Swaminathan N., Bray K.N.C. (eds.) Turbulent Premixed Flames

Подождите немного. Документ загружается.

4.3 Application of Lean Flames in Stationary Gas Turbines 343

EV Outer Liner

EV Inner Liner

Cooling Air

HPT Vane

EV Outer Carrier

EV Inner Carrier

EV Burner

Figure 4.63. Alstom EnVironmental Burner in combustor cooled by external convection.

(Courtesy of Alstom Power)

uses turbulators to enhance mixing. There are many other designs extant. [225]

is an excellent 118-page report by the Electric Power Research Institute, Califor-

nia, which describes General Electric, Siemens Westinghouse, Mitsubishi, and

Alstom heavyweight DLN combustion systems and how to optimise their fuel

controls to minimise emissions and combustion instability. Dual fuel designs for

large engines rely on a degree of water or steam injection to limit NO

emission.

2. Aeroderivative engines

Rolls-Royce Trent and RB211 derivatives employ several radially mounted

tubular combustors, within each of which the fuel is admitted in two or three

axially displaced stages; see Fig. 4.58. In addition there is a centrally mounted

primary pilot injector to improve low power efﬁciency, lean stability, and com-

bustion noise. This design style arose from a desire to achieve three objectives:

a. to ﬁt the combustor into the available space between compressor outlet and

turbine inlet,

b. to simplify the fuel system and engine control system by limiting the number

of fuel stages to two or three (there are three stages in the industrial Trent

engine), and

c. to provide sufﬁcient residence time to consume CO, unavoidably produced

at some operating conditions because of the relatively small number of

fuelling stages.

The advantage of axial staging over parallel staging is that the primary stage

sufﬁciently preheats the gas entering the secondary stage to ensure efﬁcient

combustion at very lean secondary AFRs. The disadvantages are the combustor

length required and the difﬁcult aerodynamic problem of achieving a uniform

air feed to the fuel–air premixing ducts. A detailed discussion on the staged

mixture injection and combustion is provided in Subsection 5.1.1.

GE adopted a different strategy for their LM6000 engine, which is a derivative

of their CF6 aeroengine. The LM6000 combustor is a short, axial, annular com-

bustor that ﬁts into the same length as the aerocombustor, although it is almost

344 Lean Flames in Practice

Main Fuel

Burner

Igniter

Supplemental

Fuel Burner

Pilot Fuel

Burner

Figure 4.64. KHI DLN research combustor for engines of 8–18 MW output.

three times the depth of is its aeroengine counterpart. It contains 75 LPP fuel

injection modules in three concentric rows. Because of the very short burning

length available for CO consumption, the combustion temperature must be held

within very narrow limits – 1825 ±25 K – to achieve desired emissions. Above

1850 K, NO

is produced and below 1800 K there is insufﬁcient time to complete

the relatively slow oxidation of CO to CO

. To achieve the narow temperature

limits, 10 parallel fuel injection stages are employed. The compressor outlet dif-

fuser has multiple passages to achieve a high degree of diffusion in a very short

length.

3. Small engines – below 25 MW

Although legislation controlling emissions from these engines is not yet univer-

sal, considerable effort has been made to meet the requirements of the most

sensitive regions of the world. Several companies have experimented with cata-

lysts and ceramic material in this ﬁeld, which is less challenging than that of the

larger engines. The operating conditions are similar but the smaller size of the

components reduces the manufacturing difﬁculties.

Because of their small size, there is more opportunity for application of the re-

sults of research performed by institutions with test facilities of limited capacity.

For example, KHI developed tubular DLN combustors, shown in Fig. 4.64, which

achieve 15-ppmv NO

in engines, with a test facility capable of operating only

up to 0.3 MPa, 725 K air inlet temperature [226]. The mass ﬂow required to feed

just one of six tubular combustors at less than 20% of engine pressure is only 3%

of the full engine ﬂow. This combustor is axially staged with a central pilot to aid

stability and efﬁciency at part load. 15-ppmv NO

is claimed from 50% to full

load, together with less than 50-ppmv CO [226]. Such a rig could be operated

using compressor bleed from an an existing engine, rather than investing in an

expensive plant.

