THE U N I V E R S I T Y OF MI C H I G A N COLLEGE OF ENGINEERING Department of Mechanical Engineering Final Report ROTATING COMBUSTION CHAMBER ENGINE HYDROCARBON EMISSIONS David E. Cole Douglas Jr;Tin ORA Project 08070 under contract with: CURTISS-WRIGHT CORPORATION WRIGHT AERONAUTICAL DIVISION WOOD-RIDGE, NEW JERSEY administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR December 1967

4~ 5 s E q r

TABLE OF CONTENTS Page LIST OF FIGURES iii NOMENCLATURE vii I. INTRODUCTION 1 A. General 1 B. Test Program 2 II. SUMMARY OF RESULTS, CONCLUSIONS AND RECOMMENDATIONS 4 A. Summary 4 B. Conclusions 6 C. Recommendations 6 III. TEST EQUIPMENT AND INSTALLATION 8 A. Dynamometer Installation 8 B. Temperature Measurement 8 C. Flow Measurement 8 D. Speed Measurement 10 E. Fuel/Air Ratio Measurement 10 F. Exhaust Hydrocarbon Measurement 11 G. Propane-Air Mixing System 14 H. Air Injection System 15 IV. RESULTS AND ANALYSIS 18 A. Base Engine Without Exhaust Reactor 18 B. Base Engine with the Exhaust Reactor 26 C. RC Engine with Exhaust Reactor and Air Injection 53 D. Exhaust Manifold Reactor 57 E. Analysis of RC Engine Exhaust Emissions 60 APPENDIX A. COMMENTS ON THE RC ENGINE INSTALLATION 62 APPENDIX B. RC ENGINE (COMBUSTION) AIR FLOW MEASUREMENTS 63 APPENDIX C. ORSAT CALIBRATION CURVES 64 APPENDIX D. BECKMAN NDIR CALIBRATION CURVE 65 APPENDIX E. CONVERSION OF CONCENTRATION HYDROCARBON EMISSIONS TO MASS EMISSIONS PER UNIT TIME 66 BIBLIOGRAPHY 67 ii

LIST OF FIGURES Figure Page 1. RC engine dynamometer installation-front quartering view. 2. RC engine dynamometer installation-rear quartering view. 9 3. Instrumented RC engine showing the WAD exhaust reactor. 10 4. Exhaust gas hydrocarbon sampling and analyzing cart-eside view. 12 5. Exhaust gas hydrocarbon sampling and analyzing cart-interior view showing water trap, and diaphragm and vacuum pumps. 12 6. Exhaust gas hydrocarbon analysis system-schematic diagram. 13 7. Exhaust gas sampling probe. 14 8. Propane and air mixing system-schematic diagram. 15 9. RC engine exhaust ports showing the air injection tubes, and pressure and temperature instrumentation. 16 10. Baseline hydrocarbon emissions as a function of intake manifold vacuum —standard RC engine without WAD reactor. 17 11. Baseline hydrocarbon emissions from the exhaust of individual rotors and the combined exhaust-standard RC engine without WAD reactor. 20 12. The effect of spark advance on the hydrocarbon emissions — RC engine without WAD reactor. 21 13. Hydrocarbon emissions as a function of air/fuel ratio with regular gasoline at 2000 rpm and several intake manifold vacuums-RC engine without WAD reactor, MBT spark advance. 22 14. Hydrocarbon emissions as a function of air/fuel ratio with propane fuel at 2000 rpm and several intake manifold vacuumsRC engine without WAD reactor, MBT spark advance. 24 15. Hydrocarbon emissions as a function of equivalence ratio with both regular gasoline and propane fuel —RC engine without WAD reactor, MBT spark advance. 25 iii

LIST OF FIGURES (Continued) Figure Page 16. Effect of coolant temperature on hydrocarbon emissions-RC engine without reactor. 27 17. Steady state hydrocarbon emissions from a 1965 Chevrolet (283 CID) engine. 28 18. Baseline hydrocarbon emissions as a function of intake manifold vacuum-standard RC engine with WAD reactor, no air injection. 30 19. Baseline horsepower, torque and dynamometer speed characteristics-standard RC engine with WAD reactor, no air injection. 31 20. Exhaust back pressure in engine exhaust ports and at the exhaust manifold reactor outlet-RC engine with WAD reactor, no air injection. 32 21. Engine bil, water, and air temperatures at 2000 rpmstandard RC engine with WAD reactor, no air injection. 33 22. Exhaust port and reactor coret<jerituires at 2000 rpm — standard RC engine with WAD reactor, no air injection. 33 23. Engine oil, water, and air temperatures at 3000 rpmstandard RC engine with WAD reactor, no air injection. 34 24. Exhaust port and reactor core temperatures at 3000 rpmstandard RC engine with WAD reactor, no air injection. 34 25. Cold start, warm-up test —hydrocarbon emissions, and reactor core and engine cooling water temperatures as a function of time-RC engine with WAD reactor, no air injection. 37 26. Cold start, warm-up test-exhaust port, reactor outlet and reactor core temperatures as a function of time —RC engine with WAD reactor, no air injection. 38 27. RC engine performance at 2000 rpm with the fuel-air ratio adjusted for minimum emissions —MBT spark advance, WAD reactor, no air injection. 39 28. RC engine performance at 3000 rpm with the fuel-air ratio adjusted for minimum emissions —MBT spark advance, WAD reactor, no air injection. 40 iv

LIST OF FIGURES (Continued) Figure Page 29. Comparison of the RC engine performance data at 2000 rpm — minimum emission F/A ratio and standard F/A ratio-MBT spark advance, WAD reactor, no air injection. 42 30. Comparison of the RC engine performance data at 3000 rpmminimum emission F/A ratio and standard F/A ratio-MBT spark advance, WAD reactor, no air injection. 43 31. Mass rate of hydrocarbon emissions with the standard carburetor and the variable fuel/air ratio carburetor adjusted for minimum emissions-RC engine with WAD reactor, no air injection.44 32. Mass rate of hydrocarbon emissions as a function of brake horsepower with the standard carburetor and variable fuel/air ratio carburetor adjusted for minimum emissions-RC engine with WAD reactor, no air injection. 45 33. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 2000 rpm, 17 in. Hg.-RC engine with WAD reactor, no air injection. 47 34. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 2000 rpm, 10 in. Hg.- RC engine with WAD reactor, no air injection. 48 35. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 2000 rpm, 5 in. Hg. —RC engine with WAD reactor, no air injection. 49 36. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 3000 rpm, 20 in. Hg.- RC engine with WAD reactor, no air injection. 50 37. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 3000 rpm, 15 in. Hg.-RC engine with WAD reactor, no air injection. 51 38. Hydrocarbon emissions and reactor core temperature as a function of brake horsepower with both spark advance and F/A ratio optimized for minimum emissions-~RC engine with WAD reactor, no air injection. 52 39. The effect of exhaust port air injection on the hydrocarbon emissions; 2000 rpm, 15 and 20 in. Hg. —RC engine with WAD reactor. 54 v

LIST OF FIGURES (Concluded) Figure Page 40. The effect of exhaust port air injection on the hydrocarbon emissions; 2000 rpm, 5 and 10 in. Hg. -RC engine with WAD reactor. 55 41. The effect of exhaust port air injection on the hydrocarbon emissions; 3000 rpm, 15 and 20 in. Hg.-RC engine with WAD reactor. 56 42. WAD reactor assembly-exhaust port side. 58 43. WAD reactor disassembly-after 97-1/2 hours of testing. 58 44. WAD reactor inner core-after 97-1/2 hours of testing. 59 45. WAD reactor inner core-end view showing heat distortion after 97-1/2 hours of testing. 59 46. RC engine mass rate of air flow as a function of intake manifold vacuum. 63 47. Exhaust gas composition as a function of air-fuel ratio. 64 48. Beckman calibration-n-hexane in nitrogen. 65 vi

NOMENCLATURE A/F- air/fuel ratio CFM - cubic feet per minute CID - cubic inches displacement CO - carbon monoxide F/A - Fuel/air ratio ft - feet OF - degrees Fahrenheit G.E. - General Electric HC - hydrocarbons Hg - Mercury (in. Hg. —measurement of pressure) HP - horsepower Hr. - hour in. - inches lbs. - pounds MT - Total mass flow rate of fuel and air MA - Mass air flow rate MBT - mean best torque mph - miles per hour mv - millivolts NDIR - non-dispersive infrared n-hexane - normal hexane, C6H14 No. - number vii

NOMENCLATURE (Concluded) N2 - nitrogen ppm - parts per million (volume) psi - pounds per square inch RC - Rotating combustion chamber rpm - revolutions per minute (crankshaft) sec. - second WAD - Wright Aeronautical Division - equivalence ratio viii

I. INTRODUCTION A. GENERAL The purpose of this research project was to investigate the exhaust hydrocarbon emissions from the Curtiss-Wright RC2-60-U5 engine. The results of a year and four month test program demonstrated that this engine can achieve a steady-state emission level of less than 75 ppm (n-hexane) with the proper control of engine operating variables and the addition of an exhaust reaction device. Thus, it is expected that the RC engine should conceivably be able to meet and surpass all existing Federal exhaust hydrocarbon emission standards. In the past few years with the rapid growth of atmospheric air pollution problems on both the local and national levels, the internal combustion engine has received increasing criticism for its suspected contribution to the problem. In some special cases, notably Los Angeles with its photochemical smog, the automotive spark-ignition engine has been blamed for the greatest proportion of the problem. California was the first state to legislate for control of engine exhaust emissions, doing such in 1961. They developed a seven-mode chassis dynamometer driving schedule designed to simulate typical driving conditions and specified a nondispersive infared technique for exhaust emission analysis. In 1967, California modified their earlier specifications and established a composite average maximum standard for unburned hydrocarbons and carbon monoxide emissions of 275 ppm and 1.5%, respectively, for reciprocating engines with greater than 140 CID. The federal government following California's direction has adopted a "Clean Air Act"l covering the continental United States. Currently the Federal standards, which also use the California simulated driving cycle, are also established at 275 ppm as n-hexane (HC) and 1.5% CO. In 1970 this will be lowered to 170 ppm as n-hexane (HC) and 1% CO. However, there is a distinct possibility that the 1970 standards will be modified and put on a mass flow basis (pounds per hour). A standard of this nature could penalize the large displacement engines with -heir greater air and fuel consumption. Considering the great national interest in this problem it is apparent that a new parameter has been thrust at the engine designer. In the past, performance and economy were of utmost importance, and because of their often conflicting nature many compromises were in order. Today a third parameter, exhaust emissions, has assumed equal if not greater importance. This additional consideration has greatly complicated the already formidable task facing our highway transportation industry. Our spark-ignition engines must at least meet the Federal standards before thought can be given to production for the consumer market. 1

