- Email: [email protected]
Active fuel particles dispersion by synthetic jet in an entrained flow gasifier of biomass: Cold flow
Powder Technology 302 (2016) 275–282
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Active fuel particles dispersion by synthetic jet in an entrained flow gasifier of biomass: Cold flow Ammar Hazim Saber a,b,⁎, Burak Göktepe c,d, Kentaro Umeki c, T. Staffan Lundström a, Rikard Gebart c a
Division of Fluid and Experimental Mechanics, Luleå University of Technology, 971 87 Luleå, Sweden University of Mosul, Mechanical Engineering Department, Mosul, Iraq c Energy Engineering, Division of Energy Science, Luleå University of Technology, 971 87 Luleå, Sweden d SP Energy Technology Centre, Industrigatan 1, 941 28 Piteå, Sweden b
a r t i c l e
i n f o
Article history: Received 3 March 2016 Received in revised form 9 August 2016 Accepted 30 August 2016 Available online 31 August 2016 Keywords: Non-spherical particles Biomass Flow control Synthetic jet Entrained flow gasifier
a b s t r a c t Pulverized fuel (PF) burners play a key role for the performance of PF fired gasification and combustion plants, by minimizing pollutant emission, fuel consumption and hence fuel costs. However, fuel diversity in power generation plants imposes limitations on the performance of existing PF burners, especially when burning solid fuel particles with poor flowability like biomass sawdust. In the present study, a vertically downward laminar flow was laden with biomass particles at different particle mass loading ratios, ranging from 0.47 to 2.67. The particle laden flow was forced by a synthetic jet actuator over a range of forcing amplitudes, 0.35–1.1 kPa. Pulverized pine particles with a sieve size range of 63–112 μm were used as biomass feedstock. Two-phase particle image velocimetry was applied to measure the velocity of the particles and air flow at the same time. The results showed that the synthetic jet had a large influence on the flow fields of both air and powdered pine particles, via a convective effect induced by vortex rings that propagate in the flow direction. The particle velocity, particle dispersion and hence inter-particle distance increased with increasing forcing amplitude. Moreover, particles accumulated within a specific region of the flow, based on their size. The effect on particle dispersion was more pronounced in the forced flows with low mass loading ratios. © 2016 Elsevier B.V. All rights reserved.
1. Introduction As global energy demand continues to increase, the use of renewable energy sources can serve as an alternative solution to reduce energy demand and to improve energy security with reduced greenhouse gas emissions as a benefit. Recently, there has been an increasing effort to integrate biomass into the major CO2 emitting sectors, e.g. energy and transportation since biomass is renewable and CO2 neutral. Gasification is one of the cleanest and most versatile ways to convert biomass into high value-added commercial products, e.g. biofuel or commodity chemical products. Entrained flow gasification (EFG) is a well proven, commercially available gasification technology for coal, producing high quality syngas with little or no tar. Although there has been recently promising progress towards entrained flow gasification of biomass, there still remains some challenges that come from biomass properties. Biomass particles differ from coal particles in many aspects such as particle size, shape, texture and chemical composition. Retrofitting existing entrained flow coal gasifiers to accommodate biomass can produce some technical hurdles. For instance, bulky and fibrous woody biomass ⁎ Corresponding author at: Division of Fluid and Experimental Mechanics, Luleå University of Technology, 971 87 Luleå, Sweden. E-mail addresses: [email protected], [email protected] (A.H. Saber).
http://dx.doi.org/10.1016/j.powtec.2016.08.071 0032-5910/© 2016 Elsevier B.V. All rights reserved.
leads to inconsistent feeding [1,2]. This is because fibrous biomass often has low bulk density and high cohesive forces [3]. Entrained flow gasifiers are operated at low air/fuel equivalence ratios (≈0.4) in order to convert energy into chemical energy of combustible gases instead of heat [4]. At low air/fuel ratios, particles are packed tightly together, occupying high volume fractions. Quantitatively, mass loading ratio (biomass feeding rate to carrier gas flow rate) can be ca. 2 for oxygen blow entrained flow gasifier compared with ca. 0.2 for air-blown burners. The review study by Sirignano [5] reported that close-packed particle arrays produce different reaction environment in terms of temperature and gas composition than that for isolated particles. Similar findings were obtained in an experimental study of Göktepe et al. [6] showing that highly packed particles produce higher volume fractions of soot than loosely packed particles or isolated particles. It is therefore important to investigate means to control the particle dispersion. A conventional way to control particle dispersion in powder burners is to impose a swirling flow that will enhance particle dispersion via centrifugal effects and prolong the residence time [7]. An alternative way is to control the flow with a synthetic jet actuator that generates a pulsed flow by periodic forcing a fluid back and forth through a small opening. In this way, momentum can be transferred to the flow with zero-net mass flux while simultaneously enhancing the mixing
276
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
Fig. 1. (a) Particle size distribution of pine sawdust with a sieving size of 63–112 μm (b) width-to-length ratio (b/l) (c) particle sphericity (d) snapshot of pulverized pine particles recorded by CAMSIZER XT.
of gases via a convective effect induced by trailing vortex rings [8]. It has been shown by many researchers that the flow fluctuations are more pronounced in low flow velocity cases, and this effect will decay and vanish downstream the excitation point. Moreover, several studies on particle-laden flows have reported that the vortex rings result in the gathering of solid particles in two different flow regions: a high shear
Biomass feeder
Mixing box
region, between vortex rings, and a high vorticity region, the vortex rings themselves. This phenomena is referred to as “preferential accumulation” and has been reported by Balanchandar and Eaton [9] and Tamburello and Amitay [10]. To the best of the authors' knowledge, there is no available literary information on the motion and dispersion behaviour of non-spherical
Seeding generator
Loudspeaker Fused silica windows (transmission: >99%) Nd:YAG laser Flat flame burner
Mirror Particle flow
(a)
(a) Flow meters
Quartz tube
Exhaust
CCD camera
Char bin
Air
Air
Gas cylinders
Fig. 2. Schematic diagram of experimental set-up.
section a-a
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
particles in vortex flows since the previous studies in the relevant area have been conducted with spherical particles. This paper is therefore focused on investigating the interactions between fine ground, nonspherical biomass particles and vortex rings for different particle mass loading ratios and for different strength of pulsating flows. 2. Material and methods
277
is defined as the caliper diameter measured over a geometrical shape. For particles smaller than 250 μm, the b/l ratio decreases with increasing particle size, see Fig. 1b. Sphericity is derived from the ratio of measured area of the particle projection to the measured circumference of the particle projection. The sphericity decreases from 0.8 to 0.3 when increasing particle sizes from 20 μm to 280 μm. A snapshot image of pulverized pine particles in Fig. 1d confirms the shape irregularities and non-sphericity for biomass particles.
