Applied Thermal Engineering 146 (2019) 328–337

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Experimental study on applying biomass-derived syngas in a microturbine T

a b,⁎ a
Paulo Sérgio Pedroso Corrêa Jr , Jianan Zhang , Electo Eduardo Silva Lora ,
Rubenildo Vieira Andradea, Luis Roberto de Mello e Pintoa, Albert Ratnerb
a
Department of Mechanical Engineering, Federal University of Itajuba, Pinheirinho, Itajubá CEP 37500-903, MG, Brazil
b
Department of Mechanical and Industrial Engineering, The University of Iowa, Iowa City, IA 52242, USA

HIGHLIGHTS

• Studied the performance of a 30 kW microturbine using syngas/natural gas blends.

• The system efficiency, gas flow rate, temperature, and emissions were presented.
• The microturbine burner was tested in an optical-access high-pressure chamber.
• Used OH∗ chemiluminescence and OH-PLIF to explain the CO and NOx emission trends.
• The change in gas components could vary the acoustic performance of the burner.

ARTICLE INFO ABSTRACT

Keywords: The current research aims at studying the effect of using biomass-derived syngas as a supplement of the natural
Microturbine gas in a microturbine. The study includes microturbine performance and burner tests. Microturbine performance
Syngas test used the blends of natural gas (NG) and the syngas (syngas A) from the air gasification of rice hulls. The
Biomass gasification microturbine test focuses on turbine performance features that include turbine efficiency, outlet temperature,
Emission
air/fuel flow rate, and NOx and CO emissions. In contrast, the burner tests were performed to interpret the
Combustion instability
microturbine test results and provide guide information on using high hydrogen content syngas (syngas B) in the
microturbine. Microturbine tests show that the efficiency drops (about 13%) when the fuel changed from pure
natural gas to 50% NG/50% syngas A blends. Furthermore, the temperature analysis shows that the temperature
at the turbine outlet experiences negligible variations despite the change in fuel composition. The emission
results indicate that when the output power is sufficiently high, the NOx emission keeps unchanged, whereas the
CO emission increases with the syngas addition. Burner tests examined the flame by utilizing OH∗ chemilu-
minescence and planar laser induced fluorescence of OH radicals. Results demonstrate that the increase in CO
emission is related to the incomplete combustion, whereas the unchanged NOx emission is associated with the
local hot zone in flames. Moreover, at a forcing frequency of 125.6 Hz, phase-averaged OH∗ images, global heat
release oscillation, and the Rayleigh index analysis indicate that the variation of fuel composition can change the
response of the flame to acoustic perturbation.

1. Introduction value. Therefore, it is a great challenge for existing systems to use

Biofuels derived from biomass, despite their lower calorific value,
are becoming increasingly attractive alternatives for the energy in-
dustry. One of the major biomass energy conversion technologies is
gasification, which produces syngas containing flammable gases such as
H2, CO, and CH4 [1,2]. The components of syngas may change due to
the characteristics of biomass, the type of gasifier, the gasification fluid,
and the operating parameters [2]. In addition to the variable compo-
sition, another fundamental feature of the syngas is its low heating
syngas.
Although the biomass-derived syngas has a low heating value, it still
shows a great potential in distributed generation systems [3]. More-
over, a microturbine system has been demonstrated to be suitable for
distributed generation systems due to its low-emission and fuel-flex-
ibility characteristics [4–6]. Therefore, the application of syngas in
microturbine systems has gained both academic and industrial atten-
tion in recent years [7–12]. However, some characteristics of the
syngas, such as its variable composition and low heating value, may

⁎
Corresponding author.
E-mail address: jianan-zhang@uiowa.edu (J. Zhang).