Looking into the future it is conceivable that the eventual emissions standards will be as low as 25 ppm as n-hexane (HC) and.25% CO. This most certainly will provide a monumental challenge to our powerplant designers and could signal an end to our fossil fuel prime movers. However, they most certainly will be in the picture for many years to come. B. TEST PROGRAM The RC engine was installed on a dynamometer test stand in the Automotive Engineering Laboratory. This was done in lieu of a chassis dynamometer vehicular test installation for several important reasons: 1. It is difficult to precisely control the engine operating variables on the chassis dynamometer. 2. More convenient access is provided to the engine for development purposes. 3. Steady-state data is easier to obtain and was believed to be sufficient for development work. The exhaust sampling and analyzing system which was designed and built in our laboratory, was based on the Beckman non-dispersive infared analyzer to provide compatibility with current instrumentation practice in the automotive industry and government laboratories. To minimize cost and decrease the installation time, the system was designed to analyze only one class of exhaust gas pollutants, the unburned hydrocarbons. Another consideration in this decision was that it was believed that the carbon monoxide emissions would follow the same general pattern as the hydrocarbon emissions and would therefore not be of significant value. The actual test schedule was divided into two separate phases: 1. Tests of the basic engine with the standard exhaust manifold. 2. Tests of the engine with the Curtiss-Wright exhaust reactor manifold. Prior to development work in each phase, an important series of tests was ~conducted on the engine to obtain the baseline emission characteristics; these to be used for comparative purposes. This data was examined statistically to gain an understanding of data reproducibility. In both phases of the test program certain engine operating parameters, which had been found in previous research2 on reciprocating engines to have significant effect on emissions, were modified to measure their effect on the RC engine hydrocarbon emissions. These variables were fuel/air ratio, sparkadvance, and coolant temperature. The engine with the WAD exhaust reactor was also tested with an exhaust port air injection system to measure air injection effectiveness in reducing emissions. 2

Several other tests were run incidentally to the primary test program: 1, The RC engine was operated on a propane-air mixture to measure the relative effectiveness of the standard engine's carburetor and induction system in obtaining a homogeneous gasoline and air mixture. 2Q A static leak test was conducted to obtain an indication of the raw mixture leakage past the leading apex seal and its contribution to the overall hydrocarbon emissions. 3 A 1965 Chevrolet 283 CID engine was tested to obtain a steady-state emission comparison with the RC engine. All testing was done under steady-state conditions at one of four engine speeds —1000, 1500, 2000, or 3000 rpm. 5

II. SUMMARY OF RESULTS, CONCLUSIONS AND RECOMMENDATIONS A. SUMMARY 1. The exhaust hydrocarbon emissions from the baseline engine without the WAD reactor or air injection are summarized in the following table: AVERAGE HYDROCARBON CONCENTRATIONS (ppm n-hexane) d Intake Manifold Vacuums (in.Hg.) Speed ()20 15 10 5 1000 1100 950 2000 450 400 300 250 3000 225 175 150 140 Note: these data are with 11365-129 Champion Spark Plugs. It was observed that generally the emissions decreased with both an increase in speed and load. 2. Approximately a 15% decrease in emissions was obtained with a 10~ retardation of the spark from the MBT setting. 3. Variation in air-fuel ratio had a very significant effect on the emissions. A 25% reduction in the hydrocarbon emissions was observed with optimization (for minimum emissions) of the fuel/air ratio in the engine without reactor and from 75% to 90% reduction in the engine with the exhaust reactor and no air injection. 4. The hydrocarbon emissions decreased slightly with an increase in engine coolant temperature. 5. Operation of the engine with a propane and air mixture showed that the minimum emission equivalence ratio was lower than that with operation on gasoline. No direct comparison of exhaust emissions could be made between the two fuels because the exhaust gases have different compositions. 6. The WAD exhaust manifold reactor, alone, added to the standard engine accounted for approximately a 50% reduction in the hydrocarbon emissions. 7. In the engine at 2000 and 3000 rpm with the exhaust reactor and no air injection,the hydrocarbon emissions ranged from 40 to 70 ppm and less than.06 lb/hr for the minimum emission fuel/air ratio. The equivalent results with the minimum emission fuel/air ratio and spark advance combination were 30 to 50 ppm and less than.05 lb/hr. 4

8. The exhaust back pressure attained a value of over 12 in. Hg. at 3000 rpm and high load conditions with the reactor on the engine. 9. Temperature increases of 100 to 200~F were measured from the engine exhaust port to the internal core of the reactor. 10. At an idle condition (1000 rpm, no load), the exhaust reactor attained 900 of its steady-state effectiveness (140 ppm n-hexane), 2 min.after a cold start. The emissions were extremely sensitive to the idle screw adjustment; thus, fuel/air ratio. 11. In the standard engine with the exhaust reactor and air injection it was not possible to decrease the exhaust hydrocarbon emissions below 70 ppm at 2000 rpm and 50 ppm at 3000 rpm by varying the exhaust port air injection rate. 12. Generally, minimum emissions from the exhaust reactor-air injection combination were obtained with an injected air rate of 10 to 15o of the engine air at 2000 rpm and 5 to 10% at 3000 rpm. 13. Several signs of wear and deterioration were observed in the WAD exhaust reactor: a. Erosion of the baffle plates in the inner core. b. Thermal stress cracking on the middle insulating cylinder at the welds. c. Warpage of the inner core about its longitudinal axis. 14. During the study the engine was run approximately 250 hours with no failures other than of one spark plug. The spark plugs, however, needed cleaning periodically (every 10 hours). 15. The WAD exhaust manifold reactor was tested for approximately 100 hours. 16. Oil consumption averaged 15-20 hours per quart, or in terms of mileage, 700-1100 miles per quart. 5

B0 CONCLUSIONS 1o The steady-state hydrocarbon emissions of the base RC engine are high, substantially above the road load emissions of an engine capable of meeting the Federal exhaust emission standards. 2o Fuel/air ratio is the engine variable which has the greatest single effect on emissions. Therefore, development of a low emission carburetion system should occupy a position of primary importance in the RC engine research program. 3~ It is imperative to utilize a retarded spark for operation at high intake manifold vacuums which correspond to both idle and deceleration conditions' 4, The mixing of the fuel and air and/or the distribution of the mixture to the rotors could be improved as demonstrated by the propane study. 5, An exhaust manifold reactor of the type used in this study appears to hold the greatest promise for successful exhaust emission control. 6, The use of the exhaust reactor without air injection combined with a fuel/air ratio adjustment for minimum emissions provides very efficient emission control, at least under steady-state conditions. This combination demonstrates the greatest promise for a marketable package when compared to an automotive reciprocating engine which may require both air injection and rich mixture ratios to obtain equivalent emission control, 70 The use of air injection does not appear to improve the emissions over those from the engine-reactor combination with the fuel/air ratio adjusted for minimum emissions0 8. In all of the tests the exhaust port temperatures were high enough to initiate a self-sustaining reaction in the exhaust manifold reactor0 Co RECOMMENDATIONS lo The induction system should be studied extensively to gain improvement in rotor-to-rotor fuel-air distribution and mixture homogeneityo Possibly, increased manifold heat, a dual manifold system such as used on the new Volvo or fuel injection would improve the situation, 2. A number of combustion chamber modifications should be considered to minimize regions of fuel-air mixture through which the flame might not propagateo These could include a decrease of the dead volume above the rotor side seals and between the rotor and rotor side housing, and relieving the.rotor surface near the leading and trailing apex seals, 6

3. The use of multiple ignition sources should be explored, perhaps with phased firing. 4. The size and shape of the WAD exhaust reactor manifold should be modified to obtain an optimum configuration. The current design, however, appeared to be extremely effective. 5. The use of nonmetallic materials should be explored for reactor construction. 6. The exhaust ports could be redesigned to provide for an increased portion of the "after engine" reaction zone within the ports themselves. This is an approach currently being followed in the automotive industry. 7. It would be desirable to measure CO under steady-state conditions to verify the hypothesis that its concentration is directly related to the unburned hydrocarbons. 8. A chromatographic analysis,,should be made of the RC engine exhaust to establish its composition. This may become an increasingly important factor with respect to the emission control specifications of the Federal Government if, as has been suggested, the standards begin to incorporate regulations of the individual constituents of the exhaust on a reactivity basis. 9. Serious consideration should be given to making a feasibility study of a hybrid combination of the RC engine and an electric generator as a possible powerplant for the small car, urban market. 10. The starter motor capacity should be increased to permit easier starting with a decreased current draw on the battery. 7

IIIo TEST EQUIPMENT AND INSTALLATION A- DYNAMOMETER INSTALLATION The RC engine was installed initially in test cell 242 of the Automotive Engineering Laboratory at The University of Michigan. It was connected to a Mid-West 300 hp, eddy current dynamometer through a 1964 Dodge Dart automatic transmission, which was used to minimize a torsional vibration problem that could occur with a more conventional 3dynamometer coupling. The Curtiss-Wright Preliminary Data Compilation Booklet- was used, with several modifications, as a guide for installation of the engine. The test installation is shown in FigSo 1, 2, and 3. Note particularly Fig. 3 which shows the engine with the exhaust reactor installedo Pertinent information regarding the dynamometer, the control instrumentation and comments on the installation are given in Appendix A. One serious deficiency of the Mid-West dynamometer was the lack of an amplidyne speed control system. This required the dynamometer operator to control both the engine load and speed to maintain a given test condition which proved to be very time consuming both from the standpoint of establishing the test condition and maintaining it. Consequently, when a dynamometer (G.E. 125 hp) with speed control became available, the RC engine was moved into test cell 244 at the Automotive laboratory. It was believed that the time saved in establishing and maintaining the test points fully justified the switch, B. TEMPERATURE MEASUREMENT All temperatures of less than 1000~F, with the exception of the rotor housing temperature, were measured with copper-constantan thermocouples used in conjunction with a Brown continuous indicating potentiometer. The rotor housing temperature was obtained with iron-constantan thermocouples and indicated on a Leeds and Northrup potentiometer, All high temperatures, those above 1000~F, were measured with the aid of chromel-alumel thermocouples used with either the Brown potentiometer or a second Leeds and Northrup potentiometero Co FLOW MEASUREMENT Engine air-flow data was required for several parts of the experimental work. It was measured using a specially'cconstructed air cart with roundedapproach circular orifices. The mass rate of air-flow was measured at the three engine speeds (1000,' 2000, and 3000 rpm) as a function of intake manifold vacuum. Test results are 8

