NOX Emission Controls for Heavy-Duty Vehicles: Toward Meeting a 1986 Standard (1981)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Chapter 2 HEAVY-DUTY DIESEL ENGINES DESCRIPTION OF THE PROBLEM This introduction discusses the diesel combustion process from both fundamental and empirical points of view, and then summarizes the various proposed NOX control techniques for diesel engines. The object is to provide the reader with the background necessary to an understanding of the empirical data presented later in the chapter. The material in this section is drawn largely from a review by Henein (l976) and recent compendia of papers on diesel combustion (Society of Automotive Engineers, l980, l98l). Diesel Combustion and Emissions The combustion process is the critical factor in determining the performance, emissions and fuel economy of diesel engines. The diesel combustion process is fundamentally different from that of gasoline engines. In the latter, a premixed charge of air and fuel is ignited by a spark. In the diesel engine, the intake charge consists of air only. This air is compressed to a high temperature (higher than that of the gasoline engine due to the higher com- pression ratio of the diesel engine). Near the end of the compression stroke a spray of liquid fuel is injected into the engine; the temperature of the compressed air causes the fuel to evaporate, ignite, and burn. (Some new "stratified-charge" gasoline engines use combustion processes similar to those of diesel engines.) The amount of fuel that is added at the end of a diesel's compression stroke is less than the amount that could be burned completely by the air present. Thus, on an overall but not necessarily a local basis, combustion takes place under lean conditions, and carbon monoxide is readily oxidized to carbon dioxide; carbon monoxide emissions are therefore not a significant problem in diesel engines. Initially, some of the injected fuel evaporates and undergoes a series of chemical reactions leading to ignition. The time between the start of injection of the fuel and the start of ignition is called the ignition delay. The length of this delay is governed by the rates of the pre-ignition chemical reactions, as well as by the physical process of atomization and vaporization. All of these in turn depend on cylinder temperature and pressure. Actual 2l

22 ignition occurs around the periphery of the incoming fuel jet. In this region, the air-fuel ratio is probably near the stoichiometric ratio (at which the air/fuel ratio could just give complete combustion). After ignition, expansion of the burning mixture and the already present air motion provide the mixing and combustion for the rest of the fuel. The energy released by combustion heats any unburned fuel, which then undergoes chemical reactions in an oxygen-deficient atmosphere; these re- actions are the start of the formation of particulate matter. Although there is some uncertainty about the exact mechanism of particulate formation in diesel engines, it is generally agreed that the chemical reactions occurring in this oxygen-deficient region form species that can react to form large molecules. Nucleation processes can then form particulate matter from these molecules. Particulate matter so formed can then agglomerate. Once formed in the fuel-rich regions of the engine cylinder, the particulates can and do undergo oxidation reactions in the air-rich regions. The extent of this oxidation is limited by residence time and the temperatures and concentrations of oxidizing species present. The particulate matter finally appearing in the exhaust contains carbonaceous material (soot) and adsorbed liquid hydro- carbons. High temperatures and oxygen, which are conducive to the oxidation of particulates, are also conducive to the formation of oxides of nitrogen. Most of the nitrogen oxides formed in combustion come from oxidation of atmospheric nitrogen, with a much smaller amount from oxidation of fuel-bound nitrogen. The mechanism for oxidation of atmospheric nitrogen, the Zeldovitch mechanism, is very temperature-sensitive; near-equilibrium amounts of NO may be formed at the high temperatures of engine combustion, but very little decomposition of NO will occur after the temperature is rapidly quenched in the expansion stroke of the engine. Exhaust hydrocarbons consist of unburned and partially burned hydrocarbons. The presence of these hydrocarbons is due to quenching of the combustion reactions, both at the walls of the combustion chamber and in the lean regions of the combustion gases (where the lower temperatures favor mechanisms that lead to partially oxidized hydrocarbons rather than the products of complete combustion). During diesel combustion pressures and temperatures in the combustion cylinder can reach 2,000 pounds per square inch (l3.8 MPa) and 4500°F (2750 K). The complete injection and combustion process requires from 4 to l0 milli- seconds. Liquid fuel is injected at pressures ranging from 3,000 to 25,000 pounds per square inch (20.7 to l72 MPa), and the temperatures of the exhaust gases are typically less than l000°F (800 K). The combustion process involves the interaction of chemical and physical rate processes. Ignition of the fuel requires good mixing of the incoming fuel jet with the air. The two basic diesel types, direct- and indirect-injection engines, are differentiated according to the ways they accomplish this mixing.

23 A direct-injection engine has a single combustion chamber (the top of the cylinder plus any cavity in the piston). Air is introduced into this type of engine with some amount of swirl, provided by the inlet system, to aid mixing. Because this motion is fairly small, the fuel system must provide most of the motion necessary for adequate mixing. Consequently, direct- injection engines use high injection pressures, plus multiple small holes in the injection nozzles. An indirect-injection diesel engine has, in addition to the main combustion chamber, a prechamber into which the fuel is injected, typically through a single-hole nozzle. Mixing is provided by two mechanisms. The first is the air motion produced in the prechamber by the rising piston. The second is the motion of the prechamber gases as they are ignited and expand into the main chamber. Most heavy-duty diesels are direct-injection engines; most light-duty diesels are indirect-injection engines. However, direct-injection engines tend to have l0 to l5 percent better fuel economy, and manufacturers have been moving toward direct-injection engines for all applications. Exhaust treatment of gaseous diesel emissions is difficult because of the low temperatures and large oxygen concentrations. Because the three-way catalyst that is used in gasoline engines depends on having very small amounts of oxygen in the exhaust gases, it cannot be used to remove NOX from diesel exhaust. One possible method for removing oxides of nitrogen from the exhaust is the addition of a suitable reducing agent such as ammonia or hydrogen. Because of the low temperatures of diesel exhaust, it is not possible to do this without the use of a catalyst. Diesel Engine Trade-offs Experimental data establish conclusively that there are trade-offs between performance and emission control in diesel engines. The reasons for these trade-offs are known only in a qualitative fashion, as described above. However, it is well known that they do exist and that they are relatively independent of control technique. Most trade-off curves are approximately hyperbolic in shape, so that the first increment of control produces only small degradation of performance while later increments cause accelerating degradation of performance. Well-recognized examples of such trade-offs are those between NOX and fuel consumption, particulates and NOX, and hydrocarbons and NOX. However, there are other, less well established ones. For example, the use of exhaust gas recircu- lation (EGR) to control the formation of oxides of nitrogen is widely suspected of degrading durability and driveability, especially in turbocharged diesel engines. There is evidence to suggest that the particulates and other com- bustion products in EGR gases may adversely influence durability. However, the evidence is not adequate to establish reasonably valid trade-off curves.