https://doi.org/10.1016/j.applthermaleng.2018.09.123
Received 25 March 2018; Received in revised form 20 June 2018; Accepted 27 September 2018
Available online 28 September 2018
1359-4311/ © 2018 Elsevier Ltd. All rights reserved.
P.S.P. Corrêa et al.
affect the operation stability of microturbines, which are normally op-
erated with natural gas that has a higher heating value and a stable
composition. Research has shown that using syngas can affect the
combustion characteristics of the microturbine burner, such as the
flame shape, the combustor temperature, and the emission. Lee et al.
[13]. and Baina et al. [7]. showed that the change in syngas composi-
tion can affect the combustor temperature and the system emission.
Furthermore, it has been demonstrated that the acoustic characteristics
of the combustor are also affected by the variation in syngas compo-
sition [13,14].
Moreover, it is also important to examine the effect of adding syngas
on the system performance of microturbines, such as the output power
and the overall efficiency. With a modeling work, Król et al. concluded
that the air flow rate, the net output power, and the temperature at the
turbine exit are critical parameters that can be used to check the system
performance [9]. Furthermore, they showed that the system efficiency
was not significantly affected by adding syngas. A similar trend was
detected by a modeling work conducted by Liu and Weng [15]. Rabou
et al. validated a model with an experimental study, which demon-
strated that the output power only declined slightly (1%) when using
the syngas with a low heating value (∼15 MJ/m3) [10]. Othman and
Boosroh found that using syngas can increase NOx and CO emissions
[16].
Although the syngas has a great potential in microturbine systems,
the research (especially experimental study) on this topic is still very
limited. Moreover, most research on the application of syngas in mi-
croturbines is either examining the burner [13] or the system [10].
There is still no work conducted to examine both the burner and system
performance. Therefore, the main task of the current research is ex-
perimentally examining the effects of using syngas on the system per-
formance and the combustion characteristics of the burner. The testing
results will be used to evaluate the potential of a microturbine to use
syngas as the supplement of natural gas.
Fig. 1. Microturbine testing system (the figure is not to scale).
329
Applied Thermal Engineering 146 (2019) 328–337
2. Experimental systems
2.1. Microturbine testing system
The microturbine testing system is located at the Federal University
of Itajuba, Brazil. The schematic of the microturbine testing system is
shown in Fig. 1. During the test, the gasifier was running simulta-
neously with the microturbine.
A bubbling fluidized bed gasifier, which fed by rice hulls, was used
to generate the syngas. The gasification takes place in a reactor, which
has an operating capacity up to 247 kWth and a maximum operating
temperature of 1200 °C, that has a height of 6000 mm and an out-dia-
meter of 990 mm. The reactor, which is made of sheet steel and covered
by refractory and insulation material, can be divided into two regions
that are named as bed and freeboard. The bed region, which has an
inner diameter of 407 mm and a height of 2660 mm, is filled with the
aluminium oxide powder (the bed itself) and contains a distributor
plate with injectors through which air is introduced into the bed. The
freeboard region is 3000 mm high and is used for reducing the velocity
of particles that are entrained in the air flow. The temperature was
monitored by thirteen K-type thermocouples that were installed in the
wall of the reactor.
The rice hulls were supplied to the reactor by a screw conveyor,
whereas air was provided to the gasifier reactor by an air compressor.
The produced syngas from the reactor was directed to a cyclone, which
has an outer cooling jacket where the larger particles were separated
[17]. After passed through a gas-washing tower, the syngas was di-
rected to a Venturi type scrubber, in which the residues were collected
in a separator. The main contaminants removed from the syngas were
tar and particles. The particulate concentration after cleaning is less
than 30 mg/Nm3, and the tar concentration is less than 100 mg/Nm3,
which ensures that there is no notable impact on the microturbine
performance. Once the syngas was cleaned, the volume flow rate of the
P.S.P. Corrêa et al. Applied Thermal Engineering 146 (2019) 328–337

Table 1 hydrogen concentration is high in the syngas. Therefore, the current

Properties of rice hulls. burner test also provides useful guide information for the future turbine
Ultimate analysis (wt%) Proximate analysis (wt%) test with hydrogen content syngas. Finally, the third goal is providing
preliminary results of the acoustic response of the burner using hy-
C H N O Ash Moisture Volatile Fixed Ashes drogen content syngas.
matters Carbon Therefore, besides syngas A, the syngas produced with the steam
41.74 4.3 0.4 36.63 16.93 10.59 57.62 14.86 16.93 gasification method, which is referred to as syngas B, was also used in
burner tests. The components of syngas B were calculated based on the
work of Hofbauer et al. [19]. Syngas B has an LHV of 12.55 MJ/m3.
Moreover, considering that fact that the composition of NG in the US is
Table 2 different from that in Brazil, the burner test used methane that is
Gas components of syngas A and B (all units are in vol%). convenient for burner testing. Some fuel properties of used NG and
Syngas A Syngas B [19] methane are compared in Table 3 and the results show that the NG used
in microturbine tests has similar properties to that of methane. More-
H2 5.1 38.1 over, the technical reference of Capstone C30 fuel requirements also
CO 13.4 28.1
implies that the burner performance won’t be significantly affected
CH4 1.8 8.6
CO2 22 22.2
when the fuel switches from NG (that has over 90% methane) to me-
N2 57.7 – thane [20]. In our tests, the NG (in Brazil) has 89.24% methane, which
C2H4 – 3 is sufficiently close to the 90% limit. Therefore, the change from NG to
methane is believed to have a minor impact on the burner performance.