Fig. 1. RC engine dcynamrometer installation-front quartering view. Fig. 2. RC engine dynamometer installation —rear quartering view. 9

Fig. 3. Instrumented RC engine showing the WAD exhaust reactor. plotted in Appendix B. As expected the flow rate was almost linear with respect to manifold vacuum at each speed. Thus, knowing the manifold vacuum at any of the three speeds it was possible to obtain the approximate air rate. Brooks rotometers were utilized in measuring the propane flow rate in the carburetor mixing and distribution studies, and the air injection rate in the exhaust reactor air-injection experiments. These units were calibrated individually with a positive displacement air bell. D. SPEED MEASUREMENT Both the RC engine crankshaft speed and the dynamometer speed were measured with Hewlett-Packard electronic counters. It was necessary to measure both speeds because of transmission slippage which is an inherent property of conventional automatic transmissions. E. FUEL/AIR RATIO MEASUREMENT All measurements of gasoline fuel/air ratio were made with the well known Orsat technique which involves measurement of the volume percentage of carbon dioxide (C02), carbon monoxide (CO), and oxygen (02) in the exhaust gas. The calibration curve of fuel/air ratio as a function of exhaust gas composition for gasoline is shown in Appendix C. 10

F. EXHAUST HYDROCARBON MEASUREMENT Perhaps the single most important measurement made in this research program was the concentration of unburned hydrocarbons'in the exhaust gas. To accomplish this end an exhaust hydrocarbon sampling and analyzing system was designed with the aid of Refs. 5-9 and built into a large, fully enclosed bench. The complete installation is shown in the photographs of Figs. 4 and 5. Note: Beckman NDIR analyzer is not shown in Fig. 5. The system is shown schematically in Fig. 6. The normal path of the exhaust gas sample is shown by the heavy black line. A viton-diaphragm, Air Shield Pump was used to draw the sample from the specially constructed stainless steel exhaust pipe probe which is shown in Fig. 7, into an ice bath which was used to condense most of the water vapor from the exhaust sample. The condensation step was necessary because the analyzer used was responsive to water and would thus given an erroneous hydrocarbon concentration measurement if the water vapor was not removed. Following the trap, a 2-in. paper element filter made by Gelman Instrument Company removed particulate matter from the sample. The flow rate through the system was regulated by E Hoke needle valve and measured with a Fischer and Porter rotameter. A Beckman model 31510 nondispersive infrared (NDIR) analyzer was used to measure the hydrocarbon concentration in the exhaust sample. Initially the Beckman instrument was equipped with a 5-1/4 in. sample cell which provided accurate hydrocarbon measurements in the range from 100 to 1500 parts per million (ppm). Due to consistently low hydrocarbon readings in the latter stages of the test program the 5-1/4 in. sample cell was replaced with a much more sensitive 13-1/2 in. cell which permitted accurate measurement of hydrocarbon concentration in the range from 20-500 ppm. Stainless steel tubing was used from the exhaust probe to the paper filter to prevent corrosion and contamination by reactive material in the high temperature exhaust gas. Copper tubing was used in the remainder of the system. A stainless steel sample tank was incorporated in the system to collect an exhaust gas sample which could then be analyzed by the Perkin-Elmer flame ionization analyzer available in the Combustion Laboratory of the Automotive Laboratory. This provided a periodic check on the continuous sampling Beckman analyzer. The sample tank and copper manifolding were evacuated by a Cenco vacuum pump prior to obtaining a sample. To insure accurate analyses with the Beckman instrument periodic calibration was required. Six different concentrations of normal hexane* (n-hexane) in nitrogen from 25 to 1200 ppm were used as the standard reference gases. These gases were purchased from Matheson in No. 2 high pressure gas cylinders. A sample calibration curve for the Beckman analyzer is shown in Appendix D. Nitrogen was used to purge the system after sampling of the exhaust and to establish a zero reference in the Beckman analyzer. *Normal hexane (n-hexane-C ^H ) is specified by current government standards as being representative of the average composition of the unburned hydrocarbons in the exhaust gas and therefore, is used as the standard reference gas. 11

Fig. 4. Exhaust gas hydrocarbon sampling and analyzing cart —side view. iiiiiii~~~~~~~~~iiii?!iii~ ~ ~ ~ ~ ~~:ii-iiiiiil l~iii~iiiii ii.........................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.:.......:ii:iii:...........ii:~ii ~....... -i i ~::::i::i::i::i.......i...................i..i......?!iiiiiiii~~~~~~~~~~~iiiiiii~~~~~~~~i!!iii!!iii~~~~~~~~~~~~iiii~~~~~~i~~iiiii!?iiiiiiiiiiiii~iiiiiiiiiijiii~.... 5. Exhaust g.s h ydrocarbon......... a.d a n alyzing.....t.....in view.....ing water trap- an d d i aphragm............ pum 12~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~t................................................................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:....................................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ii..........................................~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~ii...........................................................................~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ii..............................................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.................................... ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~:................................:.:........................~ ~ ~ ~ ~ ~ ~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Li................................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iiiiii...................................................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i~..............................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~t..................................~~~~ ~~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:i....................................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-l Fig. 5. Exhaust gas hydrocarbon sampling and analyz......................... view showing water trap, and diaphragm and vaccum pumps.~~~~~~~~~~~......................................................

All tubing /4" copper unless 4 STubing F\ROM PROBE (Altbn'/cop uls) 2" FILTER / IN EXHAUST otherwise specified MANIFOLD (/4 Nylon WATER Tubing TRA 200 ppm VACUUM PUMP 600 ppm ^ /1 ITR2-WAY VA L00 I WVALVE 1000 ppm SAMPLE CALIBRATION \ L TANK GASES NITROGEN FLO CONTRO n- hexane in N2 cell DIAPHRAGM PUMP 6- TOGGLE VALVE VENT BECKMAN NDIR MODEL 315 ANALYZER ROTAMETER Fig. 6. Exhaust gas hydrocarbon analysis system-schematic diagram. 13

1/4 1 Stainless steel tubing x.035 "wall thickness 1/16 Dia.-4 holes through 2 % C \ Fin -5- A8 [\^8l \). l, -'.Stainless steel arc 3 1 ~ l/ I weld - /81" radius Fig. 7. Exhaust gas sampling probe. The output of the Beckman analyzer, which was proportional to the hydrocarbon concentration, was read directly from a microampere meter on the amplifier section of the analyzer. This was done in lieu of using a pen recorder, such as Texas Instruments servo/riter II because of the recorder's initial expense ($1,000-$1,500) and more significantly, the long time delay for delivery (6-9 months). G. PROPANE-AIR MIXING SYSTEM A propane-air carburetion and mixing system was designedlll12'13 and built to furnish a homogeneous fuel/air mixture to the engine. The engine operation with the gasoline-air mixture furnished by the standard carburetor could then be compared to that with the propane system to obtain a measure of the standard systems effectiveness. A schematic diagram of the apparatus is shown in Fig. 8. Important components of the system are labeled. A Century gaseous fuel carburetor was used to obtain proper metering of the propane into the air. The fuel/air ratio was varied by changing the fuel supply pressure. 14

E XPLOSIVE 2' MEMBRANE FLAME rARRESTOR, SURGE TANK 2.5 CU. FT 2" WATER PIPE MIXING COLUMN A IR FILTER 4 A FLAME, PROPANE ARRESTOR CARBURETOR r - 1 THROTTLE ~ JL^ ^ AIR IN r INTAKE MANIFOLD PROPANE IN Fig. 8. Propane and air mixing system-schematic diagram. H. AIR INJECTION SYSTEM For part of the experimental work on the engine-exhaust reactor combination 1K4 an exhaust port air injection system was required. A rather simple system was devised which used shop air as the high pressure supply. The injection pressure was controlled with a conventional 0-50 psi regulator and the flow rate measured with a Brooks rotometer which was calibrated in CFM. A flow balanced, air manifold was constructed which proportioned equally the air to the two ports. The position of the air injection nozzles and also the bayonent type exhaust thermocouples in the exhaust ports is shown in Fig. 9. In this particular figure the air injection nozzles are shown at the right side of each port. All air injection testing was done, however, with the air nozzles in the central position, not in the location shown in the photograph. 15

Fig. 9. RC engine exhaust ports showing the air injection tubes, and pressure and temperature instrumentation. Each injection nozzle was made from stainless steel tubing. Air was injected from eight equally spaced holes in the tubing wall which provided good mixing of the air with the hot exhaust gases. 16

IV. RESULTS AND ANALYSIS The results of the experimental work are presented and discussed in the chronological order that they were performed. A. Base Engine without the WAD exhaust -reactor or air injection. B. Engine with the WAD reactor and without air injection. - C. Engine with the WAD reactor and air injection. -- Also included in this section of the report is: D. Discussion of the exhaust reactor system. E. General analysis of the sources of exhaust emissions from the RC engine. A. BASE ENGINE WITHOUT EXHAUST REACTOR The initial and one of the most important phases of testing was the measurement of the base-line emission characteristics of the "as received" RC engine. This was necessary to firmly establish a basis of comparison for all development work on the engine to minimize the exhaust hydrocarbon emissions. The results of this study are shown in Fig. 10 where the hydrocarbon emissions measured as parts per million (ppm)of n-hexane are plotted as a function of the intake manifold vacuum (in. Hg.) at three engine speeds (1,000, 2,000, and 3,000 rpm)o Several important conclusions can be drawn from this data: 1o The hydrocarbon emissions decrease with an increase in speed. 20 The emissions decrease with an increase in load. At 1000 rpm, which closely corresponds to an idle condition, a sharp increase in emissions was observed at intake manifold vacuums greater than 19 in. Hg. This undoubtedly was caused by misfiring brought on by excessive dilution of the intake charge with exhaust residual. In reciprocating engines this same effect is observed but it generally occurs at a higher intake manifold vacuum. An interesting feature of the 2000 and 3000 rpm curves was the apparent effect of spark-plug heat on the emissions. At both speeds a hotter plug (ChampionS365-129) resulted in a slight decrease in emissions. This may have resulted from poorer combustion initiation with the cold plugs (Champion f365-127) caused by cooler electrodes or a deposit formation on the electrodes. It was necessary to clean the cooler plugs every 3 or 4 hours to maintain smooth engine operation. A full range of engine loads was not investigated at 1000 rpm because of a conflict with the automatic transmission shift points.'r