24 Most heavy-duty diesel engines are turbocharged, and turbocharged engines have a response-time problem during transient operation because of turbocharger inertia. Figure l presents data illustrating the driveability problem without EGR; it is worse with EGR. As can be seen in figure l, when fuel flow is increased instantaneously (solid curve, "fuel pump rack position") inertia prevents immediate turbocharger response. This results in inadequate air flow and excessive smoke (solid curve, "smoke opacity"). If the fuel flow is forced to increase more slowly in spite of the driver's desire for immediate power, smoke can be controlled (dotted curve), but at the expense of drive- ability, since the engine reaches full speed several seconds later than it otherwise would. This problem is compounded during rapid vehicle accelerations, when part of the air entering the turbocharger is replaced by recycled exhaust. One could cut off EGR during acceleration, of course, but this would reduce the control of NOX emissions. CONTROL TECHNIQUES Retarding Injection Timing Retarding the timing of the injection of fuel is a universally used means of reducing NOX emissions. Note, however, that fuel consumption and particulate and hydrocarbon emissions tend to increase with injection retard. Increased injection pressures may mitigate these effects, depending on existing levels of injection pressure and emissions. Most injection systems currently vary injection timing with engine load and/or speed. Much work is being done on electronic control of injection timing, to permit more flexible and precise control. The degree of improvement to be expected from electronic controls depends on the sophistication of the mechanical controls they would replace. Data presented in a National Research Council (forthcoming) report on light- duty diesel engines show a l0- to l5-percent improvement in fuel economy due to improved injection systems. The committee contacted technical experts involved in developing these systems, who reported improvements in this range when the latest generation of mechanical controls were replaced by electronic systems. The major advantage of electronic controllers, in terms of emission control, is that they permit control system components to be programmed based on instantaneous sensing of engine variables. This allows control devices to operate fully when they have the most beneficial effect on emissions and the least deleterious effect on other aspects of engine performance. Under less advantageous conditions, the control devices can be less fully used, so that the engine can meet a given emission standard with relatively small impairment of performance. Varying the Shape of the Rate-of-Injection Curve The shape of the rate-of-injection curve affects both instantaneous fuel

25 fa o — SI - & TJ to UI 8s I! . Standard • With Fuel Controller (Aneroid) rr LU _ 3 2.00 1.75 1.50 1.25 1.00 100 80 80 40 20 2,500 2,250 2,000 1,750 1,500 1,250 1,000 110,000 90,000 70,000 S O X 5 o .^ « 2 50,000 Cu Hi 30,000 100 80 60 40 20 0. 3 O u. o_ I I I I J 8 10 246 TIME (s) Figure l Acceleration of a turbocharged diesel engine against a steady load, showing the trade-off between responses and exhaust smoke (experimental data) (Watson, l98l)

26 injection pressure (and thus droplet size) and the distribution and mixing of the fuel with the air. Proposed electronic controls for the fuel injection system are intended to control mainly injection timing. However, one can envision systems for varying the shape of the rate-of-injection curve, and laboratory systems to do this have been built. Too little data have been reported, though, to allow an estimate of the benefits in emissions control. Developing a reasonably simple and inexpensive system that can appropriately vary the shape of the rate-of-injection curve during transient operation is a formidable task. Turbocharging and Charge Cooling Almost all new heavy-duty diesel engines are turbocharged. Thus, any re- duction in emissions due to turbocharging must come from cooling the charge air rather than from turbocharging itself. In general, reducing charge temperature decreases NOx formation by reducing combustion temperatures. Modifying Engine Designs While no dramatic breakthroughs are expected, some changes in design details can reduce emissions. For example, decreasing the volume of the injection nozzle sac has already substantially decreased hydrocarbon emis- sions. Higher fuel injection pressures have also been beneficial, although there is some evidence that there may be an optimal injection pressure for a given NOX emission level, so that further increases in pressures may not monotonically reduce emissions. Exhaust Gas Recirculation (EGR) Recirculating exhaust gases through the engine air intake is a well-estab- lished NOX reduction technique in spark-ignition engines. It can also be used in diesel engines. In four-stroke-cycle diesel engines it may also reduce hydrocarbon emissions. However, EGR typically increases emissions of parti- culates and smoke, slightly increases fuel consumption, and, as indicated in the earlier discussion of trade-offs, decreases driveability and possibly durability. Another problem with EGR is fouling of turbochargers and charge cooling equipment. Catalytic Controls The presence of oxygen in the exhaust gases is necessary if an oxidation catalyst is to destroy unburned or partially burned fuel. Since diesel exhaust always contains excess oxygen, diesel oxidation catalysts never suffer from a lack of oxygen. The main concerns about oxidation catalysts have to do with contamination, primarily by particulates, and with overheating. Diesel exhaust temperatures decrease rapidly with load. During light-load operation, the