syngas was measured. The syngas produced at low pressure was then
compressed by a Scroll compressor and sent to mix with the natural gas
(NG) at a pressure of approximately 4 bar. The blended gas was then
delivered to the burner of the microturbine, which is Capstone model
C30. The microturbine, which has the maximum output power of
30 kW, is operating in a Brayton cycle with heat regeneration. The
schematic of the microturbine unit is shown in Fig. 1, whereas the
detailed geometry information can be found in the work of Bohn [18].
Air flew in the compressor at the atmospheric pressure and 21 °C and
then passed the regenerator to raise the temperature. After that, the
high temperature air was supplied to the burner of the combustor. The
burnt flue gas, whose temperature is T1 at the combustor outlet, first
flew into the regenerator to heat the air, and then was exhausted out of
the gas turbine unit at a temperature of T2. The emissions, which in-
clude CO and NOx, were measured by a Lancom III flue gas analyzer.
The ultimate analysis and proximate analysis of rice hulls were
performed according to ASTM D3176 and ASTM D5142, respectively.
The obtained values are listed in Table 1. The heating value of rice
hulls, which is 13.29 MJ/kg, was measured with a calorimeter pump
(IKA Works C2000). The syngas from rice hulls gasification with air is
referred to as syngas A in later sections. The components of syngas A,
which are listed in Table 2, were measured with a gas analyzer (Sie-
mens ultramat 23). The lower heating value (LHV) of syngas A is
2.87 MJ/m3.
2.2. High-pressure burner testing system
The high-pressure burner test was conducted at the University of
Iowa, USA. The burner test has three goals. The first one is interpreting
the results of the turbine performance test. The second one is testing the
capability of the burner to use high-hydrogen content syngas. The
current microturbine tests haven’t included syngas with high hydrogen
content for the safety concern that flame flashback may occur once the
2.2.1. Microturbine burner
The schematic of the Capstone microturbine burner and its dimen-
sion information are illustrated in Fig. 2(a). Air flows in the burner
through two types of holes shown in the upper left of Fig. 2(a). There
are four straight holes and 15 swirl slots that create swirling air flow.
The burner has two fuel inlets. The first inlet is used for supplying gas
fuel for the 12 fuel holes that shown in the upper right of Fig. 2(a). The
fuel holes are designed for mixing fuel gas and air so that this part of the
fuel is running in the premixed mode. In contrast, the other fuel inlet is
connected to the pilot fuel tube, which ensures the other part of the fuel
is operating as the non-premixed mode. The design aims to achieve low-
emission by employing the premixed mode while maintaining stable
combustion by using the non-premixed pilot flame.
2.2.2. High-pressure combustion chamber
The experimental configuration used for burner tests is illustrated in
Fig. 2(b). All burner tests were conducted at a gauge pressure of 3 bar,
which is close to the microturbine combustor one. The high-pressure
chamber has a height is 1850 mm and an internal diameter of is
300 mm. A more detailed description of the chamber can be found in
previous work [21,22]. The burner is placed in an air chamber that is
attached to the high-pressure chamber, which equips with quartz
windows that are used for optical measurement. The top of the high-
pressure chamber is acoustically closed and the exhaust flow is con-
trolled by three adjustable electronic valves which provide the ability to
pressurize the chamber continuously. The acoustic perturbation system
contains a signal generator, a power amplifier, and four loudspeakers
that were used to supply in-phase acoustic perturbation. The flow rate
of air and methane were controlled with two Hastings flow meter/
controllers with an accuracy of ± 1.4% and ± 1%, respectively. The
flow rate of syngas was adjusted with an Omega mass flow controller
(Omega 5400) with an accuracy of ± 1.5%.

Table 3
Fuel properties of NG and methane.
NG Methane Difference

Components (proportions are all by volume) CH4 (89.24%), C2H6 (7.86%), C3H8 (0.24%), CO2 (1.25%), N2 (1.34%), O2 (0.07%) CH4
Higher heating value (MJ/m3) 37.2 35.8 −3.7%
Lower heating value (MJ/m3) 41.1 39.7 −3.4%
Stoichiometric air/fuel ratio (kg/kg) 16.1 17.19 6.8%
Maximum laminar flame speed (cm/s) 38.4 38.3 −0.3%
Maximum adiabatic flame temperature (K) 2225.2 2224.4 −0.3%