4.3 Application of Lean Flames in Stationary Gas Turbines 345

Air staging involving variable geometry, impractical in aeroengines, may be pos-

sible in stationary gas turbines. Air is sometimes bypassed to enrich combustion

to reduce CO at low power, but here we are referring to redistribution of the

airﬂow within the combustor to control the local equivalence ratio. A search will

show that this subject has been researched widely because it is theoretically very

attractive. Its implementation is rendered difﬁcult by the temperature at which

moving parts must operate reliably, and much of the research has been done by

testing variable-geometry components locked in various positions by welding, to

limit leakage. Fretting or seizure is obviously a problem where several thousand

hours of operation are required. Much more attention needs to be given to the

mechanical aspects of this problem, rather than expending effort in the relatively

easy task of demonstrating that low emissions can be achieved by this approach.

II. Liquid-burning systems: Much of what was said in Subsection 4.2.9 relating to

control for aircraft gas turbines is applicable here. The relaxation of space and

weight constraints may allow ﬂuidic control, vortex ampliﬁers, etc., to be employed

to modulate air ﬂow. Half the pressure loss across these devices, however, is taken for

control stability and is not therefore available for mixing. Auto-ignition, discussed

in Subsection 4.3.9, is a far greater hazard than with gaseous fuels.

An LPP system cannot operate with liquid fuel if the inlet air temperature is

below the fuel boiling point. This is not a single temperature but a range over which

different constituents in the fuel evaporate. At air inlet temperatures below 500 K,

it is likely that a signiﬁcant amount of number 2 diesel fuel will fail to evaporate and

may be deposited on the walls of the premixing duct, where it can form deposits and

create sites for auto-ignition and ﬂashback. A diffusion pilot ﬂame is thus necessary.

If this is turned off at full load, it needs to be purged to prevent blockage.

4.3.6 Lean Blowoff

There is a limiting FAR below which combustion cannot be sustained. With the

natural gas fuel, this is around 5% by volume – an equivalence ratio of about 0.5. For

liquid fuels it is around 0.4 equivalence ratio. To achieve very low NO

it is necessary

to operate close to lean blowout, and therefore if the gas ﬂow is to be reduced to

reduce power, then some injectors must be turned off. This is called fuel staging. This

problem makes it difﬁcult to sustain combustion during a generator emergency load

reduction. Pilot diffusion burners ease the problem greatly, but there is a signiﬁcant

penalty. The lean blowoff limit is insensitive to pressure but becomes richer as

inlet temperature falls. As the lean blowoff limit is approached the ﬂame becomes

less stable and combustion instabilities, discussed in the next subsection, can occur.

This problem is also eased by using pilot diffusion ﬂame. Combustion instability is

not, however, conﬁned to operation close to the lean limit.

4.3.7 Combustion Instability

Combustion instability is critically important in DLN combustion systems. For

decades the users of gas turbines have been able to depend on their running

for long periods with little attention beyond scheduled maintenance. Owners of

346 Lean Flames in Practice

DLN-equipped systems ﬁnd that their engines require continuous monitoring and

regular retuning. The OEM (original equipment manufacturer) often provides this

monitoring service, but it is expensive. Although DLN systems have been in service

for 25 years, problems of this kind continue.

Combustion instability is a cyclic process driven by coupling between the rate of

heat release and pressure. These two parameters couple easily in a gas turbine, in that

gas volume increases with temperature but turbine nozzle capacity increases only

with the square root of temperature. Thus an increase in heat release rate reduces

mass ﬂow through the NGV, and the pressure rises in the combustor. This reduces

the mass ﬂow from the compressor. Cyclic coupling occurs for the following reasons:

1. When the fuel pressure drop across the injector is high, the fuel ﬂow rate is not

sensitive to combustor pressure. As the air mass ﬂow falls, the mixture becomes

richer, further increasing ﬁring temperature.

2. When the fuel pressure drop across the injector is low, fuel ﬂow is modulated

by variations in combustor pressure. Time delays related to the transport of fuel

into the burning zone can result in a phase relationship between heat release

and pressure favourable to instability.

