high emissions could assist
the engine designer.
1.. Sampling Valve Study
In this series of tests the Cox sampling valve, discussed in
the Instrumentation section, was used to withdraw a small fraction
98
-------
of the exhaust sample in a specially constructed exhaust system
near the exhaust port. Similar studies have been reported on
reciprocating engines CL6).. The sample was withdrawn at known
times with respect to the crankshaft position and thus the position
of the rotor. The data are presented as a function of crank angle
relative to the start of exhaust the process. Zero degrees is the
point where a given engine chamber begins the exhaust process. The
fluid mechanics in the engine exhaust system are substantially
different from a 4-stroke reciprocating engine. In the rotary
engine the exhaust flow never ceases for a substantial crank
interval as it does in a reciprocating engine. In fact, exhaust
flow is occurring from two engine chambers to (or back flow) the
exhaust system during the period the apex seal is traversing the
port. The total exhaust process for a given chamber occupies 360
crankshaft degrees in the rotary engine. The hydrocarbon concentra-
tion emissions are plotted as parts per million (ppm) of n-hexane
as a function crank angle. Carbon monoxide emissions were also
measured and plotted as a function of crank angle. Numerous test
conditions were evaluated, 2000 rpm, 8.8 and 14.2 bhp, 2500 rpm,
13 and 22.4 bhp and 3000 rpm, 18.7 and 34.3 bhp. Both hydrocarbon
and CO data corresponding to these test conditions are plotted in
Figs. 63-74. In several cases multiple sets of data were collected
to assess repeatability. Since it has been hypothesized that the
principal sources of unburned hydrocarbon emissions from the
rotary engine are: (a) exhaust of the quench layer volume associat-
ed with the large surface/volume ratio, (b) leakage of unburned
fuel past the leading apex seal of a given combustion chamber to
the preceding chamber which is in the exhaust mode, and (c) in-
complete flame propagation both in the trailing and leading regions
of a given chamber and since they may affect the same region of
the chamber, it was not possible to discriminate as to the distri-
bution of sources. The CO emissions are not significantly affected
by these factors other than that with the reduced combustion of the
fuel, a lower CO concentration should be evident. The local CO is
essentially a function of the local mixture ratio in the engine
chamber. A significant variation in CO data as a function of
sample position in the exhaust would have been indicative of mix-
ture stratification in the engine. This did not occur however.
The results of the study generally followed similar trends
for the various test conditions, with one exception at 2000 rpm,
8.8 bhp. All results were characterized by hydrocarbon peaks at
the beginning and/or end of a given "slug" of exhaust.
The hydrocarbon emissions while initially high rapidly de-
creased to a level less than 1000 ppm which in turn was followed by
a relatively rapid increase to the original level. Several
anomalies were, observed in the data. At 2000 rpm, 14.2 bhp, the
hydrocarbon increase near the tail of the exhaust sample began
near the center of the exhaust process CL80 crankangle degrees) /
and increased almost linearly to the maximum reading at 360°. At
2500 rpm, 13 bhp, the decrease, in hydrocarbons occurred later and
99
-------
3000
2000
0)
C
ra
K
0)
.c
a
o,
P.
o
n:
o
o
1000
2000 rpm
8.8 bhp
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (deg. )
360
Figure 63.
Exhaust hydrocarbon concentration as a function of crank
angle relative to the start of the exhaust process,
Mazda 13-B engine, 2000 rpm, 8.8 bhp. Data obtained
with a Cox timed sampling valve.
-------
4.0
3.0
— 2.0
O
O
2000 rpra
8.8 bhp
1.0
60 120 180 240 300
SAMPLE ANCLE RELATIVE TO START OF EXHAUST (cleg.)
Figure 64.
Carbon monoxide concentration as a function of crank
angle relative to the start of the exhaust process,
Mazda 13-B engine, 2000 rpm, 8.8 bhp. Data obtained
with a Cox timed sampling valve.
-------
3000
2000 rpm
14.2 bhp
2000
c
nl
X
-------
o
U)
4.0
3.0
2.0
o >-
o -°
1-1 fc?-
1.0
2000 rpm
14.2 bhp
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (deg.)
360
Figure 66.
