345 E. 47th St, New York, N.Y. 10017
The Society shall not be responsible for statements or opinions advanced in papers or cfiicussion at meetings of the Society or of its Divisions or
Sections, or printed in its publications. Discussion' is printed only if the paper is published in an ASME Journal. Authorization to photocopy
material for Internal. or personal use under circumstance not fang within the fair useTprovisions. of the Copyright Act is granted by ASME to
libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service provided that the base fee of 50.30
per page is paid directly to the CCC, 27 Congress Street. Salem MA 01970. Requests for special perrniesicin or bulk reproduction should be addressed .
to the ASME Technical Pubfishing Department.
All Rights Reserved
Copyright 1997 by ASME
. Printed in U.S.A
European Gas Turbines Ltd.
Lincoln, United Kingdom
The second Generation of EGT's G30 DLE combustion system
was introduced after a successful series of high pressure rig and
engine tests. This paper covers how operational problems with field
commissioning hardware on the lead DLN machine were dealt with,
leading to achievement of reliable low NOx hardware. Several
changes were applied to the early design which improved the mixing
and reduced the effects of high temperature distortion and combustor
dynamics. This resulted in increased life of the burner and changed
the characteristics of dynamics. It also led to very low emission
levels with an outstanding capability for turndown of CO with NOx
below 25 ppmvd (15% 02) over the whole load range. Further
coverage is given to the effect of field tuning, and of fuel
composition on the amplitudes and frequencies of dynamics. The
installation has been supported by on-line condition monitoring of
engine parameters, emission levels and ambient conditions, which are
also discussed. The general overview of site history is followed by a
summary of lessons learnt in field comparison to development test
Continuous Emission Monitor
Carbon Monoxide
Lean Blow Out
Full Speed No Load
Oxides of Nitrogen (NO+NO2)
Parts per million by volume, dry gas
Turbine Operating Temperature
(PTexit temperature-Inlet temperature) • constant
Turbine Entry Temperature
Route Mean Square
Unburnt Hydrocarbons
Lower Calorific Value
Temperature of gas fuel in K
Specific Gravity
Reaction zone equivalence ratio
The first industrial gas turbine with Dry Low NOx (DLN)
technology from EGT was supplied and commissioned in March 95
after an accelerated rig and engine development programme. The 4.5
MW Typhoon Gen-set forms part of a Combined Heat and Power
(CHP) plant at the Arnhem headquarters of the Dutch chemical
company Akio-Nobel. The site provides an ideal engine location for
endurance testing and evaluation, where the operation is essentially
baseload with excess electrical output sold to the grid.
Three combustor build changes over a period of 13 months from
the initial installation with NOx levels of 55 ppmvd in Phase IA,
resulted in hardware producing less than 15 ppmvd with virtually no
CO and no UHC at baseload conditions. (Table I.)
TABLE 1: Milestones in EGT's DLE Programme
Date (Commission)
March 95
Oct. 95
May 96
Single point Design, 3 cusp radial inward pilot
ignitor centre position
IM Design, revised piston Seal, cooled centre
ignitor, high dynamics damper
2M Design, profiled pre-chamber
off-centre ignitor, reduced film cooling
Flow matched components
27.6 kPafRMS
150 Hz
10.3 kPa/RMS
150 Hz
0.7 kPafRMS
360 Hz
55 ppmvd
37 ppmvd
fretting, carbon deposit
Tip temperature, mixing
parts life, Tip temperature
Pre-chamber, distortion
load shed operation
<15 ppmvd
Presented at the International Gas Turbine & Aeroengine Congress & Exhibition
Orlando, Florida — June 2–June 5, 1997
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 12/29/2014 Terms of Use: http://asme.org/terms
engine performance data and a CEM system. This equipment was
accessible on-line through a direct modem connection for condition
The combustor itself has experienced over 10000 running hours
and the life time prediction is at least 25000 hours or 10000 cycles.
As part of EGT's G30 technology programme, the 030 DLN
combustor has been specifically developed for EGT's small Gas
Turbines up to 8 MW. It is based on a simple fixed burner geometry,
using a part premixed, lean-burn approach and has been described in
detail by Norster and DePietro (1996).