Lean premixed combustion ampliﬁes this process. The FAR in the reaction

zone is uniform and leaner than stoichiometric. Thus heat release throughout the

mixture increases with FAR. When the fuel is injected directly into the combustor,

some zones are richer and some are leaner than stoichiometric. As FAR increases,

the fuel-rich volume releases less heat, damping to some extent the opposite effect in

the fuel-lean volume. It must be borne in mind that, at high pressure and temperature,

with the high turbulence generated in a gas turbine combustor, almost all the heat

release has taken place before most of the dilution air is admitted. In a DI diffusion-

ﬂame combustor, there is little heat release in regions of ﬂame where the equivalence

ratio is less than 0.8. (Relatively slow consumption of remaining CO and carbon will

continue, but with insufﬁcient heat release rate to generate instability.)

Combustion pressure oscillations are likely to grow in amplitude when the

phase relationship between heat release and pressure are within 90

◦

of each other.

This yields a large positive value for the left-hand side of relation (3.39), which is

Rayleigh’s criterion. Whether this is the case depends on reaction rates, mixture

transport times in the reaction zone, and the ability of these to couple with one

of the several pressure oscillation modes available. These are sometimes acoustic

resonances, or the combustion-related processes may couple with convection of spa-

tiotemporal ﬂuctuations of temperature along the combustor, which can generate

acoustic waves at the turbine nozzle. These waves are sometimes referred to as

entropy waves. The many possible ways in which the aerodynamic, chemical, and

acoustic processes can interact means that combustion instability is extremely likely

to occur at some point in the engine-operating envelope. Frequencies may range

from around 50 Hz to above 5 kHz. Frequencies above about 1 kHz are referred

to as screech, frequencies below 150 Hz may be referred to as humming, rumble

or buzz, frequencies between these two may be referred to as howl, and there is a

low-frequency phenomenon known as chugging at around 10 Hz, which is associated

with operation close to the lean extinction limit. This is not an acoustic phenomenon,

and it is better thought of as intermittent combustion.

4.3 Application of Lean Flames in Stationary Gas Turbines 347

In reading about this subject and in consulting combustion specialists, one may

encounter diverse opinions relating to the causes and description of the same prob-

lem. This arises from the complexity of the phenomena and the current state of

knowledge. Such opinions may sometimes be expressed with an unjustiﬁed dogma-

tism, creating the impression that a problem is well understood and that the proposed

action to mitigate it is certain to succeed. It is not in the interest of most advisors to

give the impression that they are unsure about a subject in which they are supposed

to be expert! The costs of being mistaken are, however, so high, that there is a case for

obtaining a second opinion. Differing opinions may not help to solve the problem,

but could lead to a more realistic assessment of programme risk and risk-mitigation

strategies.

The author strongly recommends a study of [225] by those concerned with the

practical aspects of this problem in a gas turbine plant, especially if they are about

to work with DLN combustors for the ﬁrst time.

In lean premixed combustors, a sustained amplitude above 30 kPa peak-to-peak

is very likely to cause damage. Half this amplitude is usually recommended as a

safety limit. Screech can cause damage at less than 10 kPa peak-to-peak, however.

If the frequency of instability matches the resonant frequency of some component

on the engine, then lower amplitudes can cause failure. Not only does the oscillation

subject the combustor to stress, but it can excite mechanical vibration in components

such as oil and fuel pipes, leading to fracture. If the amplitude exceeds about 4% of

combustor pressure it will reverse the ﬂow in the premixing duct, causing ﬂashback.

Flashback can also be caused by lower amplitudes.

Helmholtz dampers (see Subsection 3.2.2) are often employed in industrial en-

gines to reduce the amplitude of combustion instability. These consist of a volume

connected by a tube to the combustor. The frequency at which they absorb energy

is a function of the volume of the damper and the length and cross-sectional area of

the connecting tube or neck. The resonant frquency of this damper is

ω =

, (4.1)

where ω is frequency in rad/s,

c is speed of sound, A is the area of neck, L is the

length of neck, and V is the volume of the resonator cavity.

The absorption frequency bandwidth may be increased at the expense of the

attenuation efﬁciency by increasing the cross-sectional area of the neck. Often,

several absorbers are needed to absorb energy at different frequencies. They are

fairly large, having a volume in the region of 10% that of the combustor. They

must be connected to the combustor rather than to the outer air-feed annulus to

be effective and may need a small air purge to prevent overheating by ingress of

combustion products.