Carbon monoxide concentration as a function of crank
angle relative to the start of the exhaust process,
Mazda 13-B engine, 2000 rpm, 14.2 bhp. Data obtained
with a Cox timed sampling valve.
-------
3000
2500 rpm
13 bhp
20UO
-------
4.0
3.0
O
Ul
8-
1.0
2500 rpm
13 blip
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (deg.)
360
Figure 68.
Carbon monoxide concentration as a function of crank
angle relative to the start of the exhaust process,
Mazda 13-B engine, 2500 rpm, 13 bhp. Data obtained
with a Cox timed sampling valve.
-------
3000
2500 rpm
22.4 bhp
2000
O
cr>
-------
4.0
3.0
2.0
O
-J
o ^
o-0
1.0
2500 rpm
22.4 bhp
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (deg.)
360
Figure 70.
Carbon monoxide concentration as a function of crank
angle relative to the start of the exhaust process,
Mazda 13-B engine, 2500 rpm, 22.4 bhp. Data obtained
with a Cox timed sampling valve.
-------
3000
3000 rpm
18.7 blip
2000
o
O
CX)
1000
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (deg.)
. 360
Figure 71.
Exhaust hydrocarbon concentration as a function of
crank angle relative to the start of the exhaust
process/ Mazda 13-B engine, 3000 rpm, 18.7 bhp. Data
obtained with a Cox timed sampling valve.
-------
-1.0
3000 rpra
18.7 bhp
3.0
2.0
O
BT3 D &
1.0
Figure 72
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (cleg. )
360
Carbon monoxide concentration as a function of crank
angle relative to the start of the exhaust process,
Mazda 13-B engine, 3000 rpm, 18.7 bhp. Data obtained
with a Cox timed sampling valve.
-------
3000
2000
« a
-1 £:
"o"
X
1000
3000 rpm
34.4 blip
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (cleg.)
360
Figure 73,
Exhaust hydrocarbon concentration as a function of
crank angle relative to the start of the exhaust process,
Mazda 13-B engine, 3000 rpm, 34.3 bhp. Data obtained
with a Cox timed sampling valve.
-------
4.0
3. 0
0
O
L.O
3000 rpm
34.3 bhp
60 120 180 240 300
SAMPLE ANGLE RELATIVE TO START OF EXHAUST (deg.)
360
Figure 74.
Carbon monoxide concentration as a function of crank
angle relative to the start of the exhaust process,
Mazda 13-B engine, 3000 rpm, 34.4 bhp. Data obtained
with a Cox timed sampling valve.
-------
the increase near the. end of the exhaust process also occurred
later. This same trend was even more, pronounced at 250Q rpm, 22.4
bhp. Eowever, it is evident that because, the exhaust sample was
taken about 2 inches from th.e exhaust port opening the curves may
have been displaced because of changes in position of the Cox valve
opening with respect to the front end of a given column of exhaust
gas. Hence the curves may be. nearly identical but displaced in
time with the peaks still corresponding to the point of exhaust
port opening for a given chamber. The data at 3000 rpm, at both
18.7, 34.4 bhp tended to contradict this "shift" theory because
the peaks occurred close to the zero position.
At 2000 rpm, 8.8 bhp, the hydrocarbons in the leading portion
of the exhaust sample Cwhich cannot be totally separated from the
trailing portion of the prior chamber !'s exhaust sample) were approxi-
mately 2300 ppm. This decreased rapidly to approximately 1300 ppm,
then increased during the middle of the exhaust process followed by
another decrease and subsequently increased back to the level cor-
responding to the leading position. It is evident from these data
that the hydrocarbon emissions are significant at both the leading
and/or trailing section of the chamber, and at the light load
condition there is some evidence of incomplete burning in the cen-
tral region of the chamber. This region corresponds to the position
of the cycle where the exhaust flow rate is relatively rapid, i.e.
the rotor chamber volume is changing at the highest rate near 180°
crank angle from the opening position of the exhaust port for a
given chamber. A slight increase in CO was observed in the region
where the hydrocarbon emissions were generally lower as expected
because of the more complete burning of the hydrocarbons in this
region and consequently greater potential for the formation of CO.