The combustor (Figure I.) consists of a radial inflow swirler,
swirler-slot injection and a premix chamber. A pilot burner is situated
in the centre, but has a flush face with the radial swirler slots and is
mainly used for ignition and to stabilise the flame on part load
The initial phase IA hardware was unable to meet the Dutch
legislative limit of 37 ppmvd NOx, (15% 02) [actually 65g/G/ heat
input]. The hardware was installed firstly to provide a means of
commissioning the boiler and generator and secondly to gain some
field experience with DLN in commercial duty.
The immediate problems encountered on site started by reestablishing the ignition window which had moved since the
endurance run on the test beds in Lincoln. After starting the unit and
achieving Full Speed No Load (FSNL), the synchronisation and the
loading of the generator proved to be more straight forward than
experienced on the development test beds.
NOx was measured at 55 ppmvd (15%02), using both the CEM
and a portable analyser. This was consistent with the levels measured
in Lincoln.
Internal combustion pressure oscillations (dynamics) were
measured at levels of 27.6 kPa/RMS above 90% load, using
piezoelectric transducers fined to the burner ignitor ports and
pressure casing.
The most significant issues arising from the introduction of this
hardware related to ignitor life and carbon formation on the recessed
ignitor at low loads. The severity of this carbon build up was such
that after 4 hours the ignitor recess would be effectively closed off,
thereby isolating the ignitor from the gas mix (Figure 3).
Subsequently after the engine was shut down, re-ignition was not
possible. Sustained running at full load on the other hand resulted in
overtemperature of the central pilot zone, leading to oxidisation of
the pilot and overtemperature of the ignitor tips.
Figure 1. Burner Assembly
This design can be retrofitted to all Typhoon gas only engines,
since it does not require changes to the present centre casing
arrangement. (Figure 2.)
Figure 3. Oxidised and carboned Pilot Tip
Figure 2. Centre Casing Arrangement
In addition to ignitor and pilot tip temperature problems high
load operation highlighted another major issue not seen in
development testing. After one week running at full load, high levels
of low frequency dynamics (nominally 27.6 ld'a @ 150Hz) caused
The site installation has been supported with an extensive data
acquisition system, which comprises of combustion parameters,
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 12/29/2014 Terms of Use: http://asme.org/terms
severe fretting at the transition duct combustor liner sealing joint.
The transition duct and piston ring sealing arrangement though wear
coated were found to have grooves up to 2mm deep in places
(Fig.4.). As a consequence, the unit was limited to 3.4 MW to
minimise damage and the risk of material ingestion into the turbine.
burner was modulated to balance the tip temperatures, the exhaust
temperature spread, but more importantly the dynamics. Fuel tuning
option enabled the dynamics to be reduced to less than 10.3 kPa on
all burners.
Combustion inspections completed at 800, 1200 and 5000 hours
indicated the hardware was in good conditiontwith no re-occurrence
of fretting. Problems which were apparent though were tip
temperature drift with time and poor ignition after compressor cold
wash. Three pilot burners had to be exchanged after they developed a
progressive tip temperature drift characteristic with time but this was
achieved during inspection to minimise machine downtime.
Fuel Split Schafule
Figure 4. Transition Duct damaged through fretting
The problems experienced with the previous set of hardware led
to an improvement programme targeting both the high dynamics
levels and the levels of NOx in order to meet the emission
specification. The continuing development programme in Lincoln in
parallel with the field commissioning had already produced some
very encouraging test-rig results with single digit NOx figures. Phase
1B hardware was introduced based on these developments.
New features of the Phase 113 include an improved mixing main
burner with multi-point injection hardware, a central air cooled pilot
tip design featuring an air-swept ignitor. Dynamics were reduced by
improving the flame stabilisation on the pilot and by damping the
lateral pressure waves in the combustor liner. An improved seal
arrangement was also introduced to minimise leakage airflow
between transition duct and combustor and thereby control NOx.
This revised hardware was introduced in October 1995 and the
unit re-commissioned. The light-up characteristics remained
unchanged, however, the main/pilot fuel split schedule defined on
the test bed could not be applied to the site installation. Load swings
of up to 0.5 Mw were experienced at 85% load and very high tip
temperatures (transiently 1200 Deg C) were seen at 95% load. The
fuel split schedule which regulates the ratio of fuel being injected
through the main and pilot is dependent upon load and controlled as
a function of the engine operating temperature (T0p). A typical fuel
schedule is shown in Figure 5.