27 catalyst may become partially plugged; it can then overheat when the load is increased and the particulate material burns. Reduction catalysts require the addition of some fluid (typically ammonia or methane) to combine with the NOX in the exhaust. The effectiveness of the process is temperature-sensitive, and thus load-sensitive. The process may cause undue ammonia emissions, and it requires the user to take action (in maintaining supplies of the additive) with no direct benefit in return. There are substantial questions about catalyst durability, and the design of an adequate system is a serious challenge. Water Injection Water, injected in emulsion with the diesel fuel, is known to reduce NOX emissions with little increase (and possibly some decrease) in hydrocarbon and particulate emissions or fuel consumption. There is concern about corrosive effects, storage and use of water under cold conditions, and, again, the requirement of action by the user with no direct benefit in return. Turbocompounding Turbocompounding (the placing of a turbine, connected to the crankshaft, in the engine exhaust so that more power may be obtained from the engine) primarily affects engine efficiency and not in-cylinder emissions. Thus, any effect on emissions comes mainly from the production of more work at a fixed yield of pollutants. The main motive for tubocompounding is the reduction of fuel consumption rather than reductions in emissions. Using a thermally insulated engine (see below) or raising turbocharger efficiency increases the attractiveness of turbocompounding. Insulating Engines Thermally In an insulated engine, energy rejected from the cylinder exits via the exhaust rather than the cooling system. This extra energy in the exhaust is at a higher temperature than the cooling air. This increases the at- tractiveness of turbocompounding as well as reducing cooling losses. The use of insulated engines would require development of new engine materials as well as new lubricants. It appears that this technique would not increase, and might slightly decrease, in-cylinder NOX production, and that it would reduce the production of particulates. Trapping Particulates Filtering and/or aerodynamic trapping of particulates in the exhaust is undergoing intensive development for use in light-duty diesel-powered vehicles. The amounts of particulates produced require either frequent

28 cleaning of the trap or some means of oxidizing participates in the trap. Otherwise the collected particulates would increase engine back-pressure, thus increasing particulate production. Oxidation in the trap would require a tight control of the rate of combustion (probably by controlling the supply of oxygen) so that the trap will not be destroyed by heat. Sophisti- cated interactive controls, probably designed to control oxidation rates by throttling air entering the engine, would be required. Using Alternative Fuels Data from light-duty vehicles suggests that a specially tailored diesel fuel may reduce particulate and gaseous emissions. The same may be true of heavy- duty engines. Methanol and ethanol, which have very low cetane numbers, show particulate emissions significantly lower than those from diesel fuel when burned in diesel engines. If the use of methanol and ethanol becomes widespread, a substantially new engine would have to be developed. Use of Alternative Engines In laboratory tests, steady-flow engines, such as gas turbines and Stirling engines, show lower emissions but higher fuel consumption than diesel engines. l986 AVAILABILITY OF CONTROL TECHNIQUES NOX control techniques likely to be available in time to meet l986 emis- sion standards include variable injection timing and pressure, charge cooling, and possibly exhaust gas recirculation. It is not clear whether electronic control of injection timing will be available by l986; some systems may be on the market by that date, but whether they will have been available long enough to meet engine development, testing, and production lead times is uncertain. Charge cooling using ambient air will be available, since it requires mainly straightforward hardware changes. The outlook for exhaust gas recirculation is not clear; it should be feasible to recirculate small amounts of exhaust gas in heavy-duty engines without significant losses of driveability and durability, especially if electronic control systems are available. It appears unlikely that substantial amounts of exhaust gas can be recycled satisfactorily in all types of heavy-duty diesel engines. The System Problem One must consider the entire engine-fuel system and all of its operating and emissions characteristics when judging the feasibility of meeting a given level of any pollutant. As indicated earlier, one emittant is often reduced at the expense of an increase in another. Thus, it is impossible to address the effect of regulating NOX emissions without also considering the permitted levels of other emittants. For example, in a given engine, if one fixes

29 the levels of particulate and hydrocarbon emissions and the permitted increase in fuel consumption, the level of NOX emissions is set within fairly narrow limits. One must consider the levels of all pollutants as well as acceptable changes in fuel economy, durability, and driveability. AVAILABLE DATA Most of the emissions data on heavy—duty diesel engines comes from the l3-mode steady-state test cycle. (See Appendix C.) The source of this data is engine certification testing by EPA and the engine manufacturers. More limited emission data are available from the new transient test procedure. The sources of this transient cycle data are EPA and some, though not all, heavy-duty engine manufacturers. Generally accepted correlations between the two test procedures have not yet been developed. Figures 2 and 3 show steady-state emission data for hydrocarbons and NOX plotted against transient emission data for the same pollutants. (The correlation for carbon monoxide is not especially important for diesel engines, whose carbon monoxide emissions are low.) The NOX data show an approximate one-to-one correlation between the two test procedures. However, there is considerable scatter, and on average the transient cycle NOX emissions are about l0 percent lower than those from the steady-state cycle. Furthermore, data for low NOX levels are not available. Much more information is available on light-duty diesel engines. How- ever, such data cannot be used directly to obtain quantitative estimates of the emissions from heavy-duty engines. Heavy—duty diesels are generally open-chamber, turbocharged engines; most light-duty diesels are prechamber engines without turbochargers. In addition, the emission test cycles and regulatory definitions of useful life are different. Essentially all of the data on the effectiveness of the emission control technologies discussed in the previous section were taken using the steady- state test procedure. Such data are useful for engineering analyses of promising control options, but there is no established correlation with results from the new transient test procedure, especially at low NOX and hydrocarbon emission levels. With the new transient test procedure, preliminary data suggest signifi- cant variation in results from laboratory to laboratory. Table l2 illustrates the magnitude of this problem with data from three different engines tested in two laboratories. While NOX results differ by about l5 percent from laboratory to laboratory, hydrocarbon and particulate results differ by as much as a factor of two. NOX emission standards for heavy-duty diesels have been in effect since l974. (See Table 7.) Uncontrolled NQx emissions from heavy-duty engines manu- factured before l974 ranged from about ll to 20 g/bhp-h. In l974 combined hydro- carbon and NOX emissions were reduced below l6 g/bhp-h, and in l979 they were reduced further to below l0 g/bhp-h nationwide and 7.5 g/bhp-h in California.