From the majority of the hydrocarbon data it is evident that
the gases that initially leave the chamber and/or those that leave
last are rich in hydrocarbons. This appears to confirm the hypothe-
sis that important sources of hydrocarbon emissions are due to leak-
age and incomplete burning mixture near the leading or trailing
portions of the chamber. In addition these trailing/leading section
results may be influenced by the physical scrapping of the quench
layer off the outer housing surface by the trailing apex seal with
the subsequent concentration of these gases in the latter portion
that leaves the chamber. Thus all three hypothesized factors may
give rise to the high concentrations in the leading and/or trailing
portions of the exhaust sample. Of course, it must be recognized
that because of the length of exhaust system from chamber to the
sample valves, some integration of sample or mixing has undoubtedly
occurred.
2. Manifold Sampling System Study
As an adjunct to the investigation of exhaust emission profile
in the exhaust system using the Cox fast response, sampling valve, a
separate study was conducted using a more qualitative technique.
112
-------
In this investigation, a special exhaust pipe was constructed
which contained steady-state sampling probes located at measured
intervals along the pipe.. An exhaust sample was withdrawn and
analyzed from each, sample probe... The hydrocarbon emission results
from the various probes were then plotted on a concentration basis
as a function of probe position. From these data it was possible
to infer whether the hydrocarbons produced in the chamber were
coming from local regions within the engine or were basically
related to a homogeneous process. If the hydrocarbon emissions
were related to some form of homogeneous reaction, one would
anticipate that the exhaust hydrocarbon profile would not vary as
a function of probe location. A variation from probe to probe
would infer that the hydrocarbon emission at least partially re-
sulted from local processes in the chamber. Similar studies have
been performed in reciprocating engines with good results (16).
In the rotary engine, however, the dynamics of the exhaust process
are considerably different. In a reciprocating engine the exhaust
slug is periodically stationary during the time when the exhaust
valve is closed whereas in the rotary engine the exhaust flow is
always dynamic (an exhaust process is always occurring). At a
constant sample rate a given probe will take in a greater propor-
tion of the sampled gas from a part of the exhaust "slug" that is
moving more slowly than from a fast moving segment. By placing
probes at intervals along the exhaust pipe as indicated above, it
may be possible to determine a qualitative measure of the distri-
bution of hydrocarbons in the exhaust. The reader is referred to
Ref. 18 which is a rather complete study performed by John White
at the University of Michigan. The referenced report contains an
analysis of the data and in addition develops a computer simula-
tion by which the profile of the exhaust can be determined as if
the probes were moving in phase with the exhaust gas (equivalent
to use of fast response sample valve). One value of this techni-
que is that the equipment requirements are considerably less
sophisticated than utilization of the sampling valve.
Sample data are plotted in Figs. 75-77. In Figure 75 hydro-
carbon concentration is plotted as a function of exhaust probe
position at 2000 rpm for both road load and 100% road load horse-
power. Similar data is plotted in Fig. 76 at 2500 rpm road load
and 170% road load. Clearly these data suggest that there is a
considerable variation in concentration as a function of position
in the exhaust "slug" and confirms the behavior, at least qualita-
tively, observed with the Cox sampling valve. The curves at the
different loads are shifted considerably horizontally because of
the different exhaust "slug" volumes at different loads. Figure 77
shows a similar plot for three, different engine speeds at road
load conditions. Again the significant variation in concentration
within the exhaust slug is evident. The data presented here are
a relatively small sample from the study that was performed, how-
ever, all of the data in both raw and reduced form, show clearly
that the exhaust composition varies considerably along the exhaust
"slug". These results confirm the suspicion that the hydrocarbons
113
-------
a.
a.
a:
h-
"Z.
UJ
o
o
o
o
X
1800
1600
1400
1200
1000
800
600
400
200
2000 RPM
160% Road Load
I I
I
I
I
I 23456789
«3in-»|*-3in>| SAMPLE PROBE REFERENCE NUMBER
Exhaust Port
Flange
10
Figure 75.
Hydrocarbon concentration as a function of exhaust
sample probe position in the exhaust pipe, Mazda 13-B
engine, 2000 rpm, manifold sampling system.