130012001200 1000 RPM U I1008001 65-12S 0 -#3655 129 HAMPION SPARK PLUGS I a0Q: 2000 RPM: 900m 00- Z 3000 RPM 2520 15 10 5 INTAKE MANIFOLD VACUUM (inches Hg) 0 - Fig. 10. Baseline hydrocarbon emissions as a function of intake manifold vacuum- standard RC engine without WAD reactor. 18 18

The concentation emissions from the individual rotors and the average total of both rotors for the base engine are shown in Fig. 11 for 2000 and 3000 rpm. Notice that there is a significant spread, up to 50 ppm, in the emissions from the two rotors. At 2000 rpm the hydrocarbon concentrations from the two rotors cross at 13 in. Hg. intake manifold vacuum with the front rotor (#1) initially having higher emissions and then,at a load greater than 13 in. Hg;.lower emissions. The total exhaust was sampled at a point about 3-1/2 ft. downstream from where the exhaust from the individual rotors was sampled. Generally the hydrocarbon concentration in the total exhaust was approximately the average of that, from the two rotors considered individually. There was one exception though, at 2000 rpm and 12 in. Hg, gage pres'sure. Here the total was somewhat lower which was probably caused by slight oxidation of unburned hydrocarbones in the exhaust system between the sampling probe locations. Figure 12 shows the influence of spark-advance on the hydrocarbon emissions in the standard engine at 2000 rpm and 10, 15, and 20 in. Hg. manifold vacuum. The emissions were found to decrease with retardation of the spark which is consistent with the findings in reciprocating engine studies. The retardation generally leads to somewhat slower combustion and higher average temperatures-both of which promote more complete oxidation of the unburned reactants. The results are shown as a band to cover the range of emission values observed in the testing. This practice was followed in most of the following figures where it could be accomplished without excessive confusion. It must be noted that all experimental data is not absolutely reproducible and thus, a spread of data occurs at every test point. In the next series of tests a variable fuel/air ratio carburetor was installed on the RC engine. This carburetor used an external control on the enrichment valve which permitted the main metering systems' fuel jet area to be varied by the operator. The results are shown in Fig. 13 for the engine- at 2000 rpm and 10, 15 and 20 in. Hg manifold vacuum. Variation of the air/fuel ratio had a very significant effect on the hydrocarbon emissions. Approximately a 25% decrease in hydrocarbon emissions was observed with the minimum emission air/fuel ratio. The minimum emission air/ fuel ratio was approximately 2-1/2 ratios leaner than stoichiometric at 10 and 15 in Hg. manifold vacuum and 1/2 ratio leaner at 20 Hg. manifold vacuum. The curves appear to have the same characteristic shape as those from automotive type reciprocating engines. One point must be made with respect to the use of the air/fuel ratio data as a measurement of the absolute air/ fuel ratio furnished to the engine. The Orsat technique, which was use'd to measure air/fuel ratio in the research program, is not considered to be an accurate quantitative device. It is, however, adequate for relative measurements of air/fuel ratio. 19

700 A Anti-Drive (*) Rotor ^600 -lo Drive (#2) Rotor ~~~~~600 0^~0o Total (both Rotors) 0 500 C z a o Rotor I Q 400 2000 RPM _ Rotor 2 Iz^~~~~~~~ Tootlr 0 03 O2 0 Rotor 2 0 0 0 A W seaL ~ ~ ~ ~~~~- 0'-0o Total 4 200 0m~ 3)000 RPM >-e~~~~~, ~~~Rotor I 100 0 20 15" " 25 sINTAKE MANIFOLD VACUUM ( nches Hg) ions from the exhaust of individual Fig.l Baseline hydrocarbon emissios fro te without rea ctor rotors and the combined exhauststar RCe rat. 20

600 2000 RPM ENGINE SPEED 500 20" Hg E 1400 z 0 15"Hg 300 zo 0: o 100 0 I 00 STANDARD SPARK ADVANCE SETTINGS 10 15 " "Hg MAN. VACUUM 0 L I I I; I 015 I 20 200 250 300 350 400 450 500 550 TRUE SPARK ADVANCE (degrees BTC) Fig. 12. The effect of spark advance on the hydrocarbon emissions — RC engine without reactor. 21

6002000 RPM ENGINE SPEED REGULAR GASOLINE 500z 15 " Hg 1 II 20" Hg a0 400 0Q H " z 300 0 encircled points are standard carb.A/Fsettings 200 1.2 1.1 1.051.0.95.9.85.8.75 EQUIVALENCE RATIO - AL III I I I I I I I I0 II 13 13 14 15 16 17 18 19 20 AIR / FUEL Fig. 13. Hydrocarbon emissions as a function of air/fuel ratio with regular gasoline at 2000 rpm and. several intake manifold, vacuums-.sRC engine without reactor, MBT spark advance. 22

At mixture ratios leaner than the minimum emission air/fuel ratio a very sharp rise was noticed in the unburned hydrocarbons. This was caused by the misfiring which occured as the lean combustability limit was approached. Most of the air/fuel ratio and spark-advance testing was done at 2000 rpm. It was believed that this speed with variable loads was representative of most engine conditions and would suffice to show the effect of the stated variables on the exhaust emissions. The next series of tests were intended to measure how effective the induction system was at providing an equally distributed, homogeneous air/fuel mixture to the two rotors. To do this a propane-air carburetor and manifold system was designed and built which would essentially provide a single phase, homogeneous mixture to the engine. It is reasonable to expect that this was effectively accomplished in the propane-air system because propane is in the superheat region at room temperature and atmospheric pressure, and a large mixing volume was provided prior to induction into the engine. The results are plotted in Fig. 14 for the engine at 2000 rpm with three different loads and show the same characteristic shape as the hydrocarbon emission-:air/fuel ratio curves for gasoline. The only significant variation is that the minimum concentration point for 20 in. Hg. manifold vacuum occurs at a leaner mixture ratio than those of the 10 and 15-in. curves. Generally this minimum point would be expected to both increase in mangitude and occur at higher mixture ratios with an increase in intake manifold vacuum due to greater exhaust dilution and decreased absolute pressure which cause increased time losses. The gasoline and propane data is compared in Fig. 15 which shows that the minimum emission point occurs at lower equivalence ratios (leaner mixtures) with the propane fuel. Several points must be brought out about this comparison. 1. The only significant information is that the gasoline fueled engine can not operate as lean as the propane fueled engine and still achieve minimum emissions. 2. The absolute emission levels cannot be directly compared because the use of propane and gasoline as fuels result in different exhaust gas compositions which may cause the Beckman NDIR analyzer to respond nonlinearly with respect to concentrations from the two exhaust samples. 3. The results are plotted against equivalence ratio (relative fuel/air ratio) because the gasoline-air and propane-air chemical reactions have different stoichiometric air/fuel ratios, (14.8 for gasoline, 15.7 for propane). Figure 16 shows the results of operating the RC engine at several water inlet temperatures and demonstrates that high coolant temperatures are desirable to minimize the hydrocarbon concentrations. The testing was done at 1500 and 2000 rpm with 5, 10, 15, and 20 in. Hg. intake manifold vacuum. The hydrocarbon emissions are about 10) higher with a 140IF water inlet temperature 23

700 2000 RPM ENGINE SPEED Ic3~~~ ~~PROPANE FUEL x I' 600 c E Q. 0. 500 z O \ 0: z 400 cS)~~j~ 300 \ 91"20" Hg 0 I20 1 0" Hg 100 222 1.1 20.0 19.1 182 173 AIR/FUEL RATIO Fig. 14. Hydrocarbon emissions as a function of air/fuel ratios with propane fuel at 2000 rpm and several intake manifold vacuums-RC engine without reactor, MBT spark advance. 24

2000 RPM ENGINE SPEED REGULAR GASOLINE - -- PROPANE ^600 C 0.., c E 500 / O. / aI I \ \ //, z 20 H/ -2 o A&. I0 100 20"H 0 0, I I I I!.60.70.80.90 1.00 LIO EQUIVALENCE RATIO- 4 Fig. 15. Hydrocarbon emissions as a function of equivalence ratio with both regular gasoline and propane fuel —RC engine without reactor, MBT spark advance. 25

800 700 140 ~F 1500 RPM ENGINE SPEED 1700F a 200~F- \ 600 o 500 c] 400 200 F "1700 F ELI F500- \ \ _ 2000 RPM w 3oo - ^^ \ ENGINE SPEED - 1400 F 0 z 0. 200 C0 100 0 \ II I 25" 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg) Fig. 16. Effect of coolant temperature on hydrocarbon emissions —RC engine without reactor. 26

than at a 170~F temperature. Beyond 170~F little reduction in concentration was observed with increasing engine coolant temperature. During the context of the RC engine test program it became of interest to measure the exhaust gas hydrocarbon content of a conventional automotive reciprocating engine fi.i comparative purposes. A 1965 model Chevrolet 283 CID engine was selected for this test since it was available in the Automotive Laboratory and had approximately the same power rating as the RC engine. The results of this test are shown in Fig. 17. The se general trend of hydrocarbon emissions with speed and load that occurred in the RC engine was observed in the Cheverolet engine. The RC engine, however, demonstrated a greater range of emission levels, being much higher at low speeds (under 2000 rpm) and moderately lower at high speeds (greater than 2000 rpm). The emissions from the Chevrolet engine increased at high loads due to the functioning of the enrichment valve in the carburetor. The current automotive engines exhibit much lower emissions at steady state than this engine because of the addition of the so-called "Smog Package. " This suggests that the RC engine, with the proper modifications, could also be built to satisfy current Federal exhaust emission standards. B. BASE ENGINE WITH THE EXHAUST REACTOR As with the initial study of the RC engine without the exhaust reactor the first experimental study with the reactor was the establishment of baseline exhaust hydrocarbon emission and other pertinent data with standard operating conditions~ This base line information is shown in Figs. 18 through 24. The first of these figures, Fig. 18, indicates the exhaust hydrocarbon emissions at 2000 and 3000 rpm as a function of load (intake manifold vacuum). In addition to the previous practice of reporting the results on a concentration basis, the mass hydrocarbon emissions per unit time (lb/hr) are also showno A computer program was developed which utilized hydrocarbon concentral tion data along with engine air flow and fuel/air ratio data to compute the mass emissions. Petails of this procedure are shown in Appendix E. The mass emission calculations were stimulated by the fact that the Federal government is giving serious consideration to a total mass rather then concentration standard in 1970. Perhaps the most significant conclusion which may be drawn from this figure, with reference to Fig.,1 is -liat merely by the addition of the reactor approximately a 50% reduction in the hydrocarbon emissions is affectedo The hydrocarbon concentration is almost constant at 75-100 ppm at 3000 rpm and decreases almost linearly from 275 to 100 ppm with increasing load at 2000 rpm. Because of the fact that the air flow rate changes with both speed and manifold vacuum the mass emission data in the upper curves of Fig. 18 does 27