2.1. Particle characterization 2.2. Experimental setup and procedure Pine sawdust with a sieving size of 63–112 μm was used in this study. The characteristic size and the shape properties of the biomass sample were measured with a dynamic image analysis method (CAMSIZER XT, Retsch Technology GmbH) and are presented in Fig. 1. The particle size distribution of the biomass sample is given in terms of cumulative distribution, Q, see Fig. 1a. Here, particle diameter, i.e. the shortest chord among the maximum chords in all directions, is defined as the particle size since it is more relevant to the sieve size. Sphericity (SPHT) and width to length ratio (b/l) for different particle diameter are reported in Fig. 1b and c, respectively. The b/l ratio is calculated as the ratio of the shortest maximum chord of a particle projection to the maximum Feret diameter of a particle projection. The Feret diameter
The experiments were carried out with a laminar cold-flow jet in a laboratory scale, atmospheric entrained flow reactor. The experimental set-up is illustrated in Fig. 2. The experimental-setup mainly consists of an air supply unit, a wood powder feeder, a synthetic jet actuator unit, a flat flame burner (FFB), a quartz reactor tube, a char bin, an air seeding unit, and an air exhaust unit. Pine sawdust was fed into the flow with a syringe pump feeder and was entrained into a central tube of a commercial burner (Mckenna Flat Flame Burner by Holthuis & Associates) with a carrier air flow (0.43 l min− 1). Particle feeding rates used in this study were 14.5 g h−1, 25.5 g h−1, 47.3 g h−1 and 91.0 g h−1 that corresponded
Fig. 3. The mean binary particle images obtained at mass loading ratio of 0.47 by averaging over 200 sample images at a distance of 2–30 mm from the burner outlet for (a) base case (b) 35 Hz (c) 289 Hz (d) 451 Hz. Note that the images are false-color.
278
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
to the mass loading ratios of biomass to air as 0.47, 0.75, 1.39 and 2.67. The loading ratios were calculated with respect to the flow rate of the carrier air. The burner has two co-axial water-cooled porous discs that surround the central tube. In normal cases, a premixed mixture of methane and air (methane: 1.92 l min−1, air: 14.6 l min−1) is ignited and stabilized on the innermost stainless steel porous disc and Nitrogen (8.49 l min−1) is supplied from the outermost bronze porous disc to shield the flame. In this study, methane was replaced with air since the scope of the study encompassed visualization of the particle-flow interactions under cold flow conditions. More detailed information regarding the powder feeder and the flat flame burner can be obtained from the study of Göktepe et al. [6]. The synthetic jet actuator consists of a 4-ohm loudspeaker, (ND1404-5-1/4″, Dayton), which is fitted into an annular cavity with a central orifice of 4 mm. It produces a pulsed flow ejected into and out from the small orifice during periodic motions of the loudspeaker's diaphragm. The synthetic jet actuator was mounted perpendicular to the central tube of the burner via a Tee union adapter and an extension pipe. A fine porous steel mesh was attached to the entrance of the arm of the Tee-union adapter to avoid the accumulation of fine pine sawdust particles in the cavity and along an inner surface of the
extension pipe during the suction part of the actuation cycle. The loudspeaker was driven by a sinusoidal signal. The signal was generated by a data acquisition board (DT-9841-VIB, Data Translation), and was amplified by a stereo amplifier (Integrated amplifier A-10 from Pioneer) prior to being sent to the loudspeaker. A high sensitivity dynamic pressure sensor (106B, PCB) was flush mounted on the inner wall of the cavity. This sensor measures dynamic pressure fluctuations (in terms of Prms) developed inside the cavity. The sine-wave of 35 Hz with the peak to peak amplitudes of 0.15 V, 0.29 V and 0.48 V resulted in pressure fluctuations of 0.346 kPa (144.7 dB), 0.668 kPa (150.8 dB) and 1.106 kPa (154.8 dB), respectively. Accelerometer sensors (Delta Tron 4507, Bruel&Kjaer) were also attached on the external surface of the central pipeline to monitor the structural resonant frequencies. Particle and air velocities were measured at the same time with particle image velocimetry (PIV), a system from LaVision GmbH. The PIV system consists of double pulsed Nd-YAG laser with a maximum repetition rate of 100 Hz, a Lavision FlowMaster Imager Pro CCD-camera with a spatial resolution of 1280 × 1024 pixels per frame, timing controller and the PIV software program (LaVision Davis 7.2). The carrier air flow was seeded with aerosol droplets with diameters of 0.3 μm that were produced by atomizing Di-Ethyl-Hexyl-Sebacat (DEHS) with an
Fig. 4. Snapshot images of biomass particles at a distance range of 2–30 mm from the burner outlet in (a) the unforced laminar flow with the mass loading ratio of 0.47 (b) the forced initially laminar flow with the mass loading ratio of 0.47 (c) the unforced laminar flow with the mass loading ratio of 2.67 (d) the forced initially laminar flow with the mass loading ratio of 2.67. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa.
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
aerosol generator. The PIV measurements were conducted at a frequency of 34.65 Hz, with a total of 700 image pairs for each recorded set (sampled during around 20 s). Each set was repeated four times to check the repeatability of the measurements. The sampling frequency was adjusted to maintain in-phase coupling with time scales of the flow structures. A dynamic masking and image processing algorithm, which was previously used by Saber et al. [12], was employed to discriminate the biomass particles from the tracer aerosol particles. Based on the algorithm, the first step was to reverse black and white pixel colors in the raw binary images. In the next step, random noise, in the images, was removed by a 3 × 3 pixels smoothing filter. The biomass particles were detected by using an edge detection algorithm which detects the changes in contrast and the interior regions of the particles were filled by a dilute filter followed by an erosion filter. Biomass particles were sorted out from the images by using a threshold filter. The images were divided into interrogation windows of 32 × 32 pixels with a 50% overlap. A minimum of five tracer particles were detected in interrogation area which gave a reliable cross-correlation. The biomass particle velocities were measured by a particle tracking algorithm (PTV) which relies on tracking single particles in the flow domain. A constant temperature hot-wire probe (Type 55-P11 from Dantec Dynamics) was used to measure the air velocities in the unforced laminar flow. The probe has a 5 μm diameter platinum-plated tungsten wire, which is welded directly to the prongs. The probe measures one component of gas velocity and is recommended to be used for measurements in one directional flow of low intensity. Under the aforementioned flow
279
conditions, the carrier air velocity ejected from the central tube was measured to be around 0.37 m s−1, which corresponds to a Reynolds number (Re) of 140, while the air supplied from the porous disc had a velocity of about 0.1 m s−1. 3. Results and discussion The frequency used in this study was selected in a screening experiment by operating the synthetic jet actuator at a range of frequencies, starting from 10 Hz to 600 Hz with a step resolution of 10 Hz. The biomass particle-air flow with the mass loading ratio of 0.47 was continuously illuminated along the reactor height by the laser sheet and were visually observed by the high speed CCD camera in order to determine the frequencies that lead to high particle dispersion. At the same time, the pressure signals were measured in the quartz reactor and in the cavity. However, the mounting assembly, attached on one of the reactor ports, was located a distance away from the burner outlet and the dynamic pressure fluctuations were already damped out at this position. A custom designed probe with a tip of 100 mm was therefore used to sense the pressure fluctuations in the burner near field. Since this probe has a linear frequency response between 0 and 100 Hz, the pressure measurements in the reactor were performed over this frequency spectrum. Based on visual observations it can be stated that high particle dispersion was achieved at low frequencies in the range 15 to 50 Hz. The pressure probe measurements showed that the strongest flow fluctuations were at 35 Hz, which corresponded to St = 0.57 (Strouhal
(a)
u/U j
−1 0 1 2 3
y/D j =1
−1 0 1 2 3
y/D j =2
−1 0 1 2 3 −1.5
y/D j =3 −1
−0.5
0
0.5
1
1.5
r/D j
(b) 4
1.106 kPa 0.346 kPa 0.07 kPa
3.5 3
u/U j
2.5 2 1.5 1 0.5 0
10
15
20
25
30
35
y/D j
Fig. 5. Phase-averaged (a) vorticity field (b) total in-plane velocity field of air flow for Prms = 1.106 kPa.