A novel means of preventing coupling between pressure and FAR is employed

in the premixing ducts of the R-R industrial Trent engines. Natural gas is admitted

through the end of the duct, and the air is admitted through perforations in the wall.

The wavelength of instability is long compared with the length of the duct, so the air

jets are all in phase. If the transit time of the fuel through the perforated section is

at least twice the period of the oscillation then the ﬂuctuations in FAR are greatly

attenuated [227].

348 Lean Flames in Practice

Combustion instability is such a common problem in LPP combustion systems

that industrial DLN engines are continuously monitored and may have to be shut

down and adjusted if instability reaches a certain amplitude [225]. This has a big

ﬁnancial impact if it shuts down a power station engine during a period of peak

demand.

4.3.8 Flashback

Flashback occurs when the ﬂame speed exceeds the velocity of the unburnt mixture

of fuel and air. Premixing ducts are not cooled, so it is potentially hazardous. As

ﬂame speed is only a few meters per second, whereas ﬂow velocity may be as high as

100 m/s, it is seemingly not difﬁcult to avoid. It is important to avoid wakes from swirl

vanes or fuel injectors and thick boundary layers that persist along the length of the

premixing duct, and this is best achieved by contracting the passage to accelerate the

ﬂow towards the outlet. If too much swirl is imparted to the air to encourage mixing,

then the ﬂow on the centreline of the duct may reverse, allowing ﬂashback along the

core of the duct. Combustion oscillation adds to these problems. The discussion on

“Practical limitations governing the design of a premixing duct” is applicable here.

4.3.9 Auto-Ignition

This subject is discussed in Subsection 3.1.3 of Chapter 3. Here the discussion is

extended to include the effects of various fuels employed. Data provided in that sec-

tion may be used for number 2 light diesel distillate. The auto-ignition delay time for

diesel fuel will be lower than that for kerosene, but not to a degree that is signiﬁcant

when compared with the uncertainty in the kerosene data. Auto-ignition delay time

is difﬁcult to measure, and results depend on the method of measurement used. It is

therefore necessary to be conservative in applying published data. The writer would

always design for half the expected ignition delay time based on the shortest delay

time measured in published research. (There is a need to exercise some judgement

concerning the quality of published data.) The reason for this conservatism is that

in engines there are aerodynamic phenomena that can be controlled far less easily

than in a research test facility; notably large-scale turbulence and combustion-driven

pressure oscillations in the premixing ducts.

Auto-ignition can be caused by some of the features that can cause ﬂashback.

It is not always easy to diagnose which of the two has occurred. Wakes, boundary

layers, or small recirculation zones where unburnt fuel–air mixtures can reside can

provide a site for auto-ignition. Also it can be triggered, as can ﬂashback, from

hot surfaces in the premix duct arising from radiation and conduction back from

the ﬂame zone. It is far less likely in industrial gas-burning engines, both because

natural gas has a higher auto-ignition temperature than kerosene and because t he

compressor delivery pressure and temperature are lower than in aeroengines.

There is very little published auto-ignition data for methane or natural gas in

the temperature range of interest. Most experiments are conducted at temperatures

above 1100 K. On account of the heavier components in natural gas, the auto-ignition

delay time is around half that of methane – but this is only a rule of thumb. There

is considerable variation in the constituents of natural gas around the world. In

4.3 Application of Lean Flames in Stationary Gas Turbines 349

engines without regenerative heat exchangers and with pressure ratios below 25:1,

auto-ignition is extremely unlikely to occur with natural gas, provided that reasonable

care is taken in the aerodynamic design of the premixing ducts. At pressure ratios

above 30:1, although not a formidable problem, special care should be taken to

streamline passages containing fuel–air mixtures. Where inlet air is preheated by a

regenerative heat exchanger, temperatures above 850 K are potentially hazardous

whereas at pressures above 30 bars without such a preheat, an auto-ignition delay of

natural gas may be encountered. The author knows of no reliable published data on

which to base design rules. That is not to say that the leading manufacturers do not

have reliable data of their own.