30 2.5 |0 2.0 ^o O w 5$ LU ££ OQ O O £T QC CO Q < ^ UJ 12 1.0 0.5 0.5 1.0 1.5 2.0 2.5 HYDROCARBON EMISSIONS (g/bhp-h) MEASURED ON STEADY-STATE CYCLE Figure 2 Correlation of hydrocarbon emission data taken with the transient and steady-state test cycles

3l 12 •• 11 - 10 * LLJ 9 • U • • >- • m u 8 . ^_ tf< z .c — 7 • •*• ^| C/3 z 6 - • • iT • ••** * $2 5 •* * .* - cc 9 ^ 4 _ • x< O UJ z s 3 — 2 - 1 - n i i i i i i i i i i t i Figure 3 12 NOX EMISSIONS (g/bhp-h) MEASURED ON STEADY-STATE CYCLE Correlation of NOX emission data taken with the transient and steady-state test cycles

32 CO iH 3 O 4-1 oo o in ^ oo P** ^N ^4 r-* No in co oN co f, CO • • • • • • J2 Pb 0 O 0 O O O Jd • M oo Ol i-i CO •0 H C -H n CO O1 0 X oo 4-1 > J2 O 3 01 (H C cfl O co co co "*O oN *^ E? CO u x m m CM CM CM 1-1 Cd i C 3 pf< o C 4-1 ^H CO ^ CO CO Ol M! M H p CO 01 G H fe! Ol 4-1 O *o CM 00 r-| l-^ O 4-1 •H CO 01 •O >-i ts CO oo vo r^* oo oo co i CO 3 EC 0 O O O O CM rH a i CO CO 5 o w 01 01 i 0 1-1 o m r-~ o< CM co o X G K*~V O ^^ 00 lO "^ ^D ^D O u m •-I •H tj 4j Cd C! 4J 01 G •H cu a CO 01 cO t-i k i C a H o •H en G 4-1 li •H CO O 01 C O 01 O H J h i-H CM tH CM r-< CM 3 4-1 O 0) cd ,J 1 •rt M CO CM 01 C 1H 8 3 •-I CM CO i 00 1 H u CA

33 The l980 California NOX standards are 5 g/bhp-h. Table l3 shows l979 EPA certification data for heavy-duty engines with NOx emissions levels as measured on the steady-state test. The average over 38 engine families was 7.7 g/bhp-h, with a range from 5.7 to 9.7. Table l4 shows l980 California NOX certification data from the steady-state test procedure. The average over l5 diesel engine families from five manufacturers was 4.8 g/bhp-h, with a range from 3.7 to 5.7. Note that 60 percent of the heavy-duty diesel engine families available in the other 49 states are unavailable in California. The Southwest Research Insitute is conducting a study for the EPA to determine baseline emissions levels of 24 different heavy-duty diesel engines manufactured in model years between l976 and l980. Results available to date are given in Table l5. Also, two different measurement techniques and data correction procedures were used for the transient test program. The labeling of some data in Table l5 as "questionable" was done by EPA and Southwest Research Institute. The ranges of these data are 3.8-l0.4 g/bhp-h on the steady-state cycle and 4.6-8.7 g/bhp-h on the transient test procedure. The limits on NOX emissions from current production heavy-duty engines can be summarized as follows: In the 49 states outside of California, emissions range from 6 to l0 g/bhp-h. In California, the range is 4 to 6 g/bhp-h. The fuel economy penalty of the stricter NOx emissions requirements in California is probably 3-6 percent. These NOX emissions levels from current production engines compare with uncontrolled diesel engine NOX emissions of ll-20 g/bhp-h; that is, Federal standards have cut NOX emissions by about 20—50 percent, and California standards by about 40-75 percent. Uncon- trolled NOX emissions from heavy-duty diesel engines were one and one-half to three times the uncontrolled gasoline engine emissions baseline of 6.8 g/bhp-h recently established by EPA. (See Chapter 3.) Current controlled diesel emissions are comparable to this uncontrolled gasoline engine baseline level. Note that these summary numbers represent typical values and that the range in values is wide. Different diesel engine types, and different engine designs, can have significantly different NOX emission levels. ANALYSIS OF DIESEL NOX EMISSION CONTROL DATA A wide variety of diesel engines is produced for the heavy-duty vehicle market. Five domestic manufacturers, with 39 engine families, share over 90 percent of this market. Three-fourths of these engines are turbocharged , and 40 percent of the turbocharged engines are aftercooled. Because the available data on NC^ emission control technology are limited to only a few of these many engine types and designs, it is possible only in general terms to discuss the technology for reducing NOX emissions. An approximate quantitative assess- ment of the technology is possible, but the limitations of the data base and the diversity of diesel engines in production must be borne in mind. Different control approaches are unlikely to work equally well on all diesel engines. All manufacturers expect to use turbochargers on almost all engines, to use intercooling and/or af tercooling , and to use improved fuel injection

34 O CO o a1 H H 0l 9> •H 0 en « s S ^ ^ ^ !i^ « C m "• ^ •* Sout hwes t Sout hwes t 79 Research Research rtlf 1 cat Ion Institute Institute -Mode 13-Mode Transient NO' NON- u «••> s s ^ * * * « ^ e ^ s ^ s s s s r- -O in M i~« « is* tM ^H •* n -; -; -; -; 5 tD ^T o-i •fi 0* -H ^ 0 i ^ tn r- ON «n in -o «n r- «o o sO .J _] .J _J .j .J a. bu 0.0.0. a. a. a. a. a. a, M < w w in 2 XI I p p p p p p b, a. a. b. b, h U. 1 I H P P P P P < w w < < < £0X0.0.0.0.0. |BEU.bub.b, U.Ii.U,U, bnO.b.WMMMM IU«<<<< <<<< o c* *o o o o o o in o «-i o *5 1 ' ' o o o "** ^ o o" o «« r- "> o ^ M SooS m^ff.e-1-^ |ir,(^p^ | So-»olS *H •-i CM CO 35 CO w in 1 5 g. oopoooooo oo o ooooo §00 ooooo ooooooooooo oo ooooo ooooooooooo AM OOOOOOOOO OO O O O O o O 3; *• "- 2 - ' • - ' '''''' 1 • • :.' a s-a So { technlq tthods- a H _if«i«o<^isOu-ii£><-iN «i tn m Q o m m «n m c in c O>r^9*ror- r^i*iS?SoSf-i^S^ 0 S 71 opression clo < -H — 1 air/fuel contt S • emission -H O .J 1 W -n 0 • M u •88 o *i u* < • u C «M 0 *4 k i « ? nglne Turbo- Inter- After- Inj ycle charger cooled cooled Tim XXX X X X XXXX X X X X X X X XX X X X X X X on; FM - fuel aodulator; TD - throttle de crankcase ventilation; SPL • saoke puff larenthesls- dilute conti'ous sam'e. W1-, n la Riven baaed upon the obaerved averagi sir X X X X X X X XXXXX XXXX ".! 8 •*.". - in o *O •? r*. •* i IN (M O m f-^ O —* • ui o t o O in I si it'S Kami- Engl faccurer Engine F *O-HsO PzHPUl OOOOOOOOOOO H • M oo w o — — *EGR • exhaust gi ratio control; •o a c w « u-i i^- r- r- r- CT> r•. «7* e <UW<UBQ a -o x S S" • 3-0 00 w * cAv«rage. - ; - js e u vi m -o UUU UUUUU wWWWWWWWWWW Jl J< Jt Jrf J^ *^ oSoSoSuSu ouooououou - ~ ~ ac r r r 5 uuuu^uuuuuu

