Available online at www.sciencedirect.

com

Fuel 87 (2008) 21622167

www.fuelrst.com

Combustion performance test of a new fuel DME to adapt to a

gas turbine for power generation
Min Chul Lee *, Seok Bin Seo, Jae Hwa Chung, Yong Jin Joo, Dal Hong Ahn
Power Generation Laboratory, Korea Electric Power Research Institute, Korea Electric Power Corporation, Yusung-Gu, Taejon 305-380, Republic of Korea
Received 3 April 2007; received in revised form 22 November 2007; accepted 25 November 2007
Available online 2 January 2008

Abstract
Recently, DME (dimethyl ether, CH3OCH3) has attracted a great deal of attention as an alternative fuel owing to its easy transportation and cleanliness. This study was conducted to verify the combustion performance and to identify potential problems when DME is
fueled to a gas turbine. Combustion tests were conducted by comparing DME with methane, which is a major component of natural gas,
in terms of combustion instability, NOx and CO emissions, and the outlet temperature of the combustion chamber. The results of the
performance tests show that DME combustion is very clean but hard to control. The CO emission level of DME is lower than that of
methane, while the NOx emission level of DME is as low as that of methane. When ring DME, the pressure uctuation in the combustion chamber caused by combustion instability is lower than that occasioned when ring methane. From the results of the outlet temperature of combustor we have ascertained that DME combustion is more likely to ash back than methane combustion and this
property should be considered when operating a gas turbine and retrotting a burner.
2007 Elsevier Ltd. All rights reserved.
Keywords: DME (dimethyl ether); Alternative fuel; Gas turbine combustion

1. Introduction
1.1. Backgrounds
Environmentally friendly and new fuel DME, which has
appeared on the scene as a solution for the shortage of fossil fuels and worsening environmental problems is highly
attracted. DME has the potential of being put into practical use within a few years because of the advantages it
oers in terms of easier storage and transportation, fewer
pollutant emissions, and the option of utilizing existing
infrastructures of LPG, since the physical characteristics
of DME are similar to those of LPG. So, a lot of studies
have been conducted to estimate the DME as an alternative
fuel in terms of availability, performance, economics, technology, versatility, national security and environmental
emissions [14].
*

Corresponding author. Tel.: +82 42 865 5488; fax: +82 42 865 5489.
E-mail address: lmc@kepri.re.kr (M.C. Lee).

0016-2361/$ - see front matter 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2007.11.017

For the mass production of DME, mass consumers such

as thermal power plants are required. In addition to such
necessities, this study can impact on the diversication of
power generation fuel, the reduction of power generation
costs, and the reduction of air pollutant emissions.
1.2. Fuel characteristics of DME
DME is a clean fuel that can be produced from diverse
materials including natural gas, coal, coal-bed methane,
biomass, etc. DME is expected to be used as an alternative
fuel in near future because DME from biomass, so-called
Bio-DME is one of the very good solutions for mitigating
climate change problem. In addition, DME from natural
gas resources available from presently unused small or
middle scale gas wells might oer the economical production cost. Table 1 shows the physical and chemical properties of DME in comparison to those of other fuels. The
lower explosive limit is higher, while saturated vapor pressure, boiling point, and specic gravity are similar to those

M.C. Lee et al. / Fuel 87 (2008) 21622167

2163

Nomenclature
DME
LNG
LPG
DLN
Re
U
Dh
Dburner

dimethyl ether
liqueed natural gas
liqueed petroleum gas
dry low NOx
Reynolds number
mean velocity at the combustor inlet
hydraulic diameter
out outer diameter of burner

Dburner in inner diameter of burner

Dswirl in inner diameter of swirler
Dswirl out outer diameter of swirler
v
kinematic viscosity
ppm
part per million
/
equivalence ratio
T3
exit gas temperature of combustion chamber

Table 1
Properties of DME, methane, propane, and butane

Chemical formula
Boiling (C)
Liquid density (g/cm3)
Specic gravity (vs, air)
Vapor pressure (atm @ 0 C)
Flammable limits in air (%)
Ignition temperature (C)
Max. burning velocity (cm/s)
Equivalent A/F ratio (kg/kg)
Low heating value (MJ/kg)