It became apparent, that the translation of the hardware from UK
to Dutch Natural Gas was an issue that needed serious consideration.
By exhaustive fuel split mapping a schedule was established that
permitted safe and stable operation across the load range. NOx
emissions levels were recorded at 37 ppmvd (15%02) by the CEM at
ISO conditions and the Dutch emission acceptance test certified
compliance with the permit. The installation of fuel tuning valves in
the main burner pipes enabled gas flow to be tuned manually to each
burner. By selective setting of these valves, the total gas flow to each
Opeatim Porn, TOp
Figure 5. Fuel split schedule
Both the tip temperature and ignition problems were caused by
the cold wash sequence, where dirt, fouling the compressor was
flushed through to the combustor and more importantly into the pilot
tip cooling air feeds. These feeds were very small (taking less than
1% of total combustor air flow) and had a disproportionate effect on
burner performance by affecting the stoichiometry at the ignitor tip.
This degradation was observed through the online engine healthmonitoring data link which allowed sufficient time for intervention
and the prevention of potential damage.
After 5225 hours of operation with IB hardware, the final phase
2 hardware was installed in June 1996. It consisted of the exchange
of the main burner and pilot assembly with a further improved
mixing through a multi-multi main burner design, designated 2M to
signify 2 side wall injection holes. It also featured an offset ignitor
pilot design without tip cooling with a bluff front face to aid flame
stabilisation. The stoichiometry of the combustor reaction zone was
richened by an increased effective area transition duct to account for
the improvements in mixedness but the combustor barrel was not
Again there was no change in operational characteristics at lightup, but at baseload the combustor proved to be operating very close .
to LBO limits. The baseload pilot percentage with Dutch gas
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 12/29/2014 Terms of Use: http://asme.org/terms
gases dewpoints tend to be relatively consistent (CO Deg C) the
Wobbe Number can vary significantly: e.g. from Dutch Gas at 37
MlThm3 to Alaskan Gas at 48 MJ/Nm3.
Conventional burners are designed to operate within given
bounds of pressure ratio (typically 1.03 to 1.4), and the design rules
for this are well understood, however, adopting and following rules
becomes all the more important for a DLN combustor.
All the G30 development testing (both rig and engine) was
completed on UK Pipeline Quality Natural Gas in the facility in
Lincoln. Early field experience was obtained from the lead machine
in Holland running on Dutch Groningen gas which includes 14%
Nitrogen as a diluent. Table 2 illustrates the difference for these gases
in both theoretical parameters of calculated flame temperature, fuel
penetration (into premixer slot), and measured parameters of NOx
and main burner pressure drop.
(I4%Nitrogen) had to be set at a higher level to compensate for the
mixedness effects with the new burner compared with values typical
of UK gas.
The dynamics level after modifying the baseload pilot split were
still measured at less than 0.7 kPa at all load points. The pilot tip
temperatures were less than 700 °C and NOx at these higher pilot
flows was measured at 14 ppmvd (15%02), CO was measured below
2 ppmvd. (Figure 6.)
Both emissions and tip temperatures have remained stable at
commissioning values and no re-occurrence of ignition problems has
materialised since the simplified design has no air passages to
accumulate water or dirt
The introduction of the G30 DLN system into the field has been
supported and controlled to a large extent by the engineers directly
involved in the development of the combustion system. This has
meant that lessons learnt in the field i.e. operational difficulties and
the results of hours associated with site machines have been fed
directly back into the development programme. Some of these
lessons and observations are discussed below:
Effect on light up (ignition) window
A combustor ' light up window' is defined as the pilot/main fuel
ratio and gas flow required to give ignition. During development
testing it had been found that to achieve consistent DLN burner
ignition lean or ' soft' lights were required i.e. with very low gas
flows and relatively small pilot %' s. Overfuelling to create a rich or
' hard' light simply resulted in no ignition due to the creation of an
over rich zone around the ignitor tip. Translating this into the
Figure 6. 14 ppmvd NOx vs ambient
variation over 24 hours
CEM Emission vs. Ambient
6tco 0
70 OatizipaNd:$5000,9:55-afb.calsoacc- 00
- 20
NO* CO ppnwc11. RM.]