330
P.S.P. Corrêa et al.
Fig. 2. Microturbine burner and experimental configuration (the figure is not to
scale).
2.2.3. Diagnostic system
A high-resolution piezoelectric pressure sensor (PCB 106B) is in-
stalled in the chamber wall to monitor the pressure oscillation, which
was recorded at a rate of 20 kHz. The flame was measured with two
methods that are OH* chemiluminescence and planar laser induced
fluorescence of hydroxyl radicals (OH-PLIF), which provide a line-of-
sight and a cross-section view of the flame, respectively.
In the OH-PLIF system, which is shown in Fig. 3, a 10 Hz Nd:YAG laser
(Continuum Powerlite 9010) supplies a 532 nm beam, which was
changed to a UV beam with a wavelength of 283 nm by a dye laser
(Continuum ND6000) and a frequency doubler. Finally, the UV beam,
with an intensity of 30 mJ/pulse, was used for exciting the A-X(1, 0) band
of OH radicals. OH-PLIF images were captured with an ICCD camera
(Princeton Instruments) equipped with a narrow bandpass filter
(312.6F10-10). The ICCD camera, which was operating at a frequency of
3 Hz, was set to capture images in an 80 mm × 80 mm region with a
resolution of 512 × 512 pixels. The exposure time of taking OH-PLIF and
OH* chemiluminescence images is 150 ns and 60 μs, respectively. 100
images were captured for stable flames, whereas 600 images were used
for acoustically perturbed flames. More detailed information of the layout
of the flame diagnostic system can be found in previous work [23].
During the test, the camera was triggered by the Nd:YAG laser so
that the OH-PLIF image can be recorded correctly. When the camera
was triggered and the gate opened, it generated a pulse signal that was
sent to the computer. Consequently, the gate pulse from the camera and
the pressure signal were recorded simultaneously with two different
channels. The delay between pressure signal and the gate pulse signal is
approximately 9 ns. By doing this, the phase angle of pressure at which
the image is captured can be determined. Images with a similar phase
331
Applied Thermal Engineering 146 (2019) 328–337
Fig. 3. Schematic of the OH-PLIF diagnostic system.
angle were then distributed to a 30-degree-wide phase bin. Finally, all
images in the same phase bin were averaged to get phase-averaged
results.
3. Results and discussions
3.1. Microturbine system performance test
As the supplement of Natural gas (NG), syngas A was mixed with NG
with different proportions, which are 0%, 25%, and 50% by volume, for
the microturbine test. The gas microturbine performance was examined
at different partial loads, which increase from 0 kW to 30 kW by a step
of 5 kW in a period of 10 min. 0 kW represents the idling condition.
When the microturbine was operating at 0 kW, fuel gas was only sup-
plied to the pilot fuel tube, therefore the burner was under the non-
premixed combustion mode. Once the power is higher than 0 kW, both
premixed fuel holes and the pilot fuel tube were used. As a result, the
burner was under the partially premixed and partially non-premixed
mode. The measurement uncertainty is provided in Table 4.
3.1.1. Results
Fig. 4(a) demonstrates that more fuel gas is needed when NG is
blended with syngas A. It is because more fuel gas is needed to maintain
the thermal load. As the fuel gas flow increases, more air is required for
completed combustion. Therefore, the air flow rate also shows an in-
creasing trend shown in Fig. 4(b).
Fig. 5 shows the efficiency and the revolutions per minute (RPM) of
the microturbine. Fig. 5(a) shows that the microturbine using only NG
has an efficiency of 26% at the maximum load. In contrast, with the
75% NG/25% syngas A blend, the efficiency drops to 24.2%. Using the
blend of 50% NG/50% syngas A, the efficiency decreases further to
22.5%. A similar trend is also detected in the maximum output power,
which declines from 26.88 kW to 22.94 kW when the syngas proportion
increases from 0% to 50%. Fig. 5(b) illustrates that the microturbine
starts generating electricity from 45,000 rpm and reaches a peak value
of 96,000 rpm. In spite of the change in fuel composition, the micro-
turbine was operating in a stable status in all conditions without a
significant change in performance. It implies that the syngas can be
P.S.P. Corrêa et al.
Table 4
Uncertainty of microturbine performance tests.
Rotation per minute Power Temperature
Uncertainty 2.3% 14.1% 3.7%
Fig. 4. The flow rate of fuel gas and air.
Fig. 5. The microturbine efficiency (a) and RPM at different output power (b).
used as a supplement of natural gas in microturbines. However, the
resulting drop in efficiency and output power must be taken into ac-
count.
Fig. 6 illustrates the temperature at the turbine outlet (T1) and the
system outlet (T2), which were measured at the two locations marked
in Fig. 2. The change in fuel composition has a negligible effect on T1,
whereas T2 increases once the output power increases to around 20 kW.
The increase in T2 is related to the gas flow rate change shown in Fig. 4,
which demonstrates that the overall gas flow increases once the syngas/
NG blends were used. The increased gas flow rate will generate more
flue gas. However, the regenerator hasn’t been modified to handle the
increased flue gas so the energy in flue gas cannot be recovered effi-
ciently, which consequently result in a higher exhaust temperature T2
at the outlet of the regenerator. When the output power is the same,
larger air input means that the compressor will consume more energy,
which consequently causes a relatively low efficiency detected in
Fig. 5(a).
332
Applied Thermal Engineering 146 (2019) 328–337
Fuel flow Air flow Efficiency NOx CO
11.6% 0.3% 17.4% 2% 2%
Fig. 7 shows NOx and CO emission results. The sharp reduction in
measured values after the peak is a consequence of the shift from a non-
premixed combustion regime to a partially premixed partially non-
premixed one. When the idling condition is not considered, it can be
seen that both NOx and CO emission drops with the increasing output
power. Furthermore, Fig. 7(a) shows that cases with syngas have a
higher CO emission than that with pure NG. In contrast, the NOx
emission illustrated in Fig. 7(b) is higher in cases using syngas when the
output power is less than 20 kW, above which the NOx emission is al-
most the same for all cases. It is interesting to find that under high load
conditions (> 20 kW) the NOx emission is almost the same even when
CO increases. It is controversial to traditional trend found in combus-
tors, which is NOx is usually decreases when CO increases [24]. Similar
effects of using syngas as the fuel on emissions are also detected in the
work of other researchers [10,13,16]. However, there is no explanation
of the change in emission in previous work because of the lack of the
detailed information of the flame.
P.S.P. Corrêa et al. Applied Thermal Engineering 146 (2019) 328–337

Fig. 6. The temperature at the combustor outlet (T1) and the microturbine outlet (T2).

Fig. 7. NOx and CO emissions.