35 TABLE l4 l980 NOX Levels for Diesel Engines Certified for Sale in California (May l980) NOX l980 Control Certification Transient System NO* NC- b Manufacturer Engine Family X X Caterpillar 12 13 4.7 EGR 4.8 l4 5.2 l6-C 4.l Cummins 093G 3.7 092 4.7 Detroit Diesel 8V-92TA 4.9 6L-7lTA 5.2 8V-8.2 5.0 6V-92TA 4.9 International DT-466B 4.2 Harvester DTI-466B 4.9 Mack ll 5.2 10 5.7 12 4.l 4.6l (5.l5) (Average, 4.75) aObtained from telephone conversation with California Air Resources Board. "Bag-sampled and (in parentheses) dilute continuous integrated. SOURCE: U.S. Environmental Protection Agency (l980).

36 X: X! 'o M 5s e oo 01 3 01 rH Z H *4f CM CM CM CM CM CM CM CM r-j CM O 4 J 0 O 0O ON 00 rH OD X CO •.o rH NO o* cn oo oo 01 o co >o r** ^^ o CM in NO OJ rH H O • • • • • • • • • 01 Z -.•J- ON oo m t** ^ r"* r** in r*>» NO U 5 01 C/J | B -rl L 0 X T3 J3 O ,_! ON -a* o cn CM cn in CM o 0 o cd k i B ^-f- rH in CO *-3- CM r~^ ON CM NO co oo O1 CO O • t • • • • • • ON 4-1 0 X cn CM CM cn vO 00 rH rH rH NO ,H rH C/] M I— 1 01 e 1 B m B 01 O O 3 1-5 •H l< -0 o <T NO CM ^7 ON 00 ON O O NO CJ •O I- r** CM mc n oo >o ON oo N0 O in X CO O OD a- o rH 0 00 00 00 O CM o 01 OD 3 i rH k i CO CO 1 k i OD 3 01 41 Z H _ Cn B 01 3 U-i r*~ -S- CM CM CM CM CM CM CM B CM a> Z 0 •H B O ^% •H X: 1 .fl CO 1 U OD CO a •H 01 ON O ^4* ON ON ON 00 in NO 01 00 I 23 4J 4J rl M CO rH co «H U in bo DM 3 O o o o o o o o o O O 01 \-x 01 B U-I iH 01 01 rH rH ^^ ^^ **"•* X™\ ''"N S~^ ^N XN XN •5 ^— N 01 O <j- rH CM •— ( c n CTN CM co NO m 0 fe CO oo m rH oN c n **o GO oo in o CM CO u X • • • • • • • CO O m oo r** *«j >o 10 to l0 in m W z t^ S^X %»-* V^ V-r* S—' \-r- -^r- *^ n t-H OD 3 CO 01 01 OD H « 01 01 T» NO •H 4-1 •O O B G iH fN 01 O X O, 01 •H .0 o o C/N lO O* P"* lO O^ ^ 00 o NO 4J OD li B 00 ^^ rH 10 r** cn *^ r^* 4J |^ 3 B cd O • • • • • • • • 0t • i C J 3= CO -3- CO CM cn ~* 1-H rH rH NO •H k i tJC 4J H C OD fej ei 4J X B 0 (H o 01 k i -Q oo oo oo r~« CM o vo r** CO 0 rl ^ •0 >-i pT) vO r* co 10 cn rH o 00 00 A •H >: CO • 01 O 33 0 CO O O O fH rH rH rH O o OD 41 OD 01 h It 3 H H , BB 4J rH CO OD at rH Z CM -a- o ^J p,j 01 rH ON o m ^ 1 •rl Du SI 00 > B rH rH rH PH H 0 *G H H 4J rl 1 * 1 3 3 33 N 3 IH 01 i <!lo Ebb. PbU, 3 o 3 in 5 01 4J | O at o O OUcn rHCM rHCM 1 4k =lh -it * M U-I k i •H 3 i m Z oo ao Mt> OO ON U ON rH M 01 ON rH 01 r~ o r~ O r- f^ O r- cd -H ON O UH r^ CM a ON CM O* rH ON a rH CO rH CO rH rH Z rH ^^ rH ON iH si •H H 1 , .... . , , , rH CM cn <r m >o r«* 00 O<

37 k i 01 CO ,O -LJ | 8 CM CM CM CM CM CM CM CM CM CM ^fj NO CO GO O O rH in OO -* 0 o f-^ fyv ^H CM CM «0 ON I-. CM •* • « ^ ^ f • • • 55 CO * ** ^ in r- m co oo o rH 01 •o B -H 0 X •0 O CO _^J- CM f^ ^£) 0-1 ^ CM O O ki C 00 vO 00 O «o NO ^•j ON at o ^ « • • U S rH iH I-t CO rH CM CM CM en CM • S 0 0 if~) 1— ( vO Ol r» 00 «o CO <t ^-r •O P CO |^ CM 1-H «n %o NO vO CO x a 9 • • 9 • • « • • 33 0 o O O O o O o o 0 0 3 - h 4 J 01 CO • H 0 CM CM CM CM CM CM CM CM CM CM w 3 UH Ol JJ 55 0 H S O CO f^ CM O en £i •H 01 CO iH r.. <r OO ON 4J 4J vO co «n u~l fO in m co m X IH cd « • 9 • • • • • iH 0 O O O O o O O 0 0 4 J p* 3 B Ol rH St NO 00 r-H CM 00 ^, I LO • • • • m m m m cd in en TS iH o f> o o* CM OO X rH en •H <j- <r •^ NO PO CM • • O • 9 % • • • • Z m * > CO in r-~ m 00 r^ oo 01 •O e i-i 0 X J3 0 CO «o «n rH CM O ^ en 00 rH r- B NO <n v0 CM 0 NO 0 00 CO •* 33 rH iH ** CM rH CM CM CM CM , S 0 0 T3 l-j ON <j- NO o OO CM <j- vO 00 rH X O CO en NO ^o <J- NO OO r. in oo 9 • t • • • • 93 0 0 H I-l o o O o O o o >o 3 QJ y-~\ 01 vD m to «o H 2 H 1 {H I NO CM O rH X C/3 0 CO vO 0 CO O o o a m 2 1 " J S CO H H ID > 01 S •H CO 0 NO T-) 00 -H-^, •H rH rH a e g <U S IH ^ • L^ C_f B cj ^J < -H < S H 1 1 1 rH iH •I Q 30 53 3-^S O rH U rH •H u z H ° S, O C J ON CO C J 0 m M O ON "° r« ON g ON a r^ at * :a * ^ « S So oo r^ co CM rH ON pL| ON 1*1 cr> CQ ON ON | ^ O** ON E~" ON ffi V o» ON O ^fe ^it rH ^— S rH •*** rH C J rH rH ^-^ rH rH 5! rH 55 c rH rH ON DO . . . 9 , . . . . . B 0 H CM en -^ m vO r-. oo 0% W rH iH i-l i-i I-l iH iH rH rH rH