1000 C 900 - 800 C 7000ft..O1000 RPM Qz o 600 O6OO? 500 I 2000 RPM Z 400 ~ 0 ^ 3000 RPM 0 200 0o Q:'1 100 0~ 25"' 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg) Fig. 17. Steady state hydrocarbon emissions from a 1965 Chevrolet (2o3 CID) engine. 28

not follow the same pattern as the volume concentration data. Here it can be seen that even though the air flow rate is higher for a given manifold vacuum at 3000 rpm than 2000 rpm, the mass emissions are still higher at 2000 rpm except at high loads. The RC engine horsepower and torque data is shown in Fig. 19 as a matter of interest. Also included are curves of the dynamometer rpm, which is different than the engine speed because of transmission slippage. Note that this slippage, as expected, increases with load. An important variable with the addition of the somewhat restrictive exhaust manifold reactor was the back pressure at the exhaust ports. These data for 2000 and 3000 rpm are shown in Fig. 20. The system seems to become quite restrictive (12 in. Hg gage pressure) at moderate speeds and high loads. One can speculate that this back pressure might be a limiting factor in the design of a reactor, considering that the engine should function efficiently at speeds up to 6000 rpm and loads of 2-3 in. Hg. manifold vacuum. Test data was not obtained at these high speeds and loads because of an imbalance in the dynamometer coupling which caused excessive vibration above 3000 rpm. Contrary, to the adverse effect on engine power, the higher the exhaust back pressure the greater is the exhaust reactor's effectiveness. The reactivity of the unburned hydrocarbons increases functionally with pressure. A variable exhaust restriction device which would increase the back pressure at low loads and speeds, and relieve back pressure at high loads and speeds could offer additional emission control. Figures 21 and 23 indicate the pertinent low termperature data from the engine at 2000 and 3000 rpm, respectively. Included are the oil in and out, water in and the air inlet temperatures. Generally,these temperatures remaim almost constant throughout the range of loads. High temperature exhaust gas data is shown in Figs 22 and 24 also for 2000 and 3000 rpm. Relative to functioning of the reactor as an exhaust gas combustion chamber this information is of critical importance. It has been shown that in the temperature range Of approximately 1400~F-2000~F for every 400F increase in temperature there is a corresponding doubling of the reactivity of the unburned constituents in the exhaust gas. Thus, it is imperative to maintain the temperatures before and in the reactor at the highest possible and/or practical values to achieve the maximum reaction in a minimum reactor volume. The RC engine appears to be particularly adaptable to the type of exhaust reactor used in this study for a number of reasons: 1. An exhaust process occurs every crankshaft revolution in each exhaust port rather than every other revolution as in the case of the 4-stroke cycle reciprocating engine. 29

HYDROCARBON #/HR 2000 RPM.150 LI.050 *0 L_______I HYDROCARBON PPM. 3001c o 5)0L^\A aO \3000 RPM v O oZ 200 1 Q().050.0 HYDROCARBON PPM 2 50 Z 200000 RPM (0 150 25" 20" 1005" 0" 5 INTAKE MANIFOLD VACUUM ( inches Hg ) Fig. 18. Baseline hydrocarbon emissions as a function of intake manifold vacuum-standard RC engine with WAD reactor, no air injection. 30000 RPM c r 0

~ 70. 3000 RPM U) 60 0 50 40 320 RPM cr 20 10 150 2000 RPM U) m I 2500 - 3000 RPM LL 0002000 RPM 3000 3000 RPM E 2500 0 z >- 1500 - 0 25" 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg) Fig. 19. Baseline horsepower, torque and dynamometer speed characteristicsstandard RC engine with WAD reactor, no air injection. 31

14 12 2000 & 3000 RPM ENGINE SPEED I10'I #1I# # 2 EXHAUST - 3000 RPM / en #1 88#2 6 / EXHAUST- 2000 RPM a. 4- U) I |I^~~~ / g~ ^Reactor Out 3000 RPM 2 Reactor Out 2000 RPM 0,I I1 25" 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg) Fig. 20. Exhaust back pressure in engine exhaust ports and at the exhaust manifold reactor outlet-RC engine with WAD reactor, no air injection. 32

2000 RPM ENGINE SPEED 200- ^ QOil Out 2od180 160 - L \ Water In [ 140- Oil In? 1202 / Air I.take W LL 100 8025 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg) Fig. 21. Engine oil, water, and air temperatures at 2000 rpm —standard RC engine with WAD reactor, no air injection. 1700- 2000 RPM ENGINE SPEED 1600 - 0 \ Reactor Core w 2 1500- \, W *, #2 Exhaust 1400 1400~ #2 ^~^1 Exhaust 13000 25" 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg) Fig. 22. Exhaust port and reactor core temperatures at 2000 rpmstandard RC engine with WAD reactor, no air injection. 33

3000 RPM ENGINE SPEED 220 Oil Out 200 180 - Water In Oil In u 160 r.140 M / Air Intake Iz 120 a. X,F W 100ID25" 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg ) Fig. 23. Engine oil, water, and air temperatures at 3000 rpmstandard RC engine with WAD reactor, no air injection. 3000 RPM ENGINE SPEED Reactor Core 1900 1800 L. 0 W #lExhaust F 1700 #2 Exhaust 1600 25 20 15 10 5 INTAKE MANIFOLD VACUUM (inches Hg) Fig. 24. Exhaust port and reactor core temperatures at 3000 rpmstandard RC engine with WAD reactor, no air injection. 34

2. The higher emissions of the RC engine are benefical in obtaining high combustion temperatures during a self-sustaining reaction in the exhaust system. 3- The exhaust ports are in close proximity to one another which permits the use of a small size exhaust collector with a low surface-to-volume ratio which minimizes heat transfer. 4. There is a minimum length exhaust passage with a relatively low surfaceto-volume ratio from the engine combustion chamber to the exhaust reactor. This minimizes heat transfer through the port walls into the engine collant. It is interesting to note that the exhaust port and core temperatures are higher at the lightest measured load conditions than at 15 in. Hg. manifold vacuumo This is probably caused by a combination of two factors: 1. Fuel/air ratio enrichment; at very light loads the carburetor idle system is functioning and supplying a relatively richer mixture than the lean, cruising mixture ratio to the engineo 2. Spark retardation; since the intake manifold vacuum sensing port for the vacuum spark-advance system is located above the throttle valve at its closed position, spark retardation probably occurs with light loads. Ramifications of this can be seen in the emission data in Figs. 10 and 18 which show that there was a less than expected decrease in the hydrocarbon emissions at 15 in, Hgo manifold vacuum compared to the other engine loads. Also note that there was a temperature increase of about 150"F from the exhaust port to the reactor core for all cases. Several comments can be made about this measured temperature increase: 1. A major portion of this temperature increase was probably due to the energy release from the reaction of previously unburned carbon monoxide and hydrocarbons with oxygen in the reactor. 2o It is unlikely that the measured core temperature was the maximum exhaust reactor temperature since the measurement was made at the outside surface of the internal chamber. The maximum temperature probably occurs at the baffle surface on which the exhaust gases impinge or in the interior of the inner-cylinder. 3. Radiation losses from the bayonet type thermocouple to the relatively cool exhaust port walls may have caused some error in the measurement of the exhaust port temperatures. 4, It is not possible to predict the temperature increase theoretically with the data observed in this study because the carbon monoxide concentration reduction in the reactor was not measured. The CO may not be an important factor, however, with lean mixtures since its concentration in the exhaust should be quite low, less than 1%. Another exhaust reactor parameter, one that is particularily important with respect to a vehicular installation is the region of maximum temperature and the maximum temperature at the outside surface of the reactor. It was found to occur at the 90~ bend in the outlet pipe of the reactor and to at35

tain a maximum temperature of 1525~F at 3000 rpm and 5 in. Hg. manifold vacuum. This data is not shown in any of the figures. Base line data was not studied for the light load, 1000 rpm condition because of the uncertainty in the idle mixture settings which have predominant effect on the overall mixture ratio at this state. In all subsequent work the idle screws in both the standard and variable fuel/air ratio carburetor were adjusted for minimum hydrocarbon emissions (approximately 1 turn from fully closed). The engine appeared to still operate smoothly with this somewhat leaner than standard idle mixture ratio. No quantitative measurement of fuel/air ratio was made, however. This idle study gave the first indication of the engine-reactor combinations' extreme sensitivity to mixture ratio. At 1000 rpm and 20 in. Hg. manifold vacuum the hydrocarbon effluent decreased from about 1200 ppm for the engine without the reactor and with the Curtiss-Wright specified idle setting to approximately 150 ppm for the engine with the reactor and minimum emission idle adjustment, a reduction of 90%. An important aspect of any exhaust emission control device is that it must rapidly attain a satisfactory operating condition after a "cold start." This "cold start" data for the RC engine is presented graphically in Figs. 25 and 26. The engine was started and run at 1000 rpm for the duration of the test, with the transmission in neutral and the vacuum spark advance line disconnected and plugged. A 50N reduction in hydrocarbon emissions was observed after 30 seconds and a 95% reduction after two minutes. The warmup appears to be very rapid. Thus, the reactor should be near its optimum operating temperature for all "test points" of the "California Cycle." The dependence of the emissions on temperature in the reactor is quite evident with hydrocarbon concentration being inversely proportional to core temperature. Note that the engine coolant has not attained its steady-state operating temperature at the completion of the test (10 minutes after starting). The great effect of fuel air ratio on the emissions is shown in Figs. 27 and 28 for the engine-reactor combination with the fuel/air ratio set for minimum emissions. It was consistently possible to obtain hydrocarbon concentrations of less than 75 ppm, the lowest observed value being 30 ppm at 3000 rpm and 10 in. -Hg, manifold vacuum. High loads'at 3000 rpm were not investigated because the emissions were already low and the high reactor temperatures, which occurred at these conditions, could have caused a serious metallurgical failure. This test program modification was followed throughout the remaining studies. Due to erratic behavior of the Orsat apparatus which was used to measure the air/fuel ratio, it was not possible to determine the air/fuel ratio accurately. However, it appeared to be in the range of 15.5:1 to 16.5:1. Quite obviously the: mixture ratio must be leaner than stoichiometric to provide excess air which can react with the unburned pollutants in the reactor. 36