3.2. Burner test
The burner test investigated the flame shape and the acoustic re-
sponse of the flames. The flame shape of different cases is examined
with mean OH* chemiluminescence images and OH-PLIF images, which
provide the light-of-sight global view and the cross-section view, re-
spectively. As mentioned previously, the burner test aims to interpret
the microturbine test results, especially the emission results, and pre-
dict the performance of microturbine using high hydrogen content
syngas.
3.2.1. Operating conditions
The burner study used methane/syngas blends with different pro-
portions (by volume) of syngas. It needed to be mentioned that only the
cases corresponding to the output power of 5–20 kW were examined,
because a further increase in the power will cause a high risk of over-
heating the acoustic system. The details of the mass flow rate (in g/s)
and the volume flow rate (in SLPM) for different tests are listed in
Table 5.
3.2.2. Flame shape
The mean OH* chemiluminescence images are shown in Fig. 8. The
location of the burner is shown schematically in the first picture of the
first row of Fig. 8. For all examined cases, flames are lifted above the
burner. In each row, it is found that the OH* intensity increases with a
rising value of the output power, which indicates that the global heat
release rate is increasing [14]. At the same power level, the OH* che-
miluminescence of CH4/syngas A blends flames is notably weaker than

Table 5
Mass flow rate of gases (flow rates are in units of g/s and SLPM in parentheses).
5 kW 10 kW 15 kW 20 kW

Air 7.15 (332) 10.73 (497.8) 14.29 (663) 17.85 (828.4)

100% CH4 0.27 (22.7) 0.41 (34) 0.54 (45.3) 0.67 (56.6)
75% CH4/25% Syngas A 0.32 (22.9) 0.51 (36) 0.65 (46.4) 0.83 (58.9)
50% CH4/50% Syngas A 0.43 (26.3) 0.61 (49.3) 0.80 (63) 1.02 (66)
75% CH4/25% Syngas B 0.28 (22.2) 0.42 (34) 0.57 (45.6) 0.69 (55.6)
50% CH4/50% Syngas B 0.36 (27.6) 0.56 (42.5) 0.74 (56.4) 0.85 (64.7)