a NOx(corr)
+44. 414,440++
awaeafrocardandoroaattOna 5
attootararetWagtiana ag eirartabreetwogemprocaveterAw
0 0 8 0
a si s
0 0
C. o 0
24 hour log
experience on Dutch gas showed that attempting to increase the
throttle angle to give the same heat input as on UK gas was
unproductive. The effect of the increased gas volume was to swamp
the ignitor and prevent a flammable air/fuel mixture reaching the
ignitor plasma. The results showed that the critical parameter was the
pilot gas volume at ignition not the heat input and that this could be
Gas Composition and Temperature
' Natural Gas ' covers a broad range of possible compositions
depending on both, the type of gas field and the age of the well. Even
removing the direct from the ground "well gases" and analysing what
are termed " Pipeline Quality Natural Gases " leaves considerable
variation in composition and Wobbe Number. Although on these
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 12/29/2014 Terms of Use: http://asme.org/terms
temperature and better mixing were in practice offset by the need to
operate at a significant higher pilot percentage to minimise the lean
chug dynaniics and maintain stability.
In comparison to the Pipeline Quality Natural gases a typical oil
field "well gas" may include a significant proportion of higher
hydrocarbons and therefore has a relatively high Wobbe Number.
The lower burner pressure drops associated with that fuel gases
mean that fuel system feedback and coupling with combustor
dynamics would be a real possibility. In reality however, the
dewpoint of these gases tend to be at higher temperatures (i.e.+20 to
+40 °G), that it would not be handled without fuel heating, the effect,
which is to increase the gas volumetric flow rate and thus the burner
pressure ratio.
achieved by balancing both overall fuel flow and pilot/main fuel
split. Typical ignition fuel splits range from 18 - 25% pilot.
Effect on Flame Temperatures
Theoretical flame temperatures for the respective gases are
shown in Table 2 and would suggest that since thermal NOx is
temperature dependent a reduction in NOx should be measurable
between field and development results. This predicted difference was
not seen in practice on any sites or on any of the hardware standards
and did not appear to be dependent on absolute NOx levels (whether
55 or 14 ppmvd N0x). The observations and conclusions drawn was,
that because of the nature of the premixer, other factors such as
burner pressure ratio were far more important to NOx levels than the
flame temperature.
FFT Analyser Trace
TABLE 2: Gas Fuel Effect
North Sea Gas
Groningen Gas
700 •-•
Wobbe Index
psdil/m3 1
Flame Temperature [°K]
300 .
@ ORE=0. 53
NOx (156/002,dry)
15 (4)
14 (8)
(a) (pilot %)
100 150 19$ 245 295 345 385 435 485 535 585
Frequency (112]
Main burner AP [psi]
Figure 7. Pm-mix characteristic @ 360 Hz
@ 6% pilot
FIT Analyser Trace
Effect on fuel system pressure ratio (burner pressure
Initially, no hardware modifications were included to adapt the
burner to the higher gas volume flow rates of the Dutch Groningen
Gas. The result was an increase in burner pressure drop (illustrated
in Table 2 for the Phase 2 hardware). Because of the passage
injection system design of the 630, this higher pressure drop
translates into an increased jet penetration into the slot. The result of
this was seen in the field as the inability to run as low a pilot split as
achieved on the test bed in Lincoln i.e. the design stoichiometry of
the reactor was too lean to permit operation at this level of premix.
This exhibited itself on the machine as a pronounced LBO dynamic
characteristic (or lean chug).
On LIK gas a very low pilot could be sustained at baseload without
loss of flame or significant chug.
A characteristic pre-mix tone is shown in Figure 7., correctly tuned
for stoichiometty, a medium amplitude pilot driven peak (160 Hz)
and the dominant hot tone premix dynamic peak (360 Hz). Whilst
Figure 8. shows a typical chug tone at 60 Hz, for a combustor
approaching LBO limit.
As mentioned previously, there was no NOx-advantage
measurable on the Dutch gas. The benefits of lower theoretical flame
IL 600
50 80 130 180
2110 330 380 410 430 530 510
Fitqutney [Hz]
Figure 8. Lean Blow Out! Chug tone @ 60 Hz
Based on the site experience to date and experience of variable
gas composition a set of design rules have been established for
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 12/29/2014 Terms of Use: http://asme.org/terms
burner pressure drop and burner hole size for the 030 system. These
rules utilise a dewpoint temperature corrected Wobbe number.