-------
1800
^ 1600
Ql 1400
5 1200
<* 1000
LU
O
O
O
O
X
800
600
400
200
0
2500 RPM
Rood Load x
x.
170% Rood Load
I
I 23456789
«-3in.*k3in.-*| SAMPLE PROBE REFERENCE NUMBER
Exhaust Port
Flange
10
Figure 76.
Hydrocarbon concentration as a function of exhaust
sample probe position in the exhaust pipe, Mazda 13-B
engine, 2500 rpm, manifold sampling system.
-------
cr\
1800
- 1600
JL 1400
§ 1200
£ 1000
H 800
LU
2 600
o
0 400
o
x 200
0
Rood Load
2000 RPM
3000 RPM
I
I
23456789
SAMPLE PROBE, REFERENCE NUMBER
10
Exhaust Port
Flange
Figure 77.
Exhaust hydrocarbon concentration as a function of
sample probe location in the exhaust system, Mazda 13-B
engine, manifold sampling system.
-------
are formed non-uniformly within the. engine chamber. The apex seal
region is probably the most significant contributor either because
of gas leakage to the exhausting chamber, "pile-'Up" of the 'quench
layer due to physical scraping of the seal on the chamber and/or
incomplete flame propagation.
H. CHAMBER PRESSURE STUDY - MAZDA ENGINE
Cylinder pressure records were taken in the Mazda engine at
both motoring and firing conditions. The purpose was to compare
the polytropic exponents with values for piston engines and other
rotary engines reported in the literature. From .such a comparison
it was hoped differences in combustion and heat transfer might be
inferred.
A study of Fig. 11 in Section III reveals that with a single
transducer the pressure can only be measured for 360 deg. of crank
rotation. The pressure for most of the compression and about half
of the expansion stroke may be obtained in this manner. The
important portion of combustion is thus covered. It should be
noted that the pressure observed near the end of the expansion
trace is grossly distorted, for the recorded value is at a changing
level between the relatively high pressure at 120° ate and the low
pressure early in the compression stroke.
Pressure records were obtained under the following conditions:
Firing Engine: Production carburetion and ignition timing
RPM Brake Horsepower
2500 22.4
2500 13.1
2000 14.0
2000 8.8
Motoring RPM Throttle
2000 closed
2000 part
2000 WOT
2500 closed
2500 part
2500 WOT
Results - Firing
Figures 78,79,80 show pressure-crank angle, pressure-volume
and log pressure versus log volume, respectively for the engine at
2500 rpm, 22.4 bhp. The pressure crank-angle, diagram clearly shows
the 360° degree "window" in which, pressures were measured as well
as the rapid drop in cylinder pressure near the end of the measure-
ment as the apex seal moves across the transducer sampling hole.
Similar information is seen in Fig. 79. Obviously insufficient
117
-------
300
Mozdo 13-B
2500 rpm, 22.4 bhp
Firing, I2°BTC Spork
CO
I
240
Figure 78.
180
120 60 0
CRANK ANGLE (deg.)
60
120
Cylinder pressure as a function of crankangle, Mazda
13-B, at 2500 rpm, 22.4 bhp.
-------
300
Mazda 13'B
2500 rpm, 22.4 bhp
Firing, I2°BTC Spark
20 30
VOLUME (in3)
Figure 79. Cylinder pressure as a function of volume, Mazda 13-B,
at 2500 rpm, 22.4 bhp.
-------
300
NJ
O
200-
o
'U>
Q.
LJ
C/)
If)
UJ
o:
Mozdo 13-B
2500rpm,22.4bhp
Firing, 12°BTC Spark
5 10
VOLUME (in3)
30 40 50
Figure 80.
Cylinder pressure as a function of volume, logarithmic
plot, Mazda 13-B, at 2500 rpm, 22.4 bhp.
-------
information exists to determine, cycle, work., .Figure 80 shows the
pressure-volume, record on logarithmic coordinates. The varying
nature of the signal suggests the. need to ascribe, several polytropic
exponents over the cycle. The beginning and end of the trace were
not analyzed for reasons discussed previously.
Figure 81 shows all four of the logarithmic pressure-volume
plots for the fired engine data. In each case., multiple exponents
appear warranted. Similar results were reported by Froede in Ref.