DME

Methane Propane Butane

CH3OCH3
25.1
0.67
1.59
6.1
3.417
235
50
9.0
28.8

CH4
161.5
0.42
0.55
246
515
650
37
16.9
49.0

C3H8
42
0.49
1.52
9.3
2.19.4
470
43
15.7
46.3

C4H10
0.5
0.57
2.01
2.4
1.98.4
365
43
15.5
45.7

ston C30; 30 kW of output power) with DME and LPG,

and reported good interchangeability, performance and
operational stability [7]. Even though it is very important
to observe the combustion characteristics through ame
visualization when new fuel is adapted to gas turbines,
none of the above studies report any data about ame
structure of DME ame, primarily because it is dicult
to make visual combustor chamber observations. This
study reports about the ame images that show signicant
dierences of combustion characteristics between DME
and other fuels.
2. Combustion test apparatus

of propane and butane. Thus DME can be handled like

LPG which primarily contains propane and butane. However DME has a faster combustion speed, a lower autoignition temperature and lower heating value. The use of
DME may also cause problems such as ashback and combustion instability when applied to a DLN (dry low NOx)
gas turbine. Therefore, this study is being conducted to
evaluate the fuel performance as a gas turbine fuel and to
quantify the potential applications problems.

1.3. Prior studies on DME red gas turbines

To use DME in power plants, the fuel feed system and
combustor have to be retrotted. To this end, Hitachi of
Japan has developed a 25 MW class DME combustor to
test the various shape design variables of combustors with
multiple cluster burners in terms of combustion vibration,
exhaust gas, and combustion eciency. The results of these
tests were analyzed by signal/noise ratio analysis to obtain
an optimal shape [5]. Mitsubishi heavy industries (MHI)
has carried out combustion exhaust gas measurements with
a G-type combustor and reported that DME has exhaust
characteristics that are similar to those of natural gas. General electric (GE) has conducted combustion experiments
with a mixture of DME, methanol, and water using a
bench-scale combustor, and reported that such mixtures
can be used in commercial GE combustors [6]. GE has
obtained patent rights for the mixture ratio and scope of
use. Tokyo Electric Power Corporation (TEPCO) conducted the performance tests on micro gas turbines (Cap-

2.1. Introduction to the gas turbine combustion test facility

In this study, an atmospheric gas turbine combustion
test facility (60 kW of output power) was constructed for
the combustion testing of DME and methane. Fig. 1 shows
a schematic of the air compressor, air receiving tank, air
pre-heater, water sprinkler for cooling exhaust gas, atmospheric pressure combustor, control system, and external
stack, in addition to the DME supply system. The supply
system is composed of a DME tank, a DME vaporizer
and DME pre-heating line. By providing a critical ow
conditions for both the fuel and air in the upstream of
the combustor, the pressure uctuations generated in the
combustion dont aect the conditions in the upstream,
thus preventing the perturbation of air and fuel ows.
2.2. Model gas turbine combustor
The atmospheric DLN combustor was designed and
fabricated by simplifying the commercial gas turbine combustor on a small scale. Fig. 2 shows the design drawing of
the model gas turbine combustor and the part names
shown in Fig. 2 are described in Table 2. This premixed
combustor is constituted with a dump type combustion
chamber in the downstream, and a pilot fuel and a fuel
air mixture are supplied through a double pipe at the end
of the dump chamber. A combustion chamber was made
with a quartz tube to visualize the ame. The tube was
sealed with a graphite ring at the connection with the burner. A spark ignitor was installed in the middle of the

2164

M.C. Lee et al. / Fuel 87 (2008) 21622167

Fig. 1. Gas turbine (60 kW scale) combustion test facility.

chamber, as shown in Fig. 2b, and cooling air was supplied

around the chamber to prevent overheating. Fuelair premixed gas was produced by providing the air with a rotating movement in its ow direction using a swirler equipped
with vanes at tilted angle (h) of 45, and with four fuel
spokes, each of which had two injection holes of 0.7 mm
in diameter and, designed to operate at sonic ow
conditions.
3. Test methods, results and discussion
3.1. Test methods
The combustion characteristics of DME were compared
with those of methane by measuring the dynamic pressure
generated by combustion instability, pollutants including
NOx and CO, and temperatures at the combustion chamber outlet. The behavior of the ame in relation to the burner geometry was observed by varying the position of the
center blu body at the center of the burner.
3.2. Comparison test of the combustion characteristics of
methane and DME
Figs. 35 show the combustion vibration characteristics
of the methane and DME according to the equivalence
ratio and Reynolds number which are signicant parameters when designing combustor and testing performance.