Fig. 6. (a) Dimensionless mean axial air velocity component vs. crosswise distance at three positions from the burner outlet. (b) Dimensionless mean centerline air velocity profile vs. streamwise distance. The forcing frequency was 35 Hz.
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
Fig. 7 presents the effect of mass loading ratio, between 0.47 and 2.67, on the mean axial air velocity and the mean centerline air velocity at Prms = 1.106 kPa. The velocity magnitude was lower for the case with high mass loading ratio since the particles or particle aggregates tend to act like a semi solid obstacle against the air flow. In Fig. 7b, the mean centerline air velocity profile shows that the peak to peak distance and hence the distance between two vortices decreased with increasing the loading ratio. This can be interpreted as if the vortex rings slowed down, or were blocked, at high mass loading ratio. The phase-averaged streamwise velocity for each particle, normalized by jet velocity, u/Uj, for three sizes ranges of particles was calculated for 10 samples and is presented in Fig. 8a–c. Please note that the size of the circles gives a qualitative indication of the particle size while the color indicates the normalized particle velocity. As expected, particles passing through the vortex ring have higher stream-wise velocity while the particles located around the vortex ring have low velocity. It is also shown that for the larger particles the concentration is higher at the jet core and lower in the periphery of the image. The phaseaveraged lateral velocity of the particles were derived in a similar manner and plotted in Fig. 8d–f. The small particles are convected away from the jet core due to the rotating vortex rings for which the rotational direction is directed downwards near the axis and upwards at the outer periphery of the vortex rings. Some of the particles are entrained into the vortex ring edge and re-enter the jet core. However, in the core of the reverse flow zone, no particles were detected due to centrifugal
(a)
u/U j
number: St = fd/u, where f is forcing frequency, d is the nozzle diameter of biomass-air flow, and u is axial air velocity). This agrees well with reports in the literature that states that the maximum mixing enhancement for jet flow is in the range of St = 0.3–0.6 [11]. The resonance frequency for the cavity and the central pipeline were measured to be 289 Hz and 451 Hz, respectively. Fig. 3 shows the mean binary images of biomass particles for these three characteristic frequencies (35, 289 and 451 Hz). The particle in each image was detected by using the algorithm used in our previous study [6]. Then, the mean value for each pixel was calculated from 200 binary images, where 1 means the presence of particles. The particles are more likely to exist in the position with lighter color. It shows that the highest particle dispersion was achieved when the synthetic jet actuator was operated at 35 Hz. The resonant vibrations at 451 Hz in the central tube showed some effects on the particle dispersion albeit the effect was not as strong as that at 35 Hz. Compared to 35 Hz, the resonant frequency of the cavity at 289 Hz caused even higher dynamic pressure fluctuations inside the reactor, but did not enhance dispersion of the particles significantly. Fig. 4 shows the features of the unforced and forced particle-air flows at a distance range of 2–30 mm from the burner outlet. The biomass particles in the undisturbed laminar flows agglomerated to form larger particles and the agglomeration was more pronounced for high mass loading ratios as shown in Fig. 4a and c. This may be linked to cohesive forces or electrostatic forces among the biomass particles. When the air flow was pulsed at 35 Hz, vortex rings were formed by vortex shedding at the burner rim and then propagated downwards in the axial direction. The convective effect induced by the vortex rings spread the particles laterally over a large distance in the reactor as shown in Fig. 4b and d. Moreover, the vortex flows caused very little or no particle agglomeration. Fig. 5 shows the phase-averaged vorticity and total air velocity in the biomass-air flow pulsating at the forcing amplitude, Prms = 1.106 kPa. A visual inspection of the figures yields that the vortex ring divided the flow into two regions: a reversed flow zone (around the vortex rings) and a high speed zone (inside the vortex rings). The counter rotating flow zones drag a part of the surrounding air flow into the vortex centerline, leading to an increase of the total velocity in this region. The maximum velocity measured in the vortex centerline is about 2 m s−1 and drops to 0.2 m s− 1 around the boundary of the reversed flow zone, leading to high velocity gradients and hence shear stresses. It has previously been shown that high shear rates (velocity gradient larger than 30 s−1) break up particle aggregates, causing a reduction in the mean particle size [13]. Returning to Fig. 4, the high local shear stresses may thus be the reason for the break-up of particle aggregates in the burner near field. Alternatively, the break-up of particle aggregates might have already taken place in the central tube. Fig. 6 presents the effect of the forcing amplitude on the mean axial air velocity and the mean centerline air velocity. The mean axial air and the mean centerline air velocities were non-dimensionalized with the centerline velocity at the burner outlet obtained from the undisturbed laminar jet velocity, Uj (0.37 m s−1). The lateral and axial directions were normalized with the central tube diameter, Dj. Fig. 6a shows that the mean axial air velocity increased with the excitation amplitude. The flow is highly asymmetric for Prms = 1.106 kPa at y/Dj = 1. The asymmetry in the flow originates from the asymmetric position of the synthetic jet actuator and the main jet flow is deflected away from the synthetic jet side. Similar results have been reported previously [10]. Moreover, the negative local mean velocities at x/Dj = ± 1 reveals that the flow is reversed towards the burner rim under the effect of suction. In Fig. 6b, a peak to peak distance can be defined as the distance between the first maximum value and the second maximum value of the dimensionless velocity and can be interpreted as a distance between two vortex rings (cf. with Fig. 4b). From the figure, it can be concluded that the peak-to-peak distance increases with increasing the forcing amplitude. This can be interpreted as if the vortex rings will propagate faster when the forcing amplitude is increased.