1000 RPM TRANSMISSION IN DRIVE 1500 VACUUM SPARK ADVANCE DISCONNECTED 1400 ____ 1800 I XX-X- X X x X * ^ X — A x-X 1300 CORE TEMP - 1600 1200 x -1400 ll100-/ 1200 x 1000 - 1000 9 00 - 800 9-I LL 0= CL \16 - 800 160 U.z o HYDROCARBON 0 z60 7 EMISSIONS 10 o 500 z 600 \ 140 C) o 500- \ Q -130 > 200- 10 o 400- o - 100 o 0 o 0 1200 04610 1 2 3 4 5 6 7 8 9 10 TIME (MIN.) Fig. 25. Cold start, warm-up test-hydrocarbon emissions, and reactor core and engine cooling water temperatures as a function of time-RC engine with WAD reactor, no air injection. 37

1700 oole 1700 o REACTORS o.CORE 1600- / REACTOR OUT 1500- ____^D D O D ~ \ x 1400- / #2 EXHAUST PORT 1300 / LL 1200 z 1100-/ 1000 A NOTE- 1 8#2 EXHAUST PORT TEMPERATURES 90 o WERE APPROXIMATELY EQUAL 2 900 800700 600 500 0. I I I I I I I, I 1 2 3 4 5 6 7 8 9 10 TIME (MIN.) Fig. 26. Cold start, warm-up test-exhaust port, reactor outlet and reactor core temperatures as a function of time-RC engine with WAD reactor, no air injection. 38

2000 RPM ENGINE SPEED REACTOR TEMPERATURES 1900 1800 o0 Reactor Core 1700 \ o 1700 o Reactor Out 1600 ~/ j 1500 I2 Exhaust Port 10 1-1400' 2- - # 2 E)(haust Port o a BRAKE HORSE POWER 30 w20 co 30 O 100 HYDRO CARBON EMISSIONS 0<f 1- O 0 0 25" 20 15" 10" 5 INTAKE MANIFOLD VACUUM (inches Hg) Fig. 27. RC engine performance at 2000 rpm with the fuel/air ratio adjusted for minimum emissions —MBT spark advance, WAD reactor, no air injection. 39

2000 REACTOR TEMPERATURES 3000 RPM ENGINE SPEED ~~~~~~1900~~0 ^^ 1^^ 900 Reoctor Core O' 1800 - 0 Reactor Out < 1700 8- I Exhaust Port 1600 2 Exhaust Port 1500 ojL 50 BRAKE HORSE POWER 20 _ 41 - IOCC z HYDROCARBON EMISSIONS 0o 75 0 z 25 0 o0o 5 I ^ -, 25" 20" I 10" 5" INTAKE MANIFOLD VACUUM (inches Hg) Fig. 28. RC engine performance at 3000 rpm with the fuel/air ratio adjusted for minimum emissions-MBT spark advance, WAD reactor, no air injection. 40

It is interesting to note that the minimum emission air/fuel ratio was greater for the engine without the reactor than for the engine with the reactor. This is caused at least partially by a temperature phenomena. The highest exhaust temperatures are found with close to stoichiometric mixtures ratios. Thus, the combination of air/fuel ratio and high exhaust temperature must be considered rather than predominatly air/fuel ratio as is the case in the engine without the reactor. There is a distinct advantage to operation with the richer minimum emission mixture ratio of the engine-reactor combination, drivability. In a vehicle the richer mixture should provide smoother operation. Important data from Figs. 27 and 28 is replotted in Figs. 29 and 30 together with equivalent data for the standard engine-reactor combination to show directly how they compare. Notice that a great reduction was found in the unburned hydrocarbons but only a slight decrease in brake horsepower was observed with the relatively lean mixtures used. The reactor core temperatures are also somewhat greater for the lean mixture operation even though the exhaust port temperatures are lower. It can be hypothesized that if the mixing of the fuel and air is improved in the carburetor and induction system an even greater decrease in exhaust emissions might be achieved with perhaps an increase in the drivability of the engine. This might be accomplished by using increased intake manifold heating,* a smaller venturi carburetor, dual runner manifold, fuel injection or other such modifications. The fact that the engine responds so well to lean mixture operation could be a major factor in gaining acceptance for the RC engine in an automotive application. The minimum emission air/fuel ratio hydrocarbon concentration data was converted to a mass base (lb/hr) and was plotted against intake manifold vacuum in Fig. 31. Here also can be seen the magnitude of the reduction with a controlled mixture ratio. The observed hydrocarbon emission level of about.06 lb/hr. is quite low, based on experience with current automotive engines. The following figure, Fig. 32, shows the before mentioned mass emissions plotted against brake horsepower. This figure is particularly important with respect to a vehicular requirement. A given automobile requires "x" horsepower to propel it along the highway at "y" mph or "P" horsepower to give it an acceleration rate of "q" ft/sec.. Thus, if a modification is made to the engine which affects a decrease in emissions, the power gain or more commonly loss must be considered. If, as is the case with lean mixture ratios, the power is decreased, something must be done to increase the power back to the vehicle required power. Generally, this is accomplished in an engine by in*The heat riser was disconnected in all of the engine-exhaust reactor research so that a direct comparison could be made with earlier work on the base engine which, because of a special exhaust manifold, did not employ intake manifold heating. 91

2000 RPM ENGINE SPEED REACTOR CORE TEMPERATURE OF 1850 o 1800 F/A CARB. SET FOR w MINNIMUM EMISSIONS at 1750 < 1700 0r 3 STANDARD CAR B. 1650 - 1600 15501 0 I 40 BRAKE HORSEPOWER 40 I ^^ MINIMUM EMISSIONS 0HYDROCARBON EMISSIONS O 3200 STANDARD CAR B w 2 F/A CARB SET FOR MINIMUM 3:: LU ^~MINIMUM EMISSIONS r 10 m HY DROCARBON EMISSIONS 25" 20" 15" 10" 5" INTAKE MANIFOLD VACUUM (inches Hg.) Fig. 29. Comparison of the RC engine performance data at 2000 rpmminimum emission F/A ratio and. standard. F/A ratio-MET spark advance, WAD reactor, no air injection. 42

3000 RPM ENGINE SPEED OF REACTOR CORE TEMPERATURE 2000 0 LIUJ STANDARD F/A SET 1900o CARB. FOR:0 EMISSIONS 1800o_ a: ~BRAKE HORSEPOWER 3 60 LB i, 500 X 4 STANDARD CARB. ~. 40 S~ //^^~ ~F/A SET FOR 30 /~ BMINIMUM EMISSIONS m 30 20 10 I e 5 HYDROCARBON EMISSIONS mZ 150 mcr I0 ^STANDARD CARB. 0or 100 Z 50 MINIMUM EMISSIONS o /, / / / / / /... 25 ~ 20" 15" 10" 5" INTAKE MANIFOLD VACUUM inches Hg) Fig. 30. Comparison of the RC engine performance data at 3000 rpmminimum emission F/A ratio and standard F/A ratio —MBT spark advance, WAD reactor, no air injection. 43

#/HR 2000 RPM 2 Standord o0 Corb. 2.100 wr I Variable F/A Carb. O 0 <m Co.050 r I.0 25" 20" 15" 10 5" 5II #/HR.150 p 3000 RPM 2 Standard Carb. w ~ 050 - i:) INTAKE MANIFOLD VACUUM ( inches Hg) Fig. 31. Mass rate of hydrocarbon emissions with the standard carburetor and the variable fuel/air ratio carburetor adjusted for minimum emissions RC engine with WAD reactor, no air injection. 44

0i^ i ACV.150 co 2000 RPM fL~z $000 RPM STANDARD m CARB. o.107 ir a'~ 2000 RpM00F~P.50.50 / 2000 RPM o^__^_^ ^^^^^^ \ A\ V\ \A \^Y 0 RISE F/A CARlu, ~_~__.._..._ to 0~~3 30 —.-40 BRAKE HORSEPowER 5060 Fig~j~~Mass rateOf h orsepos at f hydrocarbon emissions as horburetor th the tandard carburetor a no &d~aiOusted for mu nto of brake carb~n'ilureto ajstedio ns -aiale fuel/air ratio no air injection. m inimum emissions__RC engine with WAr TfA ractor

creasing the throttle opening which in turn increases the mass flow of air and fuel to the engine. Therefore, even though the modification may cause a reduction in pollutant concentration, the increase in mass flow may nullify the gain if the mass effluent (lb/hr) of the emissions were considered. Hence, the critical importance of data such as that shown in Fig. 32. Spark-advance as demonstrated on the RC engine without the exhaust reactor has significant effect on emissions. The same relative effect was observed in the engine-reactor combination and therefore, the results are not plotted. Instead, both the fuel/air ratio and spark-advance were varied together. At each spark-advance setting the fuel/air ratio was adjusted to obtain minimum hydrocarbon emissions. These results are plotted in Figs. 33 through 37 which show important reactor temperatures, brake horsepower, and hydrocarbon concentration plotted as a function of spark-advance at 2000 and 3000 rpm and several intake manifold vacuums. Generally,the emissions decreased with retardation of the spark but not as much on percentage basis as those observed from the engine-reactor combination with the standard carburetor. All observed emissions were quite low, less than 75 ppm. The horsepower curves behave in a somewhat unpredictable manner, in some cases even showing a horsepower increase with spark retardation. This is not totally unexpected though, since both the spark advance and air/fuel ratio affect power and different fuel/air ratios were probably required to obtain minimum emissions at each spark advance. The reactor core temperature generally followed the pattern observed in the early studies being inversely proportional to the hydrocarbon concentration though several exceptions were observed. Light loads at both 2000 and 3000 rpm exhibited core temperatures in excess of 1800~F, certainly an indication of a high quality reaction. This reactor core temperature cannot, however, be used as a quantitative measure of the hydrocarbon emission concentration because the increase in temperature from the exhaust port to the core is a function of both the mass of reactants which are oxidized in the device and the mass rate of flow through it. For a given reactor temperature is a greater emission reduction would be expected at 2000 rpm and light load, for example, than at 3000 rpm and the same load since the exhaust gases are exposed to the high temperature region for a greater period of time. The variable fuel/air ratio and spark advance data is summarized in Fig. 38. Mass hydrocarbon emissions based on the minimum hydrocarbon concentration from the full range of spark advances examined and reactor core temperature are plotted against brake horsepower and thus relate to a vehicular requirement. Also shown is similar data for the base engine with the reactor. Througout the power range observed in this test the mass emissions were from 50% to 75o less than those of the base-engine-reactor combination and from 75 to 90G4 less than those from the base engine alone, certainly a significant improvement. One point of significant interest is the comparison of the mass emissions observed in this test to those from the test where only the fuel/air ratio was k6