The Reynolds number which is in proportion to the turbulent intensity was calculated with the averaged inlet velocity of the premixed gas into the combustion chamber and
the hydraulic diameter which is dened as Eq. (1) [8,9]
Reynolds number; Re

U Dh U Dburner

out

 Dburner in

m
1

where U is mean velocity at the combustor inlet, Dh is the

hydraulic diameter, Dburner out is the outer diameter of burner, Dburner in is the inner diameter of burner, and v is the
kinematic viscosity.
The dynamic pressure uctuation generated by combustion instability in the combustion chamber was measured
at the rate of 50,000 samples per second, for which the root
mean square value is shown in Fig. 3. It could be veried
that, in general, DME generates less combustion oscillation
than methane. This is because the molecular weight of DME
is three times greater than that of methane, which shapes the
ame in a lengthwise extension by inertia, and thus this distribution of heat release induces less combustion oscillation.
In addition, it may also be that DME energy density in the
gas phase is considerably less than methane.
Fig. 4 shows that the NOx increases qualitatively in
accordance with the increase of the equivalence ratio, as
expected if thermal NOx production is the primary source.
This result meets well with that the thermal NOx formation

M.C. Lee et al. / Fuel 87 (2008) 21622167

2165

Fig. 2. Drawings of a model DLN combustor.

Table 2
Description of a DLN combustor of Fig. 2

Part name

Material

r
s
t
u
v
w
x
y
z
P0, P1, P2, P3
T1, T2, T3

Center blu body with pilot fuel injection

Combustion air inlet
Air ow swirler
Fuel injection module with fuel spoke
Swirling module
Cooling air injection module
Main fuel inlet
Cooling air inlet
Combustion chamber
Dynamic pressure sensor
Temperature sensor

SUS304
SUS304
SUS304
SUS304
SUS304
SUS304
SUS304
SUS304
Quartz

is proportion to the equivalence ratio because the more fuel

input gets the higher temperature of ame in fuel lean premixed condition (Zeldovich Mechanism [10]). Both DME
and methane give superior NOx exhaust characteristics at
less than 35 ppm (at 15% O2).
Fig. 5 shows that DME generates far less CO than methane. This can be understood as the result of the excellent
combustibility expected by the lower ignition temperature,
and the high combustion speed as shown in Table 1. In
summary, DME is a clean fuel for a gas turbine, which
can be perfectly burnt, generating low levels of NOx.

= 0.60, CH4
= 0.70, CH4
= 0.80, CH4

P'_rms [kPa]

Number at Fig. 2

= 0.60, DME
= 0.70, DME
= 0.80, DME

0
10000

12000

14000

16000

18000

20000

Reynolds Number
Fig. 3. Comparison of pressure uctuation characteristics of methane and
DME for / = 0.60.8.

3.3. Combustion characteristics according to the combustor

blu body shape
Flame stabilization tests with a center blu body in the
three positions shown in Fig. 6 were carried out in a manner similar to that described in Section 3.2. Although no
peculiarities were identied in terms of NOx and CO

2166

M.C. Lee et al. / Fuel 87 (2008) 21622167

=0.6, CH4
=0.7, CH4
=0.8, CH4

40

30

NOx [ppm @ 15%O2]

NOx [ppm @ 15%O2]

35

=0.6, DME
=0.7, DME
=0.8, DME

30
25
20
15
10

Recessed, DME
Plane, DME
Protruded, DME

Protruded, CH4

25
20
15
10
5

5
0
10000

Recessed, CH4
Plane,CH4

0
12000

14000

16000

18000

20000

0.55

0.60

Reynolds Number
Fig. 4. Comparison of NOx emission characteristics of methane and
DME for / = 0.60.8.

35

=0.6, CH4
=0.7, CH4
=0.8, CH4

=0.6, DME
=0.7, DME
=0.8, DME

20
15
10
5

14000

16000

0.80

18000

Recessed, DME

Recessed, CH4
Plane,CH4

20

12000

0.75

25

25

0
10000

0.70

Fig. 7. NOx emission characteristics for methane and DME combustion

varying with the location of the center blu body (/ = 0.550.8).

CO [ppm @ 15%O2]

CO [ppm @ 15%O2]

30

0.65

Equivalence Ratio

Plane, DME
Protruded, DME

Protruded, CH4

15

10

20000
0

Reynolds Number

0.55

0.60

0.65

0.70

0.75

0.80

Equivalence Ratio

Fig. 5. Comparison of CO emission characteristics of methane and DME

for / = 0.60.8.

Fig. 8. CO emission characteristics for methane and DME combustion

varying with the location of the center blu body (/ = 0.550.8).