−1 0 1 2 3 −1 0 1 2 3 −1 0 1 2 3 −1.5
y/D j =1
y/D j =2
y/D j =3 −1
−0.5
0
0.5
1
1.5
r/D j
(b) 4
Mass loading ratio 0.47 2.67
3.5 3 2.5
u/U j
280
2
1.5 1 0.5 0
10
15
20
25
30
35
y/D j Fig. 7. (a) Axial component of the averaged air velocity profiles vs. crosswise distance and (b) mean centerline velocity profile vs. streamwise distance measured in the flow pulsating at Prms = 1.106 kPa. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa.
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
25
(a)
(b)
50-83 µm
281
1.4
(c)
83-117 µm
117-150 µm
1.2
20
y mm
1 15
0.8 0.6
10
0.4 5 0
0.2 0
5
10
15
20 0
5
10
x mm
15
20
0
5
10
x mm
15
20
x mm
0
u/Uj 0.3
25
(d)
(e)
50-83 µm
(f)
83-117 µm
117-150 µm 0.2
20
y mm
0.1 15 0 10 −0.1 5 0
−0.2
0
5
10
15
20 0
5
10
x mm
15
20
0
5
10
x mm
15
20
x mm
−0.3
v/Uj
Fig. 8. (a–c) Phase-averaged particle images obtained from 10 sample velocity fields, represented with particle streamwise velocity u/Uj and the particle size. (d–f) phase-averaged particle images represented with the particle radial velocity and the particle size. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa. Note that the size of the particles is qualitatively represented by the size of the circles.
25
(a)
17
12.5
(b)
(c) 20
16
11.5
y mm
15 15
10.5 10
14
9.5 5
13 8.5
9.5
10.5
8.5 14.5
11.5 0
(d)
180º 210º
5
10
x mm
150º
240º
0
15
20
(e)
15.5
16.5 180º
210º
150º
240º
1500
500 400
1000 200
500
270º
90º
300º
60º 30º
330º 0º
17.5
α
270º
90º
300º
60º 30º
330º 0º
Fig. 9. (a) phase-averaged particle image and schematic representation of orientation angle of a single particle with respect to flow direction, zoom-in sections of (b) main jet flow (c) of vortex flow, particles alignment angle with respect to flow direction in region of (d) main jet flow, (e) the vortex region. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa.
282
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
forces. The lateral particle velocity increased laterally towards the reverse flow zones. Fig. 9 shows the orientation of the biomass particles and particle velocity vectors calculated from the phase-averaged particle image for Prms = 1.106 kPa and for the mass loading ratio of 0.47. The particle orientation with respect to the stream-wise (or axial) direction is defined by an angle, α in Fig. 9a, and is illustrated in polar coordinates in Fig. 9d–e. For a given particle size range, the majority of the biomass particles in the high shear region shown in Fig. 9b was oriented over a range of angles from 330° to 30° to the flow direction, (see, Fig. 9d) while a larger portion of smaller particles around the vortex ring shown in Fig. 9c was aligned over a range of angles from 30° to 90° with the streamwise direction (see Fig. 9e). This shows that the biomass particles for a given size range had a tendency to align themselves along the streamlines of the flow. The average inter-particle distance is shown against the forcing amplitude in Fig. 10. The data of unforced flow (Prms = 0 kPa) was removed since the particle size distribution analyses showed the sign of agglomeration and does not represent the isolated effect of particle dispersion. Except for the flows with the highest mass loading ratio of 2.67, the inter-particle distance increased with increasing forcing amplitude. The effect was more pronounced for the flows with low mass loading ratios. For instance for a mass loading ratio of 0.47, the trend was quasi linear. The inter-particle distance decreased with increasing the mass loading ratio.
Experiments were performed with pine saw dust particles in a laboratory scale entrained flow reactor. The biomass particle flows as well as the carrier air flow were excited by a synthetic jet actuator to investigate the effect on the flow and the inter-particle distance. Velocity measurements were conducted with a PIV system, both for the continuous phase and for the particle phase. The flow was excited with three different excitation amplitudes and the effect of four different mass loading ratios were investigated. The results can be summarized as follows.
Inter−particle distance /mean particle diameter
• The synthetic jet actuator produced a continuous train of vortex rings when the forcing frequency was in the range 10 to 50 Hz. However, the strongest vortex rings and the highest particle dispersion were achieved at 35 Hz. • The biomass particles tend to agglomerate in the unforced laminar flows, but the particle aggregates disappeared in the forced flows, even at high loading ratios (2.67). 15
Mass loading ratio 0.47 1.39
0.75 2.67
10
5
0
Acknowledgment The authors would like to thank financial support by Norbottens Research Council (14-206), Kempe Foundation (SMK-1232), and Swedish Energy Agency for supporting this work.
4. Conclusions
0
• The strong vortex rings generated two different zones: one high momentum zone (between the vortex rings) and one high vorticity zone (the vortex rings themselves). The larger particles with high inertia were mainly concentrated between the vortex rings while the smaller particles were convected radially away from the jet core by the rotational motion of the vortex rings. • The particle loading ratios affected the flow features of the trailing vortex rings. For instance, the distance between the velocity peaks, interpreted as the distance between the centers of two vortex rings, increased with increasing actuator power and decreased with increasing particle mass loading ratio. • In dense particle-air flows, the presence of particles reduced the centerline air velocity while the centerline velocity increased with increasing actuator power. • The particle alignment with respect to the streamwise direction was different for the high momentum (between vortex rings) and high vorticity (inside the vortex rings) flow regions. The particles in the high momentum zone were aligned with the streamwise direction within a range of angles ±30° while the particles at the edge of the vortex rings were aligned within a range of angles between ± 30– 90°, i.e. close to the local direction of gas flow.
0.2
0.4
0.6
0.8
1
1.2
prms, kPa Fig. 10. Inter-particle distance calculated for different loading ratios, as function of Prms. The forcing frequency was 35 Hz.