333
P.S.P. Corrêa et al.
Fig. 8. Mean OH* chemiluminescence of flames.
that of CH4 flame, which implies that the overall heat release rate is
weaker in CH4/syngas A blends flame. A lower heat release rate could
result in an increase in unburnt gas, which is indicated by the increased
CO emission shown in Fig. 7(b). Moreover, the incomplete combustion
is also responsible for the drop in efficiency found in Fig. 5(a).
However, the global view provided by OH* chemiluminescence
cannot explain the phenomenon shown in Fig. 7 that the NOx emission
in the current study still keeps at a similar level even when the CO
emission increases. Therefore, OH-PLIF images were captured in cases
at an output power load of 20 kW to illustrate the detailed cross-section
view of the flame. Fig. 9 shows the instantaneous and mean OH-PLIF
images. By comparing the CH4 flame and CH4/syngas A flames, the
notable difference noticed is that the local OH fluorescence has a higher
intensity in cases with a higher proportion of syngas. For example, the
mean OH-PLIF image of the case with 50% CH4/50% syngas A has the
highest intensity in the region near the burner outlet. In flames, the OH
radical concentration is related to the strength of the reaction and the
flame temperature [25–28]. Therefore, the phenomenon that the CH4/
syngas A flames have a stronger local OH intensity implies that they
have a stronger local reaction than that of the CH4 flame. Similar
phenomena were detected in the work of Li et al. [29] Indicated by the
work of Tachibana et al. [30], the possible reason is that the increase in
the fuel flow rate causes insufficient mixing between fuel and air, which
leads to a high equivalence ratio (or the fuel-to-air ratio) near the
334
Applied Thermal Engineering 146 (2019) 328–337
burner outlet. As result, a stronger local reaction occurs in CH4/syngas
A flames and results in a local high temperature zone, which may re-
sponsible for the NOx emission in cases using NG/syngas A blends.
Moreover, because the CH4/syngas A flame has a small size, the re-
sident time of the fuel in the combustion zone is short, which also
causes an insufficient reaction and causes an increase in CO emission. It
can interpret the trend found with microturbine CO and NOx emission
results in Fig. 7.
As shown in Figs. 8 and 9, flames of CH4/syngas B blends are
generally larger and stronger than that of CH4/syngas A blends,
whereas they are weaker than that of CH4 flames at the same output
power. In current tests, the burner is free of flashback problem when
the proportion of syngas B is no larger than 50%, which implies that
syngas B could also be applied for microturbine tests. Moreover, based
on the OH-PLIF and OH* chemiluminescence images, it could be rea-
sonable to predict that the CO emission and NOx emission of syngas B
could be higher than that of the NG but should be lower than the cases
using syngas A. As indicated in the OH* chemiluminescence and OH-
PLIF images, the presence of hydrogen can increase the flame stability.
Therefore, the system performance should be improved when compared
with synthesis gas A. However, due to the high flame speed of hy-
drogen, there could be risks of flashback if syngas B is used in the mi-
croturbine, and therefore careful testing of the flashback problem is
needed. A useful reference of allowed hydrogen concentration is
P.S.P. Corrêa et al.
Fig. 9. Instantaneous (upper row) and mean (bottom row) OH-PLIF images.
reported in the work of Page et al. [31]. With a 60 kW Capstone mi-
croturbine, a maximum of 35% hydrogen in syngas allows stable op-
eration without flashback problems.
3.2.3. Flame response to acoustic perturbation
The turbine test with syngas A hasn’t detected any thermoacoustic
instability problems. Therefore, only cases using syngas B at an output
power of 20 kW, were examined. To compare the flame response, the
pressure oscillation must be excited at similar levels. At the beginning
of the test, we tried to perturb the flame with a wide range of frequency
(60–250 Hz). However, the acoustic damping was significantly affected
by the flame operation condition. As a result, even at the maximum
speaker input power, perturbation at most frequencies could not pro-
vide similar oscillation levels for two different cases. The case with
125.6 Hz is the only one can provide notable oscillation for two dif-
ferent cases. Therefore, it needs to emphasize that the purpose of the
acoustic perturbation test is examining whether using the syngas pro-
duced with the steam gasification can have a significant effect on the
instability trend, rather than the full acoustic response of the flame.
Therefore, the instability tests were only conducted at a fixed forcing
frequency of 125.6 Hz.
The oscillation of OH* chemiluminescence is proportional to the
heat release oscillation (q'). Therefore, the normalized heat release os-
cillation is represented by the OH* intensity oscillation,
qi qi qmean Ii Imean
= =
qmean qmean Imean (1)
where qmean is the mean heat release, I is the global OH* intensity
calculated by summing all pixel values of a single image, Imean is the
mean value of all phases, and the subscript i represents the ith phase.
Fig. 10 shows the phase-averaged OH* images and the heat release
oscillation curves. The flame shape demonstrates a significant differ-
ence when the phase angle changes, which consequently causes the
heat release to oscillate in a sinusoidal pattern. Indicated by the image,
the heat release has a higher value when the flame size is larger. By
comparing Figs. 10(a) and 10(b), it can be found that the heat release
oscillation shifts in phase when the fuel component changes. For ex-
ample, the phase-angle with the maximum value changes from 45° to
345° when the proportion of syngas B changes from 25% to 50%.
Moreover, a significant difference can also be noticed in the flame
shape, for example, the size of the flame at 105° shrinks notably when
the proportion of syngas B decreases from 25% to 50%.
The instability trend can be analyzed by the Rayleigh criterion,
335
Applied Thermal Engineering 146 (2019) 328–337
which states that the instability will be amplified if the pressure oscil-
lation and the heat release fluctuation are in phase, whereas the in-
stability will be damped if the pressure oscillation and the heat release
fluctuation are out-of-phase [32]. The Rayleigh index deduced from
Rayleigh criterion is usually to analyze the instability trend quantita-
tively [33,34]. A positive Rayleigh index represents that the instability
is excited, whereas a negative Rayleigh index stands for that the in-
stability is damped. For flames, the local Rayleigh index (R*) can be
calculated with Eq. (2),
n
1
R =
¯
Pi ·(OH )i
Pamp OH (2)
i=1
where R* is the local Rayleigh Index, Pamp is the amplitude of pressure
oscillation, OH̄ is the averaged OH* chemiluminescence, and Pi' is the
pressure oscillation at the ith phase. Pi' is obtained by extracting the
oscillation mode with a frequency of 125.6 Hz from the Fast Fourier
transform result of the pressure signal. Moreover, (OH*)i' is the local
OH* chemiluminescence oscillation at the ith phase, and n is the total
phase number. Once R* is obtained, the global Rayleigh index (Rg) can
be calculated by summing all R* on a single image. Similar to R*, a
positive Rg means that the global instability is excited, whereas a ne-
gative Rg indicates that the global instability is damped.
The computed Rayleigh index maps are shown in Fig. 11, which
demonstrates a significant change in the Rayleigh index distribution.
First of all, it can be noticed that the periodic structure of the Rayleigh
index map, which indicates that the shear layer instability triggered by
the acoustic perturbation is the main reason for the flame oscillation
[33,35]. In these two cases, the perturbations were provided at the
same frequency and similar levels. However, when the syngas B pro-
portion is 25%, the positive R* dominate the map. In contrast, the ne-
gative R* plays a dominant role once the proportion syngas B increases
to 50%. One of the reasons responsible for the difference is the flame
shape. As shown in Fig. 11, the flame with 25% syngas B is longer than
that with 50% syngas B. As a result, the downstream section (from the
height of approximately 2.5 cm to 4 cm) of case with 25% syngas B has
a larger positive R* zone than that of the case with 50% syngas B. As a
result, Rg changes from 0.0646 to −0.0369, which indicates that the
global instability shifts from being excited to being damped. The result
implies that the acoustic response of the flame can be significantly af-
fected by the change in fuel composition, which could consequently
change the acoustic oscillation mode of the combustor. Therefore, in
the future, it is necessary to examine the full acoustic response of flame
if the burner switches to high hydrogen content syngas.
P.S.P. Corrêa et al. Applied Thermal Engineering 146 (2019) 328–337

Fig. 10. The phase-averaged OH* images and heat release oscillation (q′/qmean) of (a) 75% CH4/ 25% syngas B, (b) 50% CH4/50% syngas B (Patm stands for the
atmospheric pressure).