38 a (u i-H XI o cfl i- 0 I-i M 01 00 >, cd B k i 4-1 73 O 0 k i 3 C J 4J 4_) 01 rH O •2 *-" B oo 3 01 Z H CO 1-1 3 Q rH 01 k i 0 I CO B 3 . a 41 O >, B O O rH U) OJ 4-1 O > C — X CO 01 3 CO CO O 4J 4J O >COi-HiHi-l 01 tf 3 01 cd MiH BC i-H CO Z xx: > oi -H -H 01 o 4J os & e • o H Z B 0 0 B Pu 01 01 2 iH B ki O 4J 01 rH H Cd O IM -r- 1- CO 4J 73 -H CO k i i-l 4J i-l B -H CO D. . 01 4J CO O < CO O X •BO01rH O 3 01 O 73 <0 B B ^ k i B co-Horns 1H u-i o • O CO rH kl 73 u x 01 73 rH « rH OD CO 3 CO B 01 CT> cd 1-1 • B O 01 •H to I4-i r- o SB CO 4J bO O O ON 01 fd O 4 J co B rH r-4 Oi CO B 01 X (d O kl 01 rH 0) rH ^l-t 1 B co LJ o> 4-i PI • cd-H a X> CO- 4-1 01 B -H B 0 0 4IOrH/^ rHB -H O 01 kl XI 4-l0lcdbO 3cd 3 £ 73 kl •H 6 IH >« a oo if • •H X cd < CO I-J -H iH 01 CJ 14-1 B CO jd 4-1 • 4-i o O X 0 CO • 4-1 -H I-l I CO O U01 coO3cd h to «H 3 01 bOrH4->X>41 a. X -HO 3 B3O4lrH X> CO 4JiH rH •HO01bS 73 01 COM-I rH0lkiki B- iH BI4-| CO a i-i ui oi ra GO ki MO g 0 x: 73 rH 01 CO cd 73 o 4-1 01 col > x: - kl CO |_f co0 co SON o ox bOXBO Or~ o kiO 01 01 bObOUiHO. 0il C4 cd C cdcd-HO coca -H 0101 t s XIX>7373ki COCO 73 COM1 •H 01 O, O Q 01 < bOi-i B 1-1 co u Bi Cf r-l 9 IH IH <4-l B B O u iH<x:-HO O* 01 4JO M 1X> 8 co o o 3 co 1-1 3Cd0101 4JZ 0* 014-1 • a, o B 01 9 o O CO 4-1 73bObOCO-H 414-1 CO JB H 01BB4J Bki X> u u B-H-HkiO 01O 3 SO 01 rH •H 01 4-1 4-1 g, ki cd O CO rH •H •Hi-ICOOE ki ex co o i-i CU 3 u BCX34J 301 COPu O •H C a o 0 co a: o I 01Cd73COT3 CO U BrH 4JCO01M-I1-I 0>i-l OcO CO cd 0) 4-1 X K tO 01 k44J 01 X 7373COBO 0 3-'*-tB 01 rH O C I-i -H 739 01 H o 014J3-HO OB O C B Z IJ CO CJ 4-1 B 14H £ 60 rH O B • B 0lklrHO O BrHiHOO 01 ^bOcO0lB 0101 -HO co kioo 0lOklO OH 4J 14J CO-HCn •H 01 OO4JkiX> iH IJ i-t>rH CO B -H 73 IB01Okl 4J< 0 rH B B B 0 X rH-Hkiocd o a, a. rH X> td» CO XI O ki B CO O 01 O COCd 01-H3 r- I-i H • • P M kl k i P • • rH 0 33 01 4J O 73 OLI : CO B3COZB » 4-1 co CO • ki XI B i-t i-i 4J td 73co OD014J901 ClOlHrHbp 01B CO Ml B) XI B733C3B 73O i-IB73 »B 01 (dcO-HO B-H cd B01 i § CO CO01coX> 0J4J 01X!i-l OO ki 3 ki 3 k i B iH OO cd -HO) OO cd B73 B B 4J O 0 ki X> 73 kl U-I 3 X73 O OB i-l 01 kl CO BO01O oo cox: o o- 01 X cd CO -H Z 4-1 kl 01O 4JI4-llH> 3 PC O 014-JCO73 Od 73 BX: H CM 3 B 4J rH >, 4-1 0lbObOSO cd k i k i 01 rH rHOB3X: 4J C bOB B B-H COO01O 4H01 B-H-H B£ CM H cd -~x cd 1 >-HrH>> cd-H C04J73 6 • rH iH C»> r-I x-x ,J bo co to 4J kiC O • x:cd B o - - CD r^ i-l iH rH PI in 1 1 i-l -H IX, CD rJ CX >> firS > vD tl i-H El rH N XB BOS OB CO Oki 0) ki O -H ra -o :cd 4-1 O O.rHO Z CO ki 01 | • ki CO 01O. COX> 4J 01 i-l 01 -H 4J 4J 01>4J01 COOUM4! 3 4J kl >•, tO klH 73 >kl CO • H B h < < we u B 3 B T3 rH 3: ro -H o. n 01 X!0 cd O o o 1 O o cd cd cd cd 41 01 Ol rH 01 B 0lrH B73M-IB Q O O tli U [X, Q •H kl CO OO X! kl 7373 X> td -H BB 003 1 01 01B CO4J 01M< c ro - rH ro a. 4-1 LJ B cdai > in B r^ r~— r^ oo r** oo t** ra M oo co cd o 73 c i-l •H O^ O*> OS O CT> O O"l -H —I rH CM -H tN -H kl01 OOON01 rHCO iH O 4 J *• a 01 •H 4J e oj ai rH e s4t u 0 1-1 td u BB -H OB ID COCO 4-1 U rHO-HiH014J rH -H 01 0100 B pd bo r-jcobobox: cd bo 3 X:-H 01 a B O rH CM PI <t <^-IBB4Jcfl CJB O- HS H O H td CM CM CM CM CM CO CO W W ^ CO XIEd O 7301 01 CO