Whilst these valves proved of considerable value in early
development testing, the experience on both the production test bed
and in the field has shown little need or benefit from fuel tuning. The
optimisation of burner pressure ratio's and improved manufacture
control of effective areas in Phase 2 hardware effectively making
tuning valves unnecessary.
Wobbelndex = LCV 18
Further site evaluation and validation will form an essential part
of establishing the upper and lower limits of the design rules.
Ignition System
One of the major concerns raised during the 030 concept
discussion was the utilisation of a no cross light tube and only one
ignitor per combustor philosophy i.e. no redundancy in ignition
capability. The concern was that if any one of the six ignitors,
ignitor leads or ignitor channels in the multiplexed energiser box
developed a fault the unit could not be started. However, after 10000
hours on the lead machine, and with another 13 machines in duty no
problems have arisen from the ignition system reliability.
The lesson being that provided the ignitors are kept cool either
by air cooling (as Phase 1B) or gas cooling (Phase 2) and kept free of
carbon plugging there will be no major issues on combustor ignition.
In addition the phase 2 hardware with offset ignitor also eliminates
the effects of cold wash being retained in the ignitor cooling
As discussed previously it is the pilot gas volume flow rate that
defines the ignition window. With increased numbers of machines
entering duty some variability in the position of the light up window
(pilot % and light up gas flow) has been seen. This is a function of
manufacture tolerance on combus-tor/bumer effective area and fuel
splitter valve and is the inevitable result of making six independent
ignition and combustion zones.
What has also become evident though is that, provided the
ignition window is adequately mapped during commissioning and
the % pilot and light-up flows set at mid range values, ignition can be
Fuel tuning
In Phase 1B and Phase 2, the main burner feeds were fitted with
fuel tuning valves. In the case of the 1B these were commercially
available diaphragm valves illustrated in Figure 9. whilst those fitted
to Phase 2 were purpose designed and made for DLN fuel tuning.
Tuning Valve
Ambient temperature effects
Variation in ambient temperature was found to have a
significant effect on the level of 68Hz-chug dynamics at baseload.
This is an effect also seen with high Wobbe Number sites (i.e. low
gas volumetric flow rate). The ambient temperature effect only
became apparent on introduction of the Phase 2 design where the
level of premix was significantly better.
The mechanism for this effect is that as ambient temperature
rises so the compressor discharge temperature rises, and in order to
maintain the turbine exhaust temperature (and thus the design TET)
the fuel must be reduced. This results in a lower temperature rise and
hence leaner combustor operating point. What this means to the
combustor is, that the hotter the ambient temperature, the leaner the
reactor stoichiometry and the closer the combustor operates to the
LBO limit. In order to counter this, an ambient biased fuel schedule
was devised in which the baseload fuel split is modulated as a
function of the inlet temperature above 15 °C. By altering the reactor
mixedness the margin before LBO is improved and chug dynamics
are reduced.
Figure 9. Alao Unit Combustion Inspection
Using these valves each of the 6 main feeds could be selectively
tuned enabling the total gas flow to each combustor to be modulated.
By iterative tuning it was possible to reduce the burner dynamics that
might arise as a result of burner or combustor effective area
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 12/29/2014 Terms of Use: http://asme.org/terms
required to achieve reliable load shed capability for machines tuned
to higher levels of premix and single digit NOx.
Operability and load handling
One of the positive benefits found on moving to the field
installation was that of speed stability. In the Lincoln test bay where
engine load was provided by a water dynamometer, shaft speed
stability was not good (particularly at simulated no load conditions.
Speed variations of +/- 200 rpm were typical at the nominal shaft
speed of 17250 rpm. Concern was raised that this variation would not
allow synchronisation with the grid, however, the first run with a
generator indicated that there was no issue on speed fluctuation. No
speed synchronisation problems have been encountered on any site
and it was evident that this was a brake problem unrelated to
combustor or control stability.