Another possible source of pre-ignition is PM from compressor tip rubs. This is

not a common cause, but cannot be ruled out from investigation of a problem. The

cause of pre-ignition is rarely obvious – partly because so much damage results that

evidence of the cause is destroyed.

4.3.10 External Aerodynamics

FAR deviations from the mean of more than 2% have a signiﬁcant effect on NO

in combustion systems designed to achieve values below 15 ppmv. There are in-

evitable variations from combustor to combustor through manufacturing tolerances

and likewise in the fuel injectors. With the passing of time, deposits can build up

in fuel passages, which increase variability. Variations in the air-feed pressure add

to the total variation. Whereas aircraft gas turbine emissions are veriﬁed only once

before leaving the factory, a stationary plant is checked annually.

The lengths of many LPP combustion systems are too great to ﬁt between the

compressor outlet and turbine inlet, so the compressor delivery air must be turned

through a large angle to feed the LPP fuel injectors. The GE LM6000 DLN combustor

avoids this problem, but its great depth requires a complex, multipassage compressor

outlet diffuser. Both radially and axially disposed tubular combustors present the

aerodynamicist with a challenge. To achieve the very low NO

required and to avoid

ﬂashback, a high degree of uniformity in the air distribution into the premixing

ducts is required. Where the airﬂow turns through a large angle and passes between

tubular combustors, it is obviously difﬁcult to achieve this. The problem is solved

by careful development, supported by CFD and experiment, and may require some

compromise in acceptance of a higher parasitic pressure loss with an associated loss

of cycle efﬁciency [228].

4.3.11 Combustion Research for Industrial Gas Turbines

To conduct combustion research at full scale and full pressure and temperature

is formidably expensive, even for a 50-MW aeroengine. For a 200-MW industrial

engine the air compression and heating plant required would be colossal. Yet auto-

ignition can be evaluated only at full pressure, temperature, and residence time.

Some aerodynamic phenomena mentioned in the previous section require testing in

an environment of realistic geometry, and some acoustic modes that drive combus-

tion instability occur in only the full annulus of an engine system. The aeroengine

industry has used extensive component rig testing since the inception of the gas

350 Lean Flames in Practice

turbine. This is justiﬁed by the high cost per unit power, the smaller size of the

combustors, airworthiness requirements, and low weight that results in the need to

design with the minimum stress safety factors. The large number of units produced –

typically over 1000 for a single model and occasionally 10 times that number – makes

extensive testing economically feasible. It has been customary for industrial engines

to be developed under special arrangements with kindly disposed customers on the

customer’s site.

Aeroderivative engines are downrated by removal of the fan, which reduces the

cycle pressure ratio by 30%, and to achieve the very long life at full load required

in stationary applications. (An aeroengine spends only a small percentage of its

life at above 85% load). Consequently the adaptation of aeroengines to industrial

application has not, until relatively recently, required extensive research.

The necessity to drastically reduce emissions has transformed this situation.

Ultra-low-NO

combustion systems cannot be designed and tested in the production

engine, although a demonstrator engine is often employed in the development of

a generic design. Extensive research is required, followed by validation testing at

high pressure and temperature. Combustion research in the industrial gas turbine

ﬁeld is now more competitive and technologically advanced than in the aeroengine

ﬁeld. As explained in Section 4.2, innovation in aeroengine combustion technology is

limited by airworthiness considerations. Emissions control cannot take precedence

over safety. In the industrial ﬁeld the conﬂict is a lot less severe.

Generic research cannot take account of engine geometry. It is necessary to

recognise the signiﬁcance of this in evaluating results. As mentioned in Subsec-

tion 4.3.10, there are many sources of non-uniformity in the air and fuel supply to an

engine combustor that are not present in a combustion rig. Emissions performance

achieved in a research rig cannot therefore be replicated in an engine.

Two topics are of great importance: combustion instability, which has already

been discussed at length, and auto-ignition of natural gas. As noted earlier, there

is little reliable data describing the auto-ignition characteristics of natural gas at

the temperatures and pressures encountered in modern gas turbines. Of particular

interest are temperatures between 800 and 900 K at pressures between 15 and 40 bars.