39 systems and more sophisticated timing controls. Many manufacturers anticipate that electronic engine controls will be available in the l980's. There is con- siderable debate in the industry about the feasibility of using exhaust gas recirculation (EGR) to reduce NOX emissions substantially. Very limited data are available on the effect of EGR on all aspects of heavy-duty engine operation. High EGR rates have been shown to reduce NOX emissions by as much as 80 percent, but also to result in much higher particulate and hydro- carbon emissions and higher fuel consumption. Lower EGR rates, reducing NOX emissions by perhaps 50 percent, have been shown to reduce hydrocarbon emissions also, but at the expense of a 50-percent increase in particulate emissions. There is much concern about the effect of EGR on engine durability. The duty-cycle load requirements of heavy-duty diesel engines are much more demanding than those of light-duty diesel engines (in which EGR is expected to be used widely). Data from U.S. engine manufacturers with various NOX emissions control systems that are feasible for production by the mid-l980s are summarized in Figures 4-6. These figures show the relations between different levels of NOX emission control and (respectively) fuel consumption, particulate emissions, and hydrocarbon emissions. These trends have been obtained from a relatively small number of individual engines. The data come from engines at low mileage, and they thus do not take into account deterioration in emissions control with increasing use. The upper edge of each band corresponds to simpler emission control systems, generally based on engines of current technology using retarded injection timing as the main control technique. The lower edge of the band corresponds to more sophisticated emission control systems, not yet in production. These include all the techniques listed in this chapter under the heading "l986 Availability of Control Techniques," including exhaust gas recirculation. Figure 4 shows the fuel consumption changes associated with different levels of NOX control. The fuel consumption data are for several different engines and control systems. The baseline fuel consumption for each engine and control system is the fuel consumption at current NOX emissions levels of 8 g/bhp-h. The baseline fuel consumption will be different for different engines. A new engine, of a different design, could have both lower fuel consumption and lower NOX emissions than another engine. The data, from the l3-mode steady-state test procedure, come from several engine manu- facturers and include both four-stroke-cycle and two-stroke-cycle engines. Figure 4 shows that reducing NOx emission levels below about 8 g/bhp-h results in fuel economy penalties that increase as NOX levels are lowered. For example, technology under development shows a fuel consumption penalty of about 4 percent at NOX emissions levels of 5 g/bhp-h, relative to the same technology at NOX levels of 8 g/bhp-h. Such an engine might have both lower NOX emissions and lower fuel consumption than an engine of less advanced technology, but data available to this committee show that even such advanced engines suffer increases in fuel consumption when modified to reduce NOX emissions. Figure 5 shows that particulate emissions (measured on the l3-mode test cycle) also increase as NOX emissions are reduced, reaching a level of about l

40 O 130 120 UI u. UI I 110 UI CO •D U. O LU 100 < 01 CJ cc Simplest Control Systems • (Current Technology) Most Complex Control Systems (Advanced Technology) 10 NOX EMISSIONS (g/bhp-h), MEASURED ON Figure 4 STEADY-STATE CYCLE Fuel consumption vs emission level

4l •gS IF 2 55 h- O ^2 _ _/ c < 0 Ja A Production Engines • Prototype Engines 2 4 6 8 NOX EMISSIONS (g/bhp-h) AS MEASURED ON TRANSIENT TEST CYCLE TO Figure 5 Particulate emission level vs NOX emission level

42 O Transient Cycle NOX EMISSIONS (g/bhp-h) Figure 6 Hydrocarbon emission level vs NOx emission level. Solid circles represent data taken with the transient test cycle, and open circj.es represent data taken with the steady-state test cycle.

43 g/bhp-h at NOX emission levels of 5 g/bhp-h. Particulate emission levels of the current 49-state engine are about 0.6 g/bhp-h. These particulate emissions are engine-out emissions from engines and emission control systems that do not include exhaust particulate traps. Whether particulate traps, which would provide additional reductions in particulate emissions, are feasible for l986 is unclear. The relationship between NOx and particulates is described in qualitative terms in the introduction to this chapter. The specific engine data in Figure 5 represent different engines and different operating conditions for the same engines. The quantitative differences indicated in the figure are consistent with the earlier qualitative description. Figure 6 demonstrates the difficulty of correlating hydrocarbon emission levels measured with the steady-state and transient test procedures; the data from the transient test are higher than those from the steady-state test. Hydrocarbon emissions rise as reductions in NOX emissions become substantial. The hydrocarbon emission levels of different engines and NOX emission control systems vary widely. For comparison, the l984 transient cycle standard for hydrocarbons is l.3 g/bhp-h. Two other factors must be discussed before these data can be compared with possible emission standards. These are the Selective Enforcement Audit (SEA) requirements (with a specified Acceptable Quality Limit, or AQL) and the deterioration in emission control over the life of the engine. EPA originally proposed that a l0-percent AQL be used for the SEA of new engines. (See Chapter l.) A new EPA proposal is a 40-percent AQL. The audit requirement, plus appropriate allowance for deterioration, requires low-mileage emission targets for individual engines to be 20-40 percent below the value of the standard. (For example, low-mileage targets for l984 model year engines are about 0.8 and 8.5 g/bhp-h for hydrocarbons and NOX, respectively, in order to meet standards of l.3 and l0.7 g/bhp-h.) Since data are available on neither engine variability at lower NOX, hydrocarbon, and particulate levels for advanced control systems nor deterioration of emission control over the useful life of the engine, only rough estimates of low-mileage targets in relation to possible standards can be made. The above data have been used by the committee in estimating the fuel consumption penalties, and the increases in HC and particulate emissions that result from different levels of NOX emission controls. Table l6 shows the estimated hydrocarbon and particulate emissions levels and fuel con- sumption penalties that should be achievable in the mid-l980s at various NOX emission levels. Only low-mileage targets are given in the table; insufficient information is available to estimate the necessary allowance for deterioration and variability to determine the appropriate standard. The values shown correspond to the lower one-third of the bands shown in Figures 4-6, since only the most promising technology will be used in production. However, it is not reasonable to assume that all production engines can reach the values at the lower edges of the bands, owing to the practical requirement that radically new technology be phased in over a period of several years. Assuming appropriate deterioration factors and margins to accommodate engine-to-engine variability (based on data from