Loading and unloading proved to be uneventful and again
significantly better on generator than on water brake with regard to
speed stability. The major handling problem proved to be that of
rapid unload or load shed. This is an issue for all DLN systems
operating with a lean premix flame but is particularly an issue for
small machines in generator drive applications. The most arduous
unload scenario is a Grid Breaker open load shed from full or
baseload to no load. This has to be achieved without loss of flame
and this test forms part of the requirement at a number of the Field
The mechanism for load shed from baseload to FSNL (for a
single shaft machine) is as follows. The Grid breaker opens and there
is an almost instant loss of load on the turbine. Since the throttle
valve for gas flow is fully open and the turbine is no longer
frequency (and hence speed) locked to the Grid the turbine speed
starts to increase. In order to prevent turbine overspeed the throttle is
closed rapidly to a sub FSNL condition (i.e. to the Governor Blow
Out limit) and then re-opens once speed rise is contained and the
turbine is restored to its FSNL condition and is ready to resynchronise.
The 030 combustor weak extinction limits with greater than
10% pilot far exceeds the capabilities of the conventional combustor,
however, in order to achieve NOx levels below 15 ppmvd, low pilot
flows are required and this inevitably compromises the flame
stability. At these pilot flows and with the high level of main burner
pre-mixing, load sheds can only be achieved by rapid increase of the
pilot fuel flow.
Preliminary issues during site testing related to the controls
recognition of combustor flameout. The temperature monitor routines
in the software which utilise information from the exhaust
temperature thermocouples were initially set up for conventional
combustor exhaust spreads and required modification to operate with
the more uniform DLN exhaust spread. During testing it was also
evident that the response time of the turbine exhaust thermocouples
and thus the ability of the control system to advise a revised fuel split
requirement was too slow.
A fuel split override logic was introduced to the software, such
that once above a pre-determined load point defined by a TOp set
point the pilot splitter valve would be stepped open from a very low
flow setting to nominally 25% open value when a load shed detected
signal was received from the throttle governor. This proved to be
very effective for current standard of machines tuned at 15 to 25
ppmvd NOx (15% 02) but the indications are, that further work is
There have been theoretical investigations into the effect of
humidity, ambient pressure and temperature on lean premix
combustors for a number of years. Some authors (Lewis 1981, Visser
and Bahlmann 1994) could relate the impact of fuel flow control due
to ambient variation correctly. In fact, their theoretical studies
suggested that the humidity effect of the EPA correction could be
omitted and that the temperature effect would be opposite to the
effect encountered with conventional diffusion flames.
The control of the single shaft Gas Turbine is based on
maintaining a constant Turbine Entry Temperature (TET) at
baseload. This is implied from a derived Turbine operating
temperature or TOp as discussed previously, consequently a
combined effect of ambient air condition and fuel flow will regulate
the flame temperature at the same value.
Determining an ambient temperature effect on NOx and CO has
been one of the major goals of the CEM and continuous performance
monitoring exercise to enable modifications to EGT's emission
correction formula.
From data collected throughout the installation Phases, EGT
derived conclusions, which imply a logarithmic relation between the
change in ambient temperature and the produced NOx. With regards
to the level of mixing it was apparent, that the 2M burner was less
sensitive to ambient effects than the IM design (Table 3.)
TABLE 3: Sensitivity to ambient Temperature
ANClx/5, T1
1M ( 37 ppmvd)
0.7 ppmvd /
2M (<I5 ppmvd )
0.15 PP""d/
This follows the general tendency, that Nthc formation becomes less
sensitive to ambient variations as premixing is approached.
Norster, E.R. and DePiefto, S.M., 1996, Dry Low Emissions
Combustion System for EGT Small Gas Turbines, Institution of
Diesel and Gas Turbine Engineers.
Norster, E.R. and DePietro, S.M. et al, 1996, Development of a
Dry Low NOx Combustion System for the EGT Typhoon,
PowerGen 1996.
Kowlcabi, M. et al, 1997, The Development of a Dry Low NOx
Combustion System for EGT Typhoon, to be presented at ASME 97
Visser, B.M and Bahlmann, F.C., 1994, Variation in the NOx
emission of Gas Turbines, ASME 94-GT-261
Alkabie, H.S. and Norster, E.R, 1994, ISO-Correction Formula
for G30-NOx Emission, EGT Internal report.
Lewis,0.D,1981, Prediction of NOx Emissions,ASME81-GTI19.
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 12/29/2014 Terms of Use: http://asme.org/terms