References [1] M. Asadullah, Barriers of commercial power generation using biomass gasification gas: a review, Renew. Sust. Energ. Rev. 29 (2014) 201–215, http://dx.doi.org/10. 1016/j.rser.2013.08.074. [2] A. Joppich, H. Salman, Wood powder feeding, difficulties and solutions, Biomass Bioenergy 16 (1999) 191–198, http://dx.doi.org/10.1016/S0961-9534(98)00082-8. [3] K. Svoboda, M. Pohořelý, M. Hartman, J. Martinec, Pretreatment and feeding of biomass for pressurized entrained flow gasification, Fuel Process. Technol. 90 (2009) 629–635, http://dx.doi.org/10.1016/j.fuproc.2008.12.005. [4] F. Weiland, H. Hedman, M. Marklund, H. Wiinikka, O. Öhrman, R. Gebart, Pressurized oxygen blown entrained-flow gasification of wood powder, Energy Fuel 27 (2013) 932–941, http://dx.doi.org/10.1021/ef301803s. [5] W.A. Sirignano, Advances in droplet array combustion theory and modeling, Prog. Energy Combust. Sci. 42 (2014) 54–86, http://dx.doi.org/10.1016/j.pecs.2014.01. 002. [6] B. Göktepe, K. Umeki, R. Gebart, Does distance among biomass particles affect soot formation in an entrained flow gasification process? Fuel Process. Technol. 141 (2016) 99–105, http://dx.doi.org/10.1016/j.fuproc.2015.06.038. [7] R.B. Wicker, J.K. Eaton, Structure of a swirling, recirculating coaxial free jet and its effect on particle motion, Int. J. Multiphase Flow 27 (2001) 949–970, http://dx. doi.org/10.1016/S0301-9322(00)00061-6. [8] L.M. Cerecedo, L. Aísa, J.A. García, J.L. Santolaya, Changes in a coflowing jet structure caused by acoustic forcing, Exp. Fluids 36 (2004) 867–878, http://dx.doi.org/10. 1007/s00348-003-0769-8. [9] S. Balachandar, J.K. Eaton, Turbulent dispersed multiphase flow, Annu. Rev. Fluid Mech. 42 (2010) 111–133, http://dx.doi.org/10.1146/annurev.fluid.010908. 165243. [10] D.A. Tamburello, M. Amitay, Active manipulation of a particle-laden jet, Int. J. Multiphase Flow 34 (2008) 829–851, http://dx.doi.org/10.1016/j.ijmultiphaseflow.2008. 02.006. [11] E. Mastorakos, M. Shibasaki, K. Hishida, Mixing enhancement in axisymmetric turbulent isothermal and buoyant jets, Exp. Fluids 20 (1996) 279–290, http://dx.doi. org/10.1007/BF00192673. [12] A. Saber, S. Lundström, G. Hellström, Influence of inertial particles on turbulence characteristics in the near and Outer Wall flow as revealed with high resolution PIV, J. Fluids Eng. 138 (2016) 091303, http://dx.doi.org/10.1115/1.4033369. [13] T. Serra, J. Colomer, B.E. Logan, Efficiency of different shear devices on flocculation, Water Res. 42 (2008) 1113–1121, http://dx.doi.org/10.1016/j.watres.2007.08.027.
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Active fuel particles dispersion by synthetic jet in an entrained flow gasifier of biomass: Cold flow Ammar Hazim Saber a,b,⁎, Burak Göktepe c,d, Kentaro Umeki c, T. Staffan Lundström a, Rikard Gebart c a
Division of Fluid and Experimental Mechanics, Luleå University of Technology, 971 87 Luleå, Sweden University of Mosul, Mechanical Engineering Department, Mosul, Iraq c Energy Engineering, Division of Energy Science, Luleå University of Technology, 971 87 Luleå, Sweden d SP Energy Technology Centre, Industrigatan 1, 941 28 Piteå, Sweden b
a r t i c l e
i n f o
Article history: Received 3 March 2016 Received in revised form 9 August 2016 Accepted 30 August 2016 Available online 31 August 2016 Keywords: Non-spherical particles Biomass Flow control Synthetic jet Entrained flow gasifier
a b s t r a c t Pulverized fuel (PF) burners play a key role for the performance of PF fired gasification and combustion plants, by minimizing pollutant emission, fuel consumption and hence fuel costs. However, fuel diversity in power generation plants imposes limitations on the performance of existing PF burners, especially when burning solid fuel particles with poor flowability like biomass sawdust. In the present study, a vertically downward laminar flow was laden with biomass particles at different particle mass loading ratios, ranging from 0.47 to 2.67. The particle laden flow was forced by a synthetic jet actuator over a range of forcing amplitudes, 0.35–1.1 kPa. Pulverized pine particles with a sieve size range of 63–112 μm were used as biomass feedstock. Two-phase particle image velocimetry was applied to measure the velocity of the particles and air flow at the same time. The results showed that the synthetic jet had a large influence on the flow fields of both air and powdered pine particles, via a convective effect induced by vortex rings that propagate in the flow direction. The particle velocity, particle dispersion and hence inter-particle distance increased with increasing forcing amplitude. Moreover, particles accumulated within a specific region of the flow, based on their size. The effect on particle dispersion was more pronounced in the forced flows with low mass loading ratios. © 2016 Elsevier B.V. All rights reserved.
1. Introduction As global energy demand continues to increase, the use of renewable energy sources can serve as an alternative solution to reduce energy demand and to improve energy security with reduced greenhouse gas emissions as a benefit. Recently, there has been an increasing effort to integrate biomass into the major CO2 emitting sectors, e.g. energy and transportation since biomass is renewable and CO2 neutral. Gasification is one of the cleanest and most versatile ways to convert biomass into high value-added commercial products, e.g. biofuel or commodity chemical products. Entrained flow gasification (EFG) is a well proven, commercially available gasification technology for coal, producing high quality syngas with little or no tar. Although there has been recently promising progress towards entrained flow gasification of biomass, there still remains some challenges that come from biomass properties. Biomass particles differ from coal particles in many aspects such as particle size, shape, texture and chemical composition. Retrofitting existing entrained flow coal gasifiers to accommodate biomass can produce some technical hurdles. For instance, bulky and fibrous woody biomass ⁎ Corresponding author at: Division of Fluid and Experimental Mechanics, Luleå University of Technology, 971 87 Luleå, Sweden. E-mail addresses: [email protected], [email protected] (A.H. Saber).
http://dx.doi.org/10.1016/j.powtec.2016.08.071 0032-5910/© 2016 Elsevier B.V. All rights reserved.