1500

Temperature [ oC ]

emissions as shown in Fig. 7 and Fig. 8, the results of the

temperature measurements taken at the combustion chamber outlet, as shown in Fig. 9, were noteworthy. While the
temperature at the downstream end of the combustion
chamber was increased almost linearly in proportion to
the equivalence ratio for positions (b) and (c) of Fig. 6, it
was observed that the exit gas temperature (T3) was
decreased at an equivalence ratio of between 0.65 and 0.7
for position (a) of Fig. 6. This is because the ow is recirculated at the recessed position and the ame exists
before the dump surface away from the temperature
measuring point, T3. This result is indicative of ash back

1450

Recessed, CH4
Plane,CH4

1400

Protruded, CH4

Recessed, DME
Plane, DME
Protruded, DME

1350
1300
1250
1200
1150
1100
0.55

0.60

0.65

0.70

0.75

0.80

Equivalence Ratio
(a) Recessed C.B.

(b) Plane C.B.

(c) Protruded C.B.

Fig. 6. Three types of center blu body.

Fig. 9. Exit gas temperature (T3) for methane and DME combustion
varying with the location of the center blu body (/ = 0.550.8).

M.C. Lee et al. / Fuel 87 (2008) 21622167

(a) Recessed Center Body(Methane)

(b) Recessed Center Body(DME)

(c) Plane Center Body(Methane)

(d) Plane Center Body(DME)

(e) Protruded Center Body(Methane)

(f) Protruded Center Body(DME)

2167

Fig. 10. Flame images for three types of center blu body at / = 0.7.

and stabilization of the ame structure within the recessed

stabilization region, and the ash back was visually
observed and showed in Fig. 10b. This phenomenon is
primarily because the ignition delays of DME/air mixtures
are much shorter than those of methane/air mixtures [11].
In addition, from the Fig. 7, an increase in NOx emissions
is also observed under this condition. Flame is moved into
the fuel nozzle and shortening the available mixing time for
fuel and air. Higher local ame temperatures are produced
as the combustion is initiated before leaner premixed conditions can be established with overall leaner equivalence
ratios. Thus high thermal NOx emissions result.
4. Conclusions
Combustion performance tests of DME and methane at
atmospheric pressure were carried out successfully. From
the results presented in this paper, the following conclusions are obtained.
(1) DME is a very good fuel for power generation with
low combustion instability and low NOx and CO
emissions. However depending upon the operational
conditions, excessive NOx and CO may be emitted
and high combustion oscillation may be occurred.
Therefore, the operation map and valve control
schedules should be modied according to the gas
turbine power output.
(2) Considering low ue gas temperature at the combustion chamber outlet (T3) of DME combustion, the
reduction of thermal failure of turbine blades and
nozzles is expected when operating DME fueled gas
turbines.

(3) From the experimental results for the center body

recessed case, thermal failure of the fuel nozzle and
burner is anticipated by the ash back when ring
DME to the gas turbines designed for natural gas.
Therefore the retrot of combustor is necessary to
use DME as a gas turbines fuel.
References
[1] Semelsberger TA, Borup RL, Greene HL. Dimethyl ether as an
alternative fuel. J Power Source 2006;156:497511.
[2] Gokalp I, Lebas E. Alternative fuels for industrial gas turbines
(AFTUR). Appl Therm Eng 2004;24:165563.
[3] Song J, Huang Z, Qiao X, Wang W. Performance of a controllable
premixed combustion engine fueled with dimethyl ether. Energy
Conver Manage 2004;45:222332.
[4] Cocco D, Tola V, Cau G. Performance evaluation of chemically
recuperated gas turbine (CRGT) power plants fuelled by DME.
Energy 2005;31:114658.
[5] Saitou T, Inoue H, Kobayashi N, Watanabe T. Development of a
multi-cluster burner for fuel grade DME. Proceedings of ASME
Turbo EXPO 2004; GT2004:53689.
[6] Basu A, Gradassi M, Sills R, Fleisch T, Puri R. Use of DME as a
gas turbine fuel. Proceedings of ASME Turbo EXPO 2001;
GT2001:0003.
[7] Tsuchiya T, Okamoto M. Evaluation and demonstration of DME as
an alternative fuel for micro-gas turbines. Proceedings of the Asian
Congress on gas turbines 2005; ACGT2005:066.
[8] Incropera FP, Dewitt DP. Introduction to heat transfer. 3rd ed. New
York: John Wiley & Sons Inc.; 1996. p. 41226.
[9] Bejan A. Convection heat transfer. 2nd ed. New York: John Wiley &
Sons Inc.; 1995. p. 93107.
[10] Warnatz J, Maas U, Dibble RW. Combustion. 2nd ed. Germany: Springer; 1999. p. 21931.
[11] Chen Z, Qin X, Ju Y, Zhao Z, Chaos M, Dryer FL. High temperature
ignition and combustion enhancement by dimethyl ether addition to
methaneair mixtures. Proc Combust Inst 2007;1:121522.