Fig. 11. Mean OH* chemiluminescence and Rayleigh index map of cases with (left) 75% CH4/25% syngas B, (right) 50% CH4/ 50% syngas B.

336
P.S.P. Corrêa et al.
4. Conclusions
The current study focuses on the application of the biomass-derived
syngas in a microturbine. The study includes microturbine performance
and burner tests. The microturbine test, which focuses on the turbine
performance, only used the syngas from air gasification (syngas A) that
was mixed with natural gas as the fuel. In contrast, two different types
of syngas, which are syngas A and the syngas (syngas B) derived from
steam gasification method, were blended with methane and used for
burner tests.
Turbine performance tests demonstrate that the microturbine effi-
ciency decreases about 13% when the fuel changes from pure natural
gas to 50% NG/50% syngas A blends. Moreover, the temperature
analysis shows that the temperature at the turbine outlet is similar
despite the change in fuel composition. Besides, the emission results
show that when the output power is sufficiently high (> 20 kW), the
NOx emission is almost unchanged in spite of the syngas addition,
whereas the CO emission increases after syngas is added.
Burner tests used OH* chemiluminescence and OH-PLIF images to
investigate the effect of the change in fuel composition on the flame
shape. Results show that the OH* chemiluminescence of the flame is
weakened once the syngas is added, which indicates incomplete com-
bustion that causes the increase in CO emission. Moreover, OH-PLIF
images suggest that the unchanged NOx emission could relate to the
local hot zone found in flames using CH4/Syngas blends. Moreover, at
the forcing frequency of 125.6 Hz, the acoustic perturbation tests in-
dicate that instability trend is changed from being excited to being
inhibited when the proportion of syngas B changes from 25% to 50%,
which implies that the change in syngas composition could cause a
variation in the pressure oscillation mode in the practical system.
Acknowledgments
The authors are very grateful to the financial support provided by
the Brazilian National Research and Development Council (CNPq), the
Research Support Foundation of the Minas Gerais State (FAPEMIG) and
the Coordinating Body for the improvement of Post graduate Studies in
Higher Education (CAPES- Project No 8887.124485/2014-00
“Theoretical and experimental study of biomass combustion in fur-
naces”) for the funding of Research and Development (R&D) projects,
the support of graduate students and the research productivity grants
that allowed the accomplishment of the research projects whose results
are included in this paper.
References
[1] E.E.S. Lora, R.V. Andrade, J.D.M. Ángel, M.A.H. Leite, M.H. Rocha, C.A.V.B. de
Sales, M.A.G. Mendoza, D.d.S.O. Coral, Gasification and pyrolysis for biomass
conversion into electricity and biofuels, Biocombustíveis 1 (2012) 411–498.
[2] A.V. Bridgwater, The technical and economic feasibility of biomass gasification for
power generation, Fuel 74 (1995) 631–653.
[3] U. Lee, E. Balu, J.N. Chung, An experimental evaluation of an integrated biomass
gasification and power generation system for distributed power applications, Appl.
Energy 101 (2013) 699–708.
[4] M.A.R. do Nascimento, L. de Oliveira Rodrigues, E.C. dos Santos, E.E.B. Gomes,
F.L.G. Dias, E.I.G. Velásques, R.A.M. Carrillo, Micro gas turbine engine: a review,
Progress in Gas Turbine Performance, InTech, 2013.
[5] A.M. Pantaleo, S. Camporeale, N. Shah, Natural gas–biomass dual fuelled micro-
turbines: comparison of operating strategies in the Italian residential sector, Appl.
Therm. Eng. 71 (2014) 686–696.
[6] M.A. Meybodi, M. Behnia, A study on the optimum arrangement of prime movers in
small scale microturbine-based CHP systems, Appl. Therm. Eng. 48 (2012)
122–135.
337
Applied Thermal Engineering 146 (2019) 328–337
[7] F. Baina, A. Malmquist, L. Alejo, T.H. Fransson, Extended operability of a com-
mercial air-staged burner using a synthetic mixture of biomass derived gas for
application in an externally fired micro gas turbine, Fuel 150 (2015) 664–671.
[8] F. Delattin, G.D. Lorenzo, S. Rizzo, S. Bram, J.D. Ruyck, Combustion of syngas in a
pressurized microturbine-like combustor: Experimental results, Appl. Energy 87
(2010) 1441–1452.
[9] D. Król, S. Poskrobko, J. Goscik, Micro cogeneration - rich-methane gasifier and
micro gas turbine, E3S Web Conf. 2017, p. 01024.
[10] L.P.L.M. Rabou, J.M. Grift, R.E. Conradie, S. Fransen, Micro gas turbine operation