2000 RPM, 17 inches Hg INTAKE MANIFOLD VACUUM Lu 1700- REACTOR CORE TEMPERATURE "F | BASE LINE DATA( w/Reactor) a: STD CARB., 440 SPARK ADVANCE.W ~ CORE TEMP- 1550 F 21650 - BRAKE HORSEPOWER - 13 HP uJ F HYDROCARBONS - 265 PPM OL g ^14 BRAKE HORSEPOWER WI 130 oO, n 200 - HYDROCARBON EMISSIONS a: 0 T 12 < - I I I III I 1 34~ 36~ 38H 40~ 42~ 44~ 46E 48~ 50~ 52~ SPARK ADVANCE OBTC Fig. 33. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 2000 rpm, 17 in. Hg. —RC engine with WAD reactor, no air injection. 47

2000 RPM, 10 inches Hg. INTAKE MANIFOLD VACUUM REACTOR CORE TEMPERATURE 1700 LL:) H 1600 a: 1500 BASELINE DATA (w/Reactor) O STD CARB., 380~SPARK ADVANCE CORE-TEMP- 15700 F BRAKE HORSEPOWER 29.6 HP 50 BRAKE HORSEPOWER HYDROCARBONS- 170 PPM cr 40CL w 30L! i / / / / / / /// // a 10_ 20 0a - HYDROCARBON EMISSTONS, 75 z 0 m < 50 __-__ __ _// 0 25>. ~~I~~~~~~~I 0 280 300 320 34~ 360 380 400 420 440 46~ SPARK ADVANCE OBTC Fig. 34. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 2000 rpm, 10 in. Hg.-RC engine with WAD reactor, no air injection. 48

2000 RPM, 5 inches Hg INTAKE MANIFOLD VACUUM REACTOR CORE TEMPERATURE 1650 LL eof LU1600 w1550 w w L BRAKE HORSEPOWER 18.5 HP HYDROCARBONS, 225 PPM 1500 0. BRR STD CARB. 440 SPARK ADVANCE - 169 LJL 15-CORE TEMP-5OF HYDROCARBON EMISSIONS225 P 0 175 0 16. w HYDROCARBON EMISSIONSPARK ADANCE BTC 0 o 25 320 340 36" 380 400 420 440 460 480 SPARK ADVANCE ~BTC Fig. 35. The effect of spark advance on performance with the //A ratio adjusted for minimum emissions; 2000 rpm, 5 in. Hg. —RC engine with WAD reactor, no air injection. 49

3000 RPM, 20 inches Hg- INTAKE MANIFOLD VACUUM REACTOR CORE TEMPERATURE LL o 2000 iW 1950 — A7/BASELINE DATA (w/Reactor) w ST'D CARB., 480 SPARK ADVANCE H CORE TEMP 1925 OF,~1900-~ ~BRAKE HORSEPOWER - 3 HP I9O"r~ ~HYDROCARBONS- 80 PPM = 4 BRAKE HORSEPOWER 4F r 252 w 100r HSYDROC ARBON EMISSIONS I0 75 a: 0L o 25 SPARK ADVANCE 0BTC Fig. 36. The effect of spark advance on performance with the F/A ratio adjusted for minimum emissions; 3000 rpm, 20 in. Hg. —RC engine with WAD reactor, no air injection. 50

3000 RPM,15 inches Hg INTAKE MANIFOLD VACUUM REACTOR CORE TEMPERATURE, 1900 0 Oi8O BASELINE DATA (w/Reactor) ac STD. CARB.,480 SPARK ADVANCE U CORE TEMP -1810~F - ^ 8BRAKE HORSEPOWER - 29.6 HP 10- HYDROCARBONS - 100 PPM 1800 o0 28 BRAKE HORSEPOWER a. 27 U 0 26?LLr 1IO HYDROCARBON EMISSIONS I00 a. a* 75 z 0 -5025 0 v 340 360 380 400 420 440 460 480 50~ 52 540 SPARK ADVANCE ^BTC Fig. 37. The effect of spark advance on performance with the i/A ratio adjusted for minimum emissions; 3000 rpm, 15 in. Hg. —RC engine with WAD reactor, no air injection. 51

-- -BASE DATA POINTS WITH REACTOR, STANDARD CARB 8 SPARK ADVANCE 0 VARIABLE F/A AND SPARK ADVANCE,BOTH SET FOR W o MINIMUM EMISSIONS, REACTOR CORE TEMPERATURE!^^2 t~~ ~(MAXIMUM OF 100 RETARD ON SPARK W FROM MBT SETTING) a- 1900 1800- ^ ^^^^3000 RPM OL 17009 H 1600 I1500 2000 RPM.150-'1 z / / Io / 2000 RPM.100 0.Ioo ^cr.40 \ 3000 RPM 0500 10 20 30 40 50 BRAKE HORSEPOWER Fig. 38. Hydrocarbon emissions and reactor core temperature as a function of brake horsepower with both spark advance and F/A ratio optimized for minimum emissions-RC engine with WAD reactor, no air injection. 52

varied. Only slight improvement can be seen which demonstrates the powerful effect of optimizing just the fuel/air ratio. C. RC ENGINE WITH EXHAUST REACTOR AND AIR INJECTION The final phase of the experimental study involved the measurement of the hydrocarbon emissions from the RC engine with both the exhaust manifold reactor, and air injection in the engine exhaust ports. Due to the limited time available for this part of the work, quantitative measurements were made only for the engine with the standard carburetor and spark advance. Data was taken at a full range of loads at 2000 rpm and only light loads (15 and 20 in. Hg. manifold vacuum) at 3000 rpm because at higher loads the emissions were already low and reactor temperatures excessive. The results are plotted in Figures 39, 40, and 41 which show the corrected hydrocarbon emission concentration as a function of air injected, measured in per cent of engine combustion air. All concentration measurements made with the Beckman NDIR analyzer were corrected to compensate for dilution caused by the added air. Generally, the emissions decreased with an increase in air-injection rate up to 10-15% of engine air after which the concentrations became almost constant. At all test points but one, emissions of less than 100 ppm were observed. The only exception was at 2000 rpm, 15 in. Hg. manifold vacuum where the hydrocarbon concentration was somewhat greater. Also, the leveling off of the curve at high air injection flow rates was not observed at this test condition. Apparently this behavior was caused by the lack of proper operating conditions (air/fuel ratio and spark advance) for an efficient reaction with low injection rates. Qualitatively data was observed for several combinations of air/fuel ratio and air-injection rate at the test points. The results demonstrated that only a slight improvement in the air-injected, standard engine emissions could be made. Perhaps the most interesting result of this phase of the study was that the RC engine with the reactor, no air-injection and optimized fuel/air ratio performed better than the RC engine with the reactor and air injection. This fact is very important when considering a possible production application for the engine. The advantages are threefold if the emissions can be controlled without using rich mixture ratios and air injection: 1. Decreased initial cost 2. Better economy (due to the use of lean mixtures) 3. Decreased complexity 53

300 2000 RPM, 20 inches Hg INTAKE MANIFOLD VACUUM z <: \ x \ I \: 200 - a. a. 100 z 0 z, 0 10 20 30 40 50 C) 300 r 2000 RPM, 15 inches Hg INTAKE MANIFOLD VACUUM 0 \ 0 100 O 0 0 K 20 30 40 50 AIR INJECTED —/ ENGINE AIR Fig. 39. The effect of exhaust port air injection on the hydrocarbon emissions; 2000 rpm, 15 and 20 in. Hg. —RC engine with WAD reactor. Q~~~~~~~~~~~~~~~)

300 2000 RPM, 10 inches Hg INTAKE MANIFOLD VACUUM z I c 200 S O I I I z 0 20 30 40 50 0 (r o z o0 m 300 2000 RPM,5 inches Hg INTAKE MANIFOLD VACUUM o L: 0 bJ'200 bJ a 100 0 0 0 10 20 30 40 50 AIR INJECTED.- % ENGINE AIR Fig. 40. The effect of exhaust port air injection on the hydrocarbon emissions; 2000 rpm, 5 and 10 in. Hg.-RC engine with WAD reactor. 55

300 x< 3000 RPM, 20 inches Hg INTAKE MANIFOLD VACUUM LJ CL 200a. az 0 100 R I z LLJ o 0 10 20 30 40 50 o 300 3000RPM 15 inches Hg INTAKE MANIFOLD VACUUM 0 w 200 100 0 0.___,...__..I,, I 0 10 20 30 40 50 AIR INJECTED -% ENGINE AIR Fig. 41. The effect of exhaust port air injection on the hydrocarbon emissions; 3000 rpm, 15 and 20 in. Hg.-RC engine with WAD reactor. %

D. DISCUSSION OF THE CURTISS- WRIGHT EXHAUST MANIFOLD REACTOR The exhaust manifold reactor used in the second phase of the experimental study was designed by the Curtiss-Wright Corporation specifically for the RC 2-60 engine. As demonstrated by the results of the emission studies this reactor performed extraordinarily well in its role as an after engine, combustion chamber. However, even though the reactor was made of a high quality turbine alloy, Hastelloy X, several failures did occur which cast some doubt on the design. Perhaps a ceramic material of some type would be as effective and undoubtedly, much less expensive. The reactor assembly is shown in Fig. 42 viewed from the side which attaches to the engine exhaust ports. In Fig. 43 it is shown disassembled (note this is after 100 hours of use). Located in the center of the photograph is a shell consisting of two concentric cylinders of Hastelloy X which enclose an insulating material. Several failures occurred in the welds at the end of this shell caused by excessive grinding of the original welds, stress risers due to the welding, and high thermal stresses. These failures were noticed on disassembly after approximately 30 and 60 hours of running. Repairs were made by Saffran Engineering Company in Detroit. The cylinder at the lower part of the picture is the reactor core into which the hot exhaust gases are directed from the engine. This core is subjected to the highest temperatures. The core is shown in more detail in Figs. 44 and 45. In Fig. 44 the baffle plates in the core interior can be seen. These eroded substantially during the test program demonstrating the highly corrosive nature of the high temperature exhaust gases. The effect of this high temperature and the significant temperature gradients within the reactor is shown in the end view of the core in Fig. 45. The originally circular cross-section has deformed into a noticeably elliptical shape. It is possible to generalize somewhat with respect to the design of any exhaust reactor and say that there are several factors which work together to determine its effectiveness. These factors are: 1. Temperature 6. Mass flow Rate 2. Volume 7. Specific reaction rate 3. Pressure of the unburned components 4. Time 5. Excess oxygen If, for example, the reactor temperature could be increased both its volume and/or the dwell time of the gases in the reactor could be decreased. Excess oxygen, however, must always be present and well mixed with the exhaust to achieve a more complete reaction. Unfortunately, the increased temperature would probably cause severe metallurgical problems, and more than likely increase the cost and decrease the expected life of the reactor. 57