44 current production engines), the standards that could be met would be between l.2 and l.4 times the low-mileage values given in Table l6. SUMMARY NOX emission levels from current heavy-duty production diesel engines sold in all states except California are about half the emission levels of uncontrolled diesel engines. Diesels sold in California have NOX emission levels about one-third the uncontrolled levels. The uncontrolled diesel NOX baseline is one and one-half to three times as high as the uncontrolled heavy-duty gasoline engine baseline of 6.8 g/bhp-h. The available data for evaluating the potential of NOX control technology to achieve levels substantially lower than current levels are limited. Only approx- imate engineering estimates can be made of this potential. Trade-offs between NOX control and hydrocarbon and particulate control, fuel consumption, and driveability are encountered as NOX standards are made more stringent. Only modest additional NOX control can be achieved without fuel economy penalties at lower hydrocarbon and particulate emission levels than those currently being achieved. The more promising emission control techniques being developed show that low-mileage NOX emission levels of 6 g/bhp-h on the transient test cycle can be achieved with engine-out particulates emissions of 0.5-0.7 g/bhp-h and hydro- carbon emissions of 0.7-l.4 g/bhp-h, and with a fuel consumption penalty of 2.5-4 percent. Low-mileage NOX levels of 4 g/bhp-h can be achieved with higher particulate and hydrocarbon emission levels and with a fuel consump- tion penalty of 7-l2 percent. The available data on engine-to-engine variability with the appropriate test procedure, and the data on deterioration in emission control over the engine's useful life, are inadequate to relate NOx, hydrocarbon, and partic- ulate emission levels of individual engines at low mileage to appropriate levels for the standards. The NOX standards corresponding to the low-mileage emission levels in the preceding paragraph would be l.2 to l.4 times the low-mileage levels.

45 <u * ii CO 3 C T> O 01 V •H 60 CJ CO tfl O CO 01 t-i I CO 0*0 4-1 O H cd hJ i-H v^ 4J 3 C o c 0 i-l O TH 4J 1H CO k i 4J C cd cx cd P^ S Vj 3 H •O CO C C 01 Cd O J2 U 4J CO C i-H >, O (U XI •H 3 CD fc T3 CO 0 •H T3 I-l B C 3 u cd co cd o co c 1H cd cu co 0 co •> S 1H co 4-1 g rH cu u cu c cu CO O hJ 4H X> IM ki C 0 cd o 1 o 1-1 01 O CO TJ I-l CO Cd T3 -H M K _• H EC W M 4J CO 4-1 3 cu c 4-1 PM cu cu C 4-1 CO o CO • 0 •H 1 4J 4-1 4-1 r*t C CO. CX 13 Q1 g s cd iH .,-j 3 cu cd 4-J a 4J > CO 0 CO -H 3 0 u cu cr o •ft O ,c cu •H I-H 1 <N CM 4-1 4-1 1 cu in I-H I SI *o CO • I m •H 0 CM t-^ rH 1* 3 « 4-1 CO 0 cu cd CO ki M 4J cd cu •H c cd 1 o 0 •H CO & cx co O e cu 4J 3 4J co 0 C 0 iH 0 S t^ O 4J cd CO oo -3* r^ i-H 01 •H L C • • • /— \ cu -c cd 0 O O rH iH & 3 4J > U .0 III ~*~* 4-4 a 2 ce o •O M vO r- OO T3 CO C 0 01 0 1 0 CD O cd cx at M CX co 3 M cd c cx •"-*. o cd bO •H CO \_ ^ CO -H L CO in r>* o CO X •H 4-1 • • • s~^ a> U -H O> O O iH J3 B 4-i C 4J 4J 1 1 1 «~s cu i-i O M tO *j* lO vO o O •H cd iH • • • XU-I O co Pu 3 0 O O 0 4-1 CO 2 • " •^ cu • « g C M cj U 0 3 5 o cd cu 6 4J 0 _w cd o C Q oo vO -<r CM a k i & cd cx ,O

46 REFERENCES Henein, N. A. l976. "Analysis of Pollutant Formation and Fuel Economy and Control in Diesel Engines." Progress in Energy Combustion Science, l:l65-207. National Research Council. Forthcoming. Diesel Technology. Diesel Impacts Study Committee, Technology Panel. Washington, D.C.: National Academy Press. Society of Automotive Engineers. l980. Diesel Combustion and Emissions. Warrendale, Pa. (P-86) . l98l. Diesel Combustion and Emissions, Part II. Warrendale, Pa. (SP-484) U.S. Environmental Protection Agency. l980. "Draft Regulatory Analysis, Environmental Impact Statement and NOX Pollutant Specific Study for Proposed Gaseous Emission Regulations for l985 and Later Model Year Heavy-Duty Engines." Office of Mobile Source Air Pollution Control. Washington, D.C.: U.S. Environmental Protection Agency. Watson, N. l98l. "Transient Performance Simulation in Analysis of Turbocharged Diesel Engines." Warrendale, Pa.: Society of Automotive Engineers. (SAE Paper No. 8l0338.)