Variation in the chemical constituents of natural gas must be included. In devising

laboratory research programmes with an eye on industrial application, physical scale

and the turbulence structure in airﬂows is important. Attempting to achieve greater

realism will increase the variability in results. The author believes that this must be

accepted. Highly repeatable data from a completely non-representative environment

is almost useless, as the engine designer does not know what safety margin to allow.

Experimental data with realistic variability provide statistical information that is of

practical use.

4.3.12 Future Trends and Research Emphasis

limits will probably become increasingly stringent, and in a few years’ time the

generally accepted limit will probably fall from 25 ppmv to 15 or even 10 ppmv. These

values have already been achieved by DLN systems in some applications. It will be

apparent from the foregoing discussion that the greatest obstacle to the achievement

of lower NO

is combustion instability. This is a very complex phenomenon, and there

References 351

is much to be done both to better understand it and to devise methods for its control.

Active control of fuel admission is a promising ﬁeld of study for liquid fuel. This

topic is discussed at length in Section 3.2. The greatest difﬁculty at present lies in the

perfecting of valves that can modulate ﬂow at the required frequency. The inertia of

liquid fuel opposes every effort to rapidly change the ﬂow rate, and reﬂected waves

from geometric features in the pipe work tend to attenuate input variations. The

closer the control valve is placed to the fuel nozzle, the better, but then the hostile

environment creates problems in t he valves. Because of the compressibility of gas it

is difﬁcult to see how active control of the gas ﬂow can be transmitted to the injector

ﬁnal oriﬁce. There is scope for the development of improved fuel–air mixing devices.

The discussion in Subsection 4.2.9 should be read in conjunction with this section,

and the discussion in that section on the topic of CFD applies equally in the industrial

gas turbine ﬁeld.

REFERENCES

[1] B. Lewis, Outlook in combustion research, Combust. Flame 14, 1– (1970).

[2] R. F. Sawyer, Science based policy for addressing energy and environmental

problems, Proc. Combust. Inst. 32, 45–56 (2009).

[3] COM(2007) 19 ﬁnal: Communication from t he Commission to the Council and

the European Parliment, “Results of the review of the Community Strategy

to reduce CO

emissions from passenger cars and light-commercial vehicles”

(2007).

[4] A. M. K. P. Taylor, Science review of internal combustion engines, Energy

Policy 36, 4657–4667 (2008).

[5] R. L. Evans, Lean-burn spark-ignited internal combustion engines, in D. Dunn-

Rankin (ed.), Lean Combustion: Technology and Control (Academic, New

York, 2008).

[6] R. W. Schefer, C. White, and J. Keller, Lean hydrogen combustion, in D. Dunn-

Rankin (ed.), Lean Combustion: Technology and Control (Academic, New

York, 2008).

[7] Light-duty automotive technology, carbon dioxide emissions, and fuel econ-

omy trends: 1975 through 2009. Executive summary. Compliance and inno-

vative strategies division and Transportation and Climate division, EPA 420-

S-09-001 (Ofﬁce of Transportation and Air Quality, U.S. Environmental Pro-

tection Agency, November 2009); http://www.epa.gov/otaq/cert/mpg/fetrends/

420s09001.pdf.

[8] Federal Register (www.regulations.gov) under Docket ID EPA-HQ-OAR-

2009-0171. Also http://www.epa.gov/climatechange/endangerment.html.

[9] National Highway Trafﬁc Safety Administration, Docket NHTSA-2009-0059,

Notice of intent to prepare an environmental impact statement for new cor-

porate average fuel economy standards (U.S. Department of Transportation

2009).

[10] Council Directive 89/458/EEC of 18 July 1989, amending with regard to Eu-

ropean emission standards for cars below 1.4 litres, Directive 70/220/EEC on

the approximation of the laws of the Member States relating to measures to be

taken against air pollution by emissions from motor vehicles, Ofﬁcial Journal

L 226, 3.8.1989, pp. 1–3.

[11] Council Directive 91/441/EEC of 26 June 1991 amending Directive 70/220/EEC

on the approximation of the laws of the Member States relating to measures

to be taken against air pollution by emissions from motor vehicles, Ofﬁcial

Journal L 242, 30/08/1991, pp. 0001–0106.