(18) for a KKM-250-5 engine.
In comparing data from Froede with our results some similari-
ties and differences are evident. A considerable amount of after-
burning is seen to exist. In the Mazda engine, an expansion exponent
of less than 0.63 persists until about 10 in3 (164 cc) cylinder
volume or 2.22 x the minimum volume. In the KKM-250-5 a similar
after-burning period exists. In the absence of such after-burning,
the expansion line at the start of expansion would be an extension
of the line later in expansion whose slope for the Mazda was 1.22-
1.23. Such an extended line is shown with dashes in Fig. 81 • Of
course all engines whether reciprocating or rotating have some
degree of afterburn due to the finite progression of combustion and
the re-association of dissociated species during expansion. However
this afterburning appears somewhat more extensive than that report-
ed by Nelson for an 8.7:1 compression ratio experimental piston
engine at part load in Fig. 14 of Ref. (19.) . In that case after-
burning ceased at an expansion volume equalling 1.8 x the minimum
volume. Moreover the after-burning in the Mazda was clearly greater
than that reported by Lancaster in Ref. (2QJ for a single cylinder
piston engine. In comparing with Froede's results on the KKM-250-5,
it is noted that the polytropic expansion exponent for the Mazda,
(ave. 1.22) is slightly lower than Froede1s result (ave. 1.27) if
his one value of 1.7 is ignored. It is also lower than Nelson's
average value of 1.29 and Lancaster's 1.30 for reciprocating engines.
No firm conclusions can be drawn from these small differences in
expansion exponents in light of the probable differences in mixture
ratio, load and measurement technique between the various researchers.
In general, a lower expansion exponent suggests more after-burning
whereas a higher value suggests greater heat transfer and leakage.
The compression exponents of this study ranged from 0.89 to
1.12. These values are quite low compared to the values of Froede
which were 1.2 or greater and those of Nelson and Lancaster which
were about 1.3. The characteristic of a lower exponent late in
compression is similar to the finding of Froede and suggests exces-
sive leakage and heat loss near top center piston position. How-
ever the problems appeared more extensive, in the Mazda engine than
the KKM-250-5 engine.. Typically the piston engine compression line
is well represented by a single polytropic exponent. Consequently,
the low and multiple valued compression exponent of this study is
distinctly different from that of the piston engine. Leakage of
unburned mixture past the. leading apex seal is thought to be a
121
-------
300
Log Pressure-Log Volume, Mozdo 13~B. Firing Conditions
300, ,
5 10
VOLUME (in3)
200
!i 100
01
(ft
20 30 40 50
60
40
20-
15
123
T 1 r
2500 rpm
I3.t bhp
5 10 20 30 10 50
VOLUME (in.3)
K)
to
300
200
° 100
U»
-3
LJ
ft-
60
en
40
20
15
\ 062 2000 rpm
Ox!025 I4-0 bhp
^-1.22
0.97-'
300
200
5 100
01
tn
Id
60
40
20
5 10
VOLUME (in3)
2O 30 40 50
\ 2000 rpm
\ 8.8 bhp
,063
1.22
1.12"
5 10
VOLUME (in3)
20 30 40 50
Figure 81.
Logarithmic plots of pressure as a function of volume,
fired Mazda 13-B engine.
-------
major source of unburned hydrocarbons in the. rotary engine and the
low observed compression exponent especially near top center
probably indicate such leakage.
Results - Motoring
Six pressure records were obtained under motoring engine con-
ditions without fuel at 2000 and 2500 rpm, various throttle open-
-ings. These are shown in Fig. 82. in viewing these curves it is
well to keep in mind that the expansion lines are lower than the
compression lines. This is a common observance in engines of all
types without combustion and is due to heat transfer and leakage.
As in the fired case, multiple exponents may be assigned. Lower
exponents near top center on compression reflect leakage and heat
transfer in that portion of the cycle. The larger exponents during
expansion reflect the same phenomena. In general the motoring
exponents are higher than the firing exponents. This results from
the absence of fuel since the polytropic exponent of a fuel-air
mixture is lower than that of pure air. Lancaster's results scale
to a compression exponent of 1.3 and an expansion exponent of 1.43
in a reciprocating engine. In general it is concluded that the
motoring exponents are lower on compression and higher on expansion
than a piston engine and this suggests greater leakage and heat
losses in the rotary engine. However because of the unique seal-
ing system of the rotary engine, one must be careful about drawing
conclusions on sealing in a non-fired test.