with biomass producer gas and mixtures of biomass producer gas and natural gas,
Energy Fuels 22 (2008) 1944–1948.
[11] H. Sadig, S.A. Sulaiman, M.A. Said, S. Yusup, Performance and emissions of a micro-
gas turbine fueled with LPG/producer gas in a dual fuel mode, Applied Mechanics
and Materials, Trans Tech Publ, 2015, pp. 482–486.
[12] P. Laranci, E. Bursi, F. Fantozzi, Numerical analysis of a microturbine combustion
chamber modified for biomass derived syngas, ASME 2011 Turbo Expo: Turbine
Technical Conference and Exposition, American Society of Mechanical Engineers,
2011, pp. 541–548.
[13] M.C. Lee, S.B. Seo, J.H. Chung, S.M. Kim, Y.J. Joo, D.H. Ahn, Gas turbine com-
bustion performance test of hydrogen and carbon monoxide synthetic gas, Fuel 89
(2010) 1485–1491.
[14] M.C. Lee, J. Yoon, S. Joo, J. Kim, J. Hwang, Y. Yoon, Investigation into the cause of
high multi-mode combustion instability of H2/CO/CH4 syngas in a partially pre-
mixed gas turbine model combustor, Proc. Combust. Inst. 35 (2015) 3263–3271.
[15] L. Aiguo, W. Yiwu, Effects of lower heat value fuel on the operations of micro-gas
turbine, Energy Power Eng. 1 (2009) 28.
[16] N.F. Othman, M.H. Boosroh, Effect of H2 and CO contents in syngas during com-
bustion using Micro Gas Turbine, IOP Conference Series: Earth Environ. Sci. 32
(2016) 012037.
[17] R. Andrade, Biomass Gasification: A Theoretical and Experimental Analysis, in, Vol.
Ph.D., Universidade Federal de Itajubá, UNIFEI. Thesis, 2007.
[18] D. Bohn, Micro gas turbine and fuel cell—a hybrid energy conversion system with
high potential, Micro Gas Turb. (2005) 1–46.
[19] H. Hofbauer, T. Fleck, G. Veronik, Gasification feedstock database, IEA Bioenergy,
Task XIII Thermal Gasification of Biomass, 1997.
[20] Capstone MicroTurbine Fuel Requirements http://www.globalmicroturbine.com/
pdf/fuelspecs.pdf.
[21] J. Zhang, A. Ratner, Experimental study of the global and local flame responses to
acoustic perturbation, J. Propul. Power 33 (2016) 479–489.
[22] J. Zhang, A. Ratner, Effect of pressure variation on acoustically perturbed swirling
flames, Proc. Combust. Inst. 36 (2017) 3881–3888.
[23] D.M. Kang, F.E.C. Culick, A. Ratner, Combustion dynamics of a low-swirl com-
bustor, Combust. Flame 151 (2007) 412–425.
[24] A.H. Lefebvre, Gas Turbine Combustion, CRC Press, 1998.
[25] J. Heinze, U. Meier, T. Behrendt, C. Willert, K.-P. Geigle, O. Lammel, R. Lückerath,
PLIF thermometry based on measurements of absolute concentrations of the OH
radical, Z. Phys. Chem. 225 (2011) 1315–1341.
[26] R. Marsh, J. Runyon, A. Giles, S. Morris, D. Pugh, A. Valera-Medina, P. Bowen,
Premixed methane oxycombustion in nitrogen and carbon dioxide atmospheres:
measurement of operating limits, flame location and emissions, Proc. Combust.
Instit. 36 (2017) 3949–3958.
[27] B. Ge, S. Zang, P. Guo, Y. Tian, Experimental investigation of double-swirled non-
premixed syngas flames by planar laser-induced fluorescence, Energy Fuels 26
(2012) 1585–1591.
[28] M. Huang, Y. Xiao, Z. Zhang, W. Shao, Y. Xiong, Y. Liu, Z. Liu, F. Lei, Effect of air/
fuel nozzle arrangement on the MILD combustion of syngas, Appl. Therm. Eng. 87
(2015) 200–208.
[29] L. Yang, Z. Wang, Y. Zhu, Z. Li, J. Zhou, Z. Huang, K. Cen, Premixed jet flame
characteristics of syngas using OH planar laser induced fluorescence, Chin. Sci. Bull.
56 (2011) 2862.
[30] S. Tachibana, K. Kanai, S. Yoshida, K. Suzuki, T. Sato, Combined effect of spatial
and temporal variations of equivalence ratio on combustion instability in a low-
swirl combustor, Proc. Combust. Inst. 35 (2015) 3299–3308.
[31] D. Page, B. Shaffer, V. McDonell, Establishing Operating Limits in a Commercial
Lean Premixed Combustor Operating on Synthesis Gas Pertaining to Flashback and
Blowout, 2012, pp. 647–656.
[32] J. Rayleigh, The explanation of certain acoustical phenomena, Nature 18 (1878)
319–321.
[33] Y. Huang, A. Ratner, Experimental investigation of thermoacoustic coupling for
low-swirl lean premixed flames, J. Propul. Power 25 (2009) 365.
[34] P. Sun, Y. Yuan, B. Ge, Y. Tian, Z. Zhang, S. Zang, Combustion oscillation char-
acteristics and flame structures in a lean premixed prevaporized combustor, Energy
Fuels 31 (2017) 10060–10067.
[35] D.W. Davis, P.L. Therkelsen, D. Littlejohn, R.K. Cheng, Effects of hydrogen on the
thermo-acoustics coupling mechanisms of low-swirl injector flames in a model gas
turbine combustor, Proc. Combust. Inst. 34 (2013) 3135–3143.