Fig. 42. WAD reactor assembly-viewed exhaust port side. Fig. 43. WAD reactor disassembly-after 97-1/2 hours of testing. 58

Fig. 44. WAD reactor inner core-after 97-1/2 hours testing. Fig. 45. WAD reactor inner core-end. view showing heat distortion after 97-1/2 hours testing. 9 — Fig. 45. WAD reactor inner core —end view showing heat distortion~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iiiliii~liiii:-iiii after 97~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-1/2 hours testing.i

E. ANALYSIS OF THE RC ENGINE EXHAUST EMISSIONS The RC engine presents a difficult analysis problem to one concerned with explaining the whys and wherefores of its exhaust emissions. The exhaust emissions from the combustion chamber of the RC engine as well as other spark-ignition engines are a result of several processes. 1. Perhaps the most 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. Remaining in this so-called "quench zone" is raw fuel/air mixture, partially burned and/or cracked fuel, carbon monoxide, and other less noxious products. There is apparently little tendency to form oxides of nitrogen, however, in this relatively low temperature region. During the exhaust process some of these "quench zone" products are scavenged from the chamber walls and released to the atmosphere, and some are retained and partially burned on the succeeding cycle. This emission source is particularly significant in the RC engine for two basic reasons: a. The RC engine has a high surface/volume ratio in the combustion chamber and thus, a large quench zone volume. b. The trailing apex seal of a given rotating combustion chamber "scrapes" the quench layer off the rotor housing and into the peripheral exhaust port. 2. The second possible source arises from another quench process. In this case if two relatively cool surfaces enclosing a mixture are brought progressively together, a certain separation distance (quench distance) will be reached where a flame will not propagate between them. The volume contained in these quenched regions is filled primarily with unburned fuel/air mixture which has an extremely high concentration of unburned hydrocarbons. If this volume is relatively large and is vented unburned to the atmosphere during the engine cycle the overall hydrocarbon emissions may be greatly affected. The RC engine in its present form has substantial combustion chamber volume contained between surfaces in closer proximity than this "quench distance." These regions include: a. The volume above the side seal and between the rotor and rotor housing. b. The leading and trailing portions of the rotor surface which are close to the rotor housing. c. The elevated rib bordering the rotor cut-out may also be too close to the housing. 3. Leakage of unburned mixture to the atmosphere can affect the emissions significantly. Raw fuel/air mixture when considered alone has an unburned 60

hydrocarbon concentration of from 17,000 to 25,000 ppm. Thus,even a small percentage loss of mixture can have a great effect on the overall concentration of hydrocarbons in the exhaust. A conventional four-stroke cycle automotive engine loses only a very small portion of unburned mixture past the closed exhaust valve. The blowby is not considered because it is rebreathed with current emission controls. The RC engine, however, shows significant leakage of raw mixture past the leading apex seal, which is lost directly through the exhaust port to the atmosphere. 4. Incomplete combustion in the reaction zone can result in emissions, particularly with rich fuel/air mixtures and/or poor mixing of the fuel and airo The RC engine would appear to have a minimum problem from this source because it operates well on lean mixtures and has intense combustion chamber turbulence which promotes homogeneity and complete combustion. 61

APPENDIX A. COMMENTS ON THE RC ENGINE INSTALLATION 1. A 1964 Dodge Dart automatic transmission P/N 2466092 and torque converter P/N 2204403 were connected to the engine as safety devices to prevent starter motor overload due to the large rotor inertia of the MidWest dynamometer and to minimize a possible torsional vibration problem. 2. A spicer coupling, which is used on most of the dynamometers in the Automotive Lab, was welded to a Dodge universal joint to connect the transmission to the dynamometer. This was done in lieu of machining the Thomas coupling specified by Curtiss-Wright. 3. Lake Shore engine stands were used with rubber motor mounts in place of the pedestal mounts indicated by blueprints Nos. 318118 and 319423. 4. All of the iron-constantan thermocouples, except for the one mounted in the skin of the combustion chamber of the anti-drive rotor, were replaced with copper-constantan and chromel-alumel thermocouples which matched the reference junctions of the Brown potentiometer in the dynamometer console. 5. In the experimental studies without the WAD exhaust reactor the exhaust manifold supplied with the engine was replaced with individual steel pipes which were attached directly to the exhaust ports. Several feet downstream from the ports the pipes were brought together' into a single exhaust pipe. This system permitted either individual sampling of the exhaust gases from each rotor, or sampling of the total engine exhaust. 62

APPENDIX B. RC ENGINE COMBUSTION AIR FLOW MEASUREMENTS 700 - 600- 3000 RPM 500 * 400 - T;~ ^^^~~, 2000 RPM 0 c 300 200 1000 RPM 100 0 I 25" 20" 15" 10" 5" INTAKE MANIFOLD VACUUM ( inches Hg) Fig. 46. RC engine mass rate of air flow as a function of intake manifold vacuum. 63

APPENDIX C. ORSAT CALIBRATION CURVES 15 14 ~02 13 12 II10- \ / FUEL-C8H17 o_ 8 0- 7~ 7 00 6 _n w CO 3 5 (>-> D 4w 0' 3 /^ 2 / \ 8 9 10 11 12 13 14 15 16 17 18 19 20 I \I IAIR/FUE RATIO I.10.090.080.070.060 050 FUEL/AIR RATIO Fig. 4.7. Exhaust gas composition as a function of air/fuel ratio. 64

APPENDIX D. BECKMAI NDIR CALIBRATION CURVE 100 (n 80 w w 0 60 z 40 U) w L / 0 20 20 0 400 800 1200 1500 PPM C6HI4N N2 BY VOLUME Fig. 48. Beckman: calibration-n-hexane in nitrogen. ^ /~~~~~~~~6

APPENDIX E. CONVERSION OF CONCENTRATION HYDROCARBON EMISSIONS TO MASS EMISSIONS PER UNIT TIME Assume the dry exhaust gas is composed only of these components with their respective molecular weights; C02 - 44, N2 - 28, and C6H14 - 86. For each mole of CO2 there are approximately 6 moles of N2. Molecular weight of C02 and N2 mixture = (x44) + (6x28) 303 (E-l) If x = ppm C6Hh4 (n-hexane), then 86x is proportional to the mass of C6H14, and 30.3 (10 -x) is proportional to the mass of CO2 + N2, Mass CH 86 xH ___6 14 _ 86x ____ (E-2) Total mass 30.3(106-x)+86 x if x < 1000 ppm, less than a. 5 error will be introduced if the above relation is simplified to; Mass C6H14 86 x ~^ ~ (E-3) Total Mass 30. x 10 Engine air flow + fuel flow is; MT = MA + F/A MA = (1+F/A) MA where (E M = total mass flow rate and MA = air flow rate MT = (1+F/A) MA Thus, the mass rate of hydrocarbon emissions is MC6H4 = MA(1+F/A)(~6 x) 50.3x106 x = ppm n-hexane (measured) MC6H14 -* hr. n-hexane (E-5) 66

BIBLIOGRAPHY 1. Public Law 88-206, "The Clean Air Act", Approved December 17, 1963, U.S. Department of Health, Education, and Welfare. 2. Jackson, M. W., Wiese, W. M., and Wentworth, J. T., "Influence of AirFuel Ratio, Spark Timing and Combustion Chamber Design on Exhaust Hydrocarbon Emissions'", SAE paper 486B, March 1962. 3. Curtiss-Wright Corporation, Wright Aeronautical Division, "Preliminary Data Compilation for Curtiss-Wright RC2-60-U5 Rotating Combustion Engine", March 1965. 4. Obert, E. F., "Internal Combustion Engines - Analysis and Practice", International Textbook Company, Scranton, Penn., 1944. 5. "An Analytical Procedure for Simultaneous Measurement of Hydrocarbons, Carbon Monoxide and Carbon Dioxide in Automobile Exhaust Gas", Vehicle Emission Measurement Panel of the Vehicle Combustion Products Committee, Automobile Manufacturers Association, AMA Engineering Notes 64-, October, 1963. 6. "Control of Air Pollution from New Motor Vehicles and New Motor Vehicle Engines", Federal Register, Vol. 31, 61, March 30, 1966, and Federal Register, Vol. 32, 24, February 4, 1967. 7. Daniel, W. A., and Wentworth, J. T., "Exhaust Gas Hydrocarbons-Genesis and Exodus', SAE paper 486B, presented in Detroit, Mich., March, 1962. 8. Bennett, P. A., Jackson, M. W., Murphy, C. K., and Randall, R. A., "Reduction of Air Pollution by Control of Emissions from Automotive Crankcases", SAE Transactions, 68, 514, (1960). 9. Was, W. L., and Stanyar, A. E., "Air Cleaner Control of Crankcase Emission Gases," SAE paper 648E, presented in Detroit, Mich., Jariuary, 1963~ 10. Beckman Instructions Manual 1307-C,'Models IF 215, IR 315, and-IR 415 Infrared Analyzers", Beckman Instruments Inc, Fullerton, California, May, 1966. 11. Robison, J. A., and Brehob, W. M., "The Influence of Improved Mixture Quality on Engine Exhaust Emissions and Performance", presented at the Western States Combustion Institute Meeting, October 25-26, 1965. 67

12. Wisby, T. E., "How to Convert to LP-Gas Carburetion," Ross-Martin Co., Tulsa, Oklahoma, 1951. 13. Abell, C., "Butane-Propane Power Manual and Principles of LP Gas Carburetiont, Jenkins Publications Inc., Los Angeles, California, 1952. 14. Steinhagen, W. K., Niepoth, G. W., and Mick, S. H., "Design and Development of the General Motors Air-Injection Reactor System", SAE paper 660106, presented in Detroit, Michigan, January, 1966. 15. Schnabel, Jo W., Yingst, J. E. Heinen, C. M., and Fagley, W. So "Development of a Flame Type Afterburner", SAE paper 486G, presented in Detroit, Mich., March 1962. 68

U TVESI OF MICHIGAN 3 9015 02841 2602 THE UNIVERSITY OF MICHIGAN DATE DUE L^t(6 1H>+