352 Lean Flames in Practice

[12] 1999/125/EC: Commission Recommendation of 5 February 1999 on the reduc-

tion of CO

emissions from passenger cars, Ofﬁcial Journal L 040, 13/02/1999,

pp. 0049–0050.

[13] T. Lake, J. Stokes, R. Osborne, and R. Murphy, The lean boost combustion

system for improved fuel economy, in H. Zhao (ed.), Advanced direct injection

combustion engine technologies and development, Gasoline and gas engines

(Woodhead, Oxford, 2009), Vol. 1.

[14] P. Leduc, B. Dubar, A. Ranini, and G. Monnier Downsizing of gasoline engine:

An efﬁcient way to reduce CO

emissions, Oil Gas Sci. Technol. Rev. IFP 58,

115–127 (2003).

[15] Regulation (EC) 443/2009 of the European Parliment and of the Council of 23

April 2009, setting emission performance standards for new passenger cars as

part of the Community’s integrated approach to reduce CO

emissions from

light-duty vehicles, Ofﬁcial Journal L 140, 5.6.2009, p. 2.

[16] The plain English guide to the Clean Air Act, publication EPA-456/K-07-

001, U.S. Environmental Protection Agency, Ofﬁce of Air Quality Planning

and Standards, Research Triangle Park, NC (2007). http://www.epa.gov/air/peg/

understand.html.

[17] Council Directive 70/220/EEC of 20 March 1970 on the approximation of the

laws of the Member States relating to measures to be taken against air pollution

by gases from positive-ignition engines of motor vehicles, Ofﬁcial Journal L 076,

06/04/1970, pp. 0001–0022.

[18] Directive 2008/50/EC of the European Parliament and of the Council of 21

May 2008 on ambient air quality and cleaner air for Europe, Ofﬁcial Journal L

152, 11.6.2008, pp. 1–44.

[19] Decision 1600/2002/EC of the European Parliament and of the Council of 22

July 2002 laying down the Sixth Community Environment Action Programme,

Ofﬁcial Journal L 242, 10/09/2002, pp. 0001–0015.

[20] A summary of Directive 2008/50/EC of the European Parliament and of the

Council of 21 May 2008 on ambient air quality and cleaner air for Europe. http://

europa.eu/legislation

summaries/environment/air pollution/ev0002 en.htm.

[21] Title 40 of the United States Code of Federal Regulations: Protection of En-

vironment. Part 50-National Primary And Secondary Ambient Air Quality

Standards. Summarised at http://www.epa.gov/ttn/naaqs.

[22] Primary National Ambient Air Quality Standards for Nitrogen Dioxide; Fi-

nal Rule. Environmental Protection Agency 40 CFR, Parts 50 and 58, Fed-

eral Register, Vol. 75, No. 26, Tuesday, February 9, 2010, Rules and Reg-

ulations. Summarised at http://www.epa.gov/air/nitrogenoxides/pdfs/20100124

presentation.pdf.

[23] J. Bachmann, L. Zaragoza, J. Cohen, and J. Haines, Review Of The National

Ambient Air Quality Standards For Particulate Matter, Assessment Of Scien-

tiﬁc and Technical Information, OAQPS Staff Paper, Strategies and Standards

Division, Ofﬁce of Air Quality Planning and Standards, U.S. Environmental

Protection Agency, Research Triangle Park, N. C. 27711, EPA-450/5-82-001.

Available at http://nepis.epa.gov.

[24] J. Bachmann, Air today, yesterday and tomorrow: Part 2. Presentations on air

quality management and health effects (2008). Available at http://www.epa.gov/

air/caa/.

[25] J. Bachmann, Will the circle be unbroken: A history of the U.S. National Am-

bient Air Quality Standards, J. Air Waste Manage. Assoc. 57, 652–697 (2007).

[26] http://www.epa.gov.ttn/atw/pollsour.html.

[27] Environmental fact sheet, Air toxics from Motor Vehicles, EPA400-F-92-004.

Available at http://www.epa.gov/oms/f02004.pdf.

[28] http://ec.europa.eu/environment/air/quality/standards.htm. Summarising Air

Quality Standards in Directive 2008/50/EC of the European Parliament and