leads to inconsistent feeding [1,2]. This is because fibrous biomass often has low bulk density and high cohesive forces [3]. Entrained flow gasifiers are operated at low air/fuel equivalence ratios (≈0.4) in order to convert energy into chemical energy of combustible gases instead of heat [4]. At low air/fuel ratios, particles are packed tightly together, occupying high volume fractions. Quantitatively, mass loading ratio (biomass feeding rate to carrier gas flow rate) can be ca. 2 for oxygen blow entrained flow gasifier compared with ca. 0.2 for air-blown burners. The review study by Sirignano [5] reported that close-packed particle arrays produce different reaction environment in terms of temperature and gas composition than that for isolated particles. Similar findings were obtained in an experimental study of Göktepe et al. [6] showing that highly packed particles produce higher volume fractions of soot than loosely packed particles or isolated particles. It is therefore important to investigate means to control the particle dispersion. A conventional way to control particle dispersion in powder burners is to impose a swirling flow that will enhance particle dispersion via centrifugal effects and prolong the residence time [7]. An alternative way is to control the flow with a synthetic jet actuator that generates a pulsed flow by periodic forcing a fluid back and forth through a small opening. In this way, momentum can be transferred to the flow with zero-net mass flux while simultaneously enhancing the mixing
276
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
Fig. 1. (a) Particle size distribution of pine sawdust with a sieving size of 63–112 μm (b) width-to-length ratio (b/l) (c) particle sphericity (d) snapshot of pulverized pine particles recorded by CAMSIZER XT.
of gases via a convective effect induced by trailing vortex rings [8]. It has been shown by many researchers that the flow fluctuations are more pronounced in low flow velocity cases, and this effect will decay and vanish downstream the excitation point. Moreover, several studies on particle-laden flows have reported that the vortex rings result in the gathering of solid particles in two different flow regions: a high shear
Biomass feeder
Mixing box
region, between vortex rings, and a high vorticity region, the vortex rings themselves. This phenomena is referred to as “preferential accumulation” and has been reported by Balanchandar and Eaton [9] and Tamburello and Amitay [10]. To the best of the authors' knowledge, there is no available literary information on the motion and dispersion behaviour of non-spherical
Seeding generator
Loudspeaker Fused silica windows (transmission: >99%) Nd:YAG laser Flat flame burner
Mirror Particle flow
(a)
(a) Flow meters
Quartz tube
Exhaust
CCD camera
Char bin
Air
Air
Gas cylinders
Fig. 2. Schematic diagram of experimental set-up.
section a-a
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
particles in vortex flows since the previous studies in the relevant area have been conducted with spherical particles. This paper is therefore focused on investigating the interactions between fine ground, nonspherical biomass particles and vortex rings for different particle mass loading ratios and for different strength of pulsating flows. 2. Material and methods
277
is defined as the caliper diameter measured over a geometrical shape. For particles smaller than 250 μm, the b/l ratio decreases with increasing particle size, see Fig. 1b. Sphericity is derived from the ratio of measured area of the particle projection to the measured circumference of the particle projection. The sphericity decreases from 0.8 to 0.3 when increasing particle sizes from 20 μm to 280 μm. A snapshot image of pulverized pine particles in Fig. 1d confirms the shape irregularities and non-sphericity for biomass particles.
2.1. Particle characterization 2.2. Experimental setup and procedure Pine sawdust with a sieving size of 63–112 μm was used in this study. The characteristic size and the shape properties of the biomass sample were measured with a dynamic image analysis method (CAMSIZER XT, Retsch Technology GmbH) and are presented in Fig. 1. The particle size distribution of the biomass sample is given in terms of cumulative distribution, Q, see Fig. 1a. Here, particle diameter, i.e. the shortest chord among the maximum chords in all directions, is defined as the particle size since it is more relevant to the sieve size. Sphericity (SPHT) and width to length ratio (b/l) for different particle diameter are reported in Fig. 1b and c, respectively. The b/l ratio is calculated as the ratio of the shortest maximum chord of a particle projection to the maximum Feret diameter of a particle projection. The Feret diameter
The experiments were carried out with a laminar cold-flow jet in a laboratory scale, atmospheric entrained flow reactor. The experimental set-up is illustrated in Fig. 2. The experimental-setup mainly consists of an air supply unit, a wood powder feeder, a synthetic jet actuator unit, a flat flame burner (FFB), a quartz reactor tube, a char bin, an air seeding unit, and an air exhaust unit. Pine sawdust was fed into the flow with a syringe pump feeder and was entrained into a central tube of a commercial burner (Mckenna Flat Flame Burner by Holthuis & Associates) with a carrier air flow (0.43 l min− 1). Particle feeding rates used in this study were 14.5 g h−1, 25.5 g h−1, 47.3 g h−1 and 91.0 g h−1 that corresponded
Fig. 3. The mean binary particle images obtained at mass loading ratio of 0.47 by averaging over 200 sample images at a distance of 2–30 mm from the burner outlet for (a) base case (b) 35 Hz (c) 289 Hz (d) 451 Hz. Note that the images are false-color.
278
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
to the mass loading ratios of biomass to air as 0.47, 0.75, 1.39 and 2.67. The loading ratios were calculated with respect to the flow rate of the carrier air. The burner has two co-axial water-cooled porous discs that surround the central tube. In normal cases, a premixed mixture of methane and air (methane: 1.92 l min−1, air: 14.6 l min−1) is ignited and stabilized on the innermost stainless steel porous disc and Nitrogen (8.49 l min−1) is supplied from the outermost bronze porous disc to shield the flame. In this study, methane was replaced with air since the scope of the study encompassed visualization of the particle-flow interactions under cold flow conditions. More detailed information regarding the powder feeder and the flat flame burner can be obtained from the study of Göktepe et al. [6]. The synthetic jet actuator consists of a 4-ohm loudspeaker, (ND1404-5-1/4″, Dayton), which is fitted into an annular cavity with a central orifice of 4 mm. It produces a pulsed flow ejected into and out from the small orifice during periodic motions of the loudspeaker's diaphragm. The synthetic jet actuator was mounted perpendicular to the central tube of the burner via a Tee union adapter and an extension pipe. A fine porous steel mesh was attached to the entrance of the arm of the Tee-union adapter to avoid the accumulation of fine pine sawdust particles in the cavity and along an inner surface of the
extension pipe during the suction part of the actuation cycle. The loudspeaker was driven by a sinusoidal signal. The signal was generated by a data acquisition board (DT-9841-VIB, Data Translation), and was amplified by a stereo amplifier (Integrated amplifier A-10 from Pioneer) prior to being sent to the loudspeaker. A high sensitivity dynamic pressure sensor (106B, PCB) was flush mounted on the inner wall of the cavity. This sensor measures dynamic pressure fluctuations (in terms of Prms) developed inside the cavity. The sine-wave of 35 Hz with the peak to peak amplitudes of 0.15 V, 0.29 V and 0.48 V resulted in pressure fluctuations of 0.346 kPa (144.7 dB), 0.668 kPa (150.8 dB) and 1.106 kPa (154.8 dB), respectively. Accelerometer sensors (Delta Tron 4507, Bruel&Kjaer) were also attached on the external surface of the central pipeline to monitor the structural resonant frequencies. Particle and air velocities were measured at the same time with particle image velocimetry (PIV), a system from LaVision GmbH. The PIV system consists of double pulsed Nd-YAG laser with a maximum repetition rate of 100 Hz, a Lavision FlowMaster Imager Pro CCD-camera with a spatial resolution of 1280 × 1024 pixels per frame, timing controller and the PIV software program (LaVision Davis 7.2). The carrier air flow was seeded with aerosol droplets with diameters of 0.3 μm that were produced by atomizing Di-Ethyl-Hexyl-Sebacat (DEHS) with an
Fig. 4. Snapshot images of biomass particles at a distance range of 2–30 mm from the burner outlet in (a) the unforced laminar flow with the mass loading ratio of 0.47 (b) the forced initially laminar flow with the mass loading ratio of 0.47 (c) the unforced laminar flow with the mass loading ratio of 2.67 (d) the forced initially laminar flow with the mass loading ratio of 2.67. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa.
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
aerosol generator. The PIV measurements were conducted at a frequency of 34.65 Hz, with a total of 700 image pairs for each recorded set (sampled during around 20 s). Each set was repeated four times to check the repeatability of the measurements. The sampling frequency was adjusted to maintain in-phase coupling with time scales of the flow structures. A dynamic masking and image processing algorithm, which was previously used by Saber et al. [12], was employed to discriminate the biomass particles from the tracer aerosol particles. Based on the algorithm, the first step was to reverse black and white pixel colors in the raw binary images. In the next step, random noise, in the images, was removed by a 3 × 3 pixels smoothing filter. The biomass particles were detected by using an edge detection algorithm which detects the changes in contrast and the interior regions of the particles were filled by a dilute filter followed by an erosion filter. Biomass particles were sorted out from the images by using a threshold filter. The images were divided into interrogation windows of 32 × 32 pixels with a 50% overlap. A minimum of five tracer particles were detected in interrogation area which gave a reliable cross-correlation. The biomass particle velocities were measured by a particle tracking algorithm (PTV) which relies on tracking single particles in the flow domain. A constant temperature hot-wire probe (Type 55-P11 from Dantec Dynamics) was used to measure the air velocities in the unforced laminar flow. The probe has a 5 μm diameter platinum-plated tungsten wire, which is welded directly to the prongs. The probe measures one component of gas velocity and is recommended to be used for measurements in one directional flow of low intensity. Under the aforementioned flow
279
conditions, the carrier air velocity ejected from the central tube was measured to be around 0.37 m s−1, which corresponds to a Reynolds number (Re) of 140, while the air supplied from the porous disc had a velocity of about 0.1 m s−1. 3. Results and discussion The frequency used in this study was selected in a screening experiment by operating the synthetic jet actuator at a range of frequencies, starting from 10 Hz to 600 Hz with a step resolution of 10 Hz. The biomass particle-air flow with the mass loading ratio of 0.47 was continuously illuminated along the reactor height by the laser sheet and were visually observed by the high speed CCD camera in order to determine the frequencies that lead to high particle dispersion. At the same time, the pressure signals were measured in the quartz reactor and in the cavity. However, the mounting assembly, attached on one of the reactor ports, was located a distance away from the burner outlet and the dynamic pressure fluctuations were already damped out at this position. A custom designed probe with a tip of 100 mm was therefore used to sense the pressure fluctuations in the burner near field. Since this probe has a linear frequency response between 0 and 100 Hz, the pressure measurements in the reactor were performed over this frequency spectrum. Based on visual observations it can be stated that high particle dispersion was achieved at low frequencies in the range 15 to 50 Hz. The pressure probe measurements showed that the strongest flow fluctuations were at 35 Hz, which corresponded to St = 0.57 (Strouhal
(a)
u/U j
−1 0 1 2 3
y/D j =1
−1 0 1 2 3
y/D j =2
−1 0 1 2 3 −1.5
y/D j =3 −1
−0.5
0
0.5
1
1.5
r/D j
(b) 4
1.106 kPa 0.346 kPa 0.07 kPa
3.5 3
u/U j
2.5 2 1.5 1 0.5 0
10
15
20
25
30
35
y/D j
Fig. 5. Phase-averaged (a) vorticity field (b) total in-plane velocity field of air flow for Prms = 1.106 kPa.
Fig. 6. (a) Dimensionless mean axial air velocity component vs. crosswise distance at three positions from the burner outlet. (b) Dimensionless mean centerline air velocity profile vs. streamwise distance. The forcing frequency was 35 Hz.
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
Fig. 7 presents the effect of mass loading ratio, between 0.47 and 2.67, on the mean axial air velocity and the mean centerline air velocity at Prms = 1.106 kPa. The velocity magnitude was lower for the case with high mass loading ratio since the particles or particle aggregates tend to act like a semi solid obstacle against the air flow. In Fig. 7b, the mean centerline air velocity profile shows that the peak to peak distance and hence the distance between two vortices decreased with increasing the loading ratio. This can be interpreted as if the vortex rings slowed down, or were blocked, at high mass loading ratio. The phase-averaged streamwise velocity for each particle, normalized by jet velocity, u/Uj, for three sizes ranges of particles was calculated for 10 samples and is presented in Fig. 8a–c. Please note that the size of the circles gives a qualitative indication of the particle size while the color indicates the normalized particle velocity. As expected, particles passing through the vortex ring have higher stream-wise velocity while the particles located around the vortex ring have low velocity. It is also shown that for the larger particles the concentration is higher at the jet core and lower in the periphery of the image. The phaseaveraged lateral velocity of the particles were derived in a similar manner and plotted in Fig. 8d–f. The small particles are convected away from the jet core due to the rotating vortex rings for which the rotational direction is directed downwards near the axis and upwards at the outer periphery of the vortex rings. Some of the particles are entrained into the vortex ring edge and re-enter the jet core. However, in the core of the reverse flow zone, no particles were detected due to centrifugal
(a)
u/U j
number: St = fd/u, where f is forcing frequency, d is the nozzle diameter of biomass-air flow, and u is axial air velocity). This agrees well with reports in the literature that states that the maximum mixing enhancement for jet flow is in the range of St = 0.3–0.6 [11]. The resonance frequency for the cavity and the central pipeline were measured to be 289 Hz and 451 Hz, respectively. Fig. 3 shows the mean binary images of biomass particles for these three characteristic frequencies (35, 289 and 451 Hz). The particle in each image was detected by using the algorithm used in our previous study [6]. Then, the mean value for each pixel was calculated from 200 binary images, where 1 means the presence of particles. The particles are more likely to exist in the position with lighter color. It shows that the highest particle dispersion was achieved when the synthetic jet actuator was operated at 35 Hz. The resonant vibrations at 451 Hz in the central tube showed some effects on the particle dispersion albeit the effect was not as strong as that at 35 Hz. Compared to 35 Hz, the resonant frequency of the cavity at 289 Hz caused even higher dynamic pressure fluctuations inside the reactor, but did not enhance dispersion of the particles significantly. Fig. 4 shows the features of the unforced and forced particle-air flows at a distance range of 2–30 mm from the burner outlet. The biomass particles in the undisturbed laminar flows agglomerated to form larger particles and the agglomeration was more pronounced for high mass loading ratios as shown in Fig. 4a and c. This may be linked to cohesive forces or electrostatic forces among the biomass particles. When the air flow was pulsed at 35 Hz, vortex rings were formed by vortex shedding at the burner rim and then propagated downwards in the axial direction. The convective effect induced by the vortex rings spread the particles laterally over a large distance in the reactor as shown in Fig. 4b and d. Moreover, the vortex flows caused very little or no particle agglomeration. Fig. 5 shows the phase-averaged vorticity and total air velocity in the biomass-air flow pulsating at the forcing amplitude, Prms = 1.106 kPa. A visual inspection of the figures yields that the vortex ring divided the flow into two regions: a reversed flow zone (around the vortex rings) and a high speed zone (inside the vortex rings). The counter rotating flow zones drag a part of the surrounding air flow into the vortex centerline, leading to an increase of the total velocity in this region. The maximum velocity measured in the vortex centerline is about 2 m s−1 and drops to 0.2 m s− 1 around the boundary of the reversed flow zone, leading to high velocity gradients and hence shear stresses. It has previously been shown that high shear rates (velocity gradient larger than 30 s−1) break up particle aggregates, causing a reduction in the mean particle size [13]. Returning to Fig. 4, the high local shear stresses may thus be the reason for the break-up of particle aggregates in the burner near field. Alternatively, the break-up of particle aggregates might have already taken place in the central tube. Fig. 6 presents the effect of the forcing amplitude on the mean axial air velocity and the mean centerline air velocity. The mean axial air and the mean centerline air velocities were non-dimensionalized with the centerline velocity at the burner outlet obtained from the undisturbed laminar jet velocity, Uj (0.37 m s−1). The lateral and axial directions were normalized with the central tube diameter, Dj. Fig. 6a shows that the mean axial air velocity increased with the excitation amplitude. The flow is highly asymmetric for Prms = 1.106 kPa at y/Dj = 1. The asymmetry in the flow originates from the asymmetric position of the synthetic jet actuator and the main jet flow is deflected away from the synthetic jet side. Similar results have been reported previously [10]. Moreover, the negative local mean velocities at x/Dj = ± 1 reveals that the flow is reversed towards the burner rim under the effect of suction. In Fig. 6b, a peak to peak distance can be defined as the distance between the first maximum value and the second maximum value of the dimensionless velocity and can be interpreted as a distance between two vortex rings (cf. with Fig. 4b). From the figure, it can be concluded that the peak-to-peak distance increases with increasing the forcing amplitude. This can be interpreted as if the vortex rings will propagate faster when the forcing amplitude is increased.
−1 0 1 2 3 −1 0 1 2 3 −1 0 1 2 3 −1.5
y/D j =1
y/D j =2
y/D j =3 −1
−0.5
0
0.5
1
1.5
r/D j
(b) 4
Mass loading ratio 0.47 2.67
3.5 3 2.5
u/U j
280
2
1.5 1 0.5 0
10
15
20
25
30
35
y/D j Fig. 7. (a) Axial component of the averaged air velocity profiles vs. crosswise distance and (b) mean centerline velocity profile vs. streamwise distance measured in the flow pulsating at Prms = 1.106 kPa. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa.
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
25
(a)
(b)
50-83 µm
281
1.4
(c)
83-117 µm
117-150 µm
1.2
20
y mm
1 15
0.8 0.6
10
0.4 5 0
0.2 0
5
10
15
20 0
5
10
x mm
15
20
0
5
10
x mm
15
20
x mm
0
u/Uj 0.3
25
(d)
(e)
50-83 µm
(f)
83-117 µm
117-150 µm 0.2
20
y mm
0.1 15 0 10 −0.1 5 0
−0.2
0
5
10
15
20 0
5
10
x mm
15
20
0
5
10
x mm
15
20
x mm
−0.3
v/Uj
Fig. 8. (a–c) Phase-averaged particle images obtained from 10 sample velocity fields, represented with particle streamwise velocity u/Uj and the particle size. (d–f) phase-averaged particle images represented with the particle radial velocity and the particle size. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa. Note that the size of the particles is qualitatively represented by the size of the circles.
25
(a)
17
12.5
(b)
(c) 20
16
11.5
y mm
15 15
10.5 10
14
9.5 5
13 8.5
9.5
10.5
8.5 14.5
11.5 0
(d)
180º 210º
5
10
x mm
150º
240º
0
15
20
(e)
15.5
16.5 180º
210º
150º
240º
1500
500 400
1000 200
500
270º
90º
300º
60º 30º
330º 0º
17.5
α
270º
90º
300º
60º 30º
330º 0º
Fig. 9. (a) phase-averaged particle image and schematic representation of orientation angle of a single particle with respect to flow direction, zoom-in sections of (b) main jet flow (c) of vortex flow, particles alignment angle with respect to flow direction in region of (d) main jet flow, (e) the vortex region. The forcing frequency was 35 Hz and amplitude was Prms = 1.106 kPa.
282
A.H. Saber et al. / Powder Technology 302 (2016) 275–282
forces. The lateral particle velocity increased laterally towards the reverse flow zones. Fig. 9 shows the orientation of the biomass particles and particle velocity vectors calculated from the phase-averaged particle image for Prms = 1.106 kPa and for the mass loading ratio of 0.47. The particle orientation with respect to the stream-wise (or axial) direction is defined by an angle, α in Fig. 9a, and is illustrated in polar coordinates in Fig. 9d–e. For a given particle size range, the majority of the biomass particles in the high shear region shown in Fig. 9b was oriented over a range of angles from 330° to 30° to the flow direction, (see, Fig. 9d) while a larger portion of smaller particles around the vortex ring shown in Fig. 9c was aligned over a range of angles from 30° to 90° with the streamwise direction (see Fig. 9e). This shows that the biomass particles for a given size range had a tendency to align themselves along the streamlines of the flow. The average inter-particle distance is shown against the forcing amplitude in Fig. 10. The data of unforced flow (Prms = 0 kPa) was removed since the particle size distribution analyses showed the sign of agglomeration and does not represent the isolated effect of particle dispersion. Except for