123
-------
Log Pressure -Log Volume, Mazda 13-B, Motoring
300,
3 60
(/>
I/)
E 40
a
2
20OOrpm
Closed Throtle
300
2000 rpm
Port Thfotle
9 ioo[
- [
20-
5 IO 20 3O 10 SO
VOLUME (in3)
200-
2 100
!
5 60
5 10 20 30 4O SO
VOLUME (in3]
2O 30 40 SO
3OO- -i
K)
.£>
9
-------
REFERENCES
1. Jones, C., The Curtiss-tfright Rotating Combustion Engines
Today, SAE Trans., v. 73, 1965
2. Personal communications with representative of the General
Motors Corporation.
3. 1975 Gas Mileage Guidefor New Car Buyers, U.S. Environmental
Protection Agency and Federal Energy Admin., 1974-75.
4. 1975 Mazda shop manual.
5. Papa, L. J., Gas Chromatography - Measuring Exhaust Hydrocarbons
Down to Parts Per Billion, SAE paper 670494, SAE Progress in
Technology, v. 14, 1971.
6. Oberdorfer, P. E., The Determination of Aldehydes in Automobile
Exhaust Gas, SAE paper 670123, SAE Progress in Technology Series,
v. 14, 1971.
7. Sampling Procedure for Determining Exhaust Aldehyde Emissions,
Internal Report, University of Michigan, W. E. Lay Automotive
Laboratory
8. Obert, E. F., Internal Combustion Engines, Intext Co., Scranton,
Pa., 1968, pg. 481.
9. Taylor, C. F., Taylor, E. S., The Internal Combustion Engine,
Intext Co., Scranton, Pa., 1970, pg. 411-342.
10. Cole, D. E., Jones, C., Reduction of Emissions from the Curtiss-
Wright Rotating Combustion Engine with an Exhaust Reactor, SAE
paper 700074, SAE Progress in Technology Series, v. 14, 1971.
11. Patterson, D. J., Kadlec, R. H., Kinetics of Oxidization and
Quenching of Combustibles in Exhaust Systems of Gasoline Engine,
Final report No. 310830-3-F, University of Michigan, 1972.
12. Jackson, M. W., Wiese W. M. and Wentworth, J. T., The Influence
of Air-Fuel Ratio, Spark Timing and Combustion Chamber Deposits
on Exhaust Hydrocarbon Emissions, SAE paper 486A, SAE Techni-
cal Progress Series, v. 6, 1964,
13, Hagen, D. F., Holiday, G, W., The Effects of Engine Operating
and Design Variables on Exhaust Emissions, SAE paper 486C, SAE
Technical Progress Series, v, 6", 1964,
14. Daniel, W. A,, Why Engine Variables Affect Exhaust Hydrocarbon
Emission? SAE paper 700108, SAE. Progress in Technology Series,
v, 14, 1971.
125
-------
15. Cole, D. E,, Emission Characteristics of the W.anke.1 Rotary
Engine, SME paper MM73-&54, 1973,.
16. Daniel, W. A., Wentworth., J, T. , Exhaust Gas Hydrocarbons -
Genesis and Exodus, SAE 4863, SAE Progress in Technology
Series, v. 6, 1964,
17. White, J. H., Investigation and Analysis of the Exhaust
Chamber Hydrocarbon Stratification in an Existing Mazda 13-B
Rotary Engine masters thesis, University of Michigan, Ann Arbor,
Mich., 1976.
18. Froede, Walter G., The NSU-Wankel Rotating Combustion Engine,
SAE paper 288A, January 1961.
19. Caris, D. F., Nelson, E. E., A New Look at High Compression
Engines, SAE paper 61A, June 1968.
20. Lancaster, D. R.,et al, Measurement and Analysis of Engine
Pressure Data, SAE paper 750026, February 1975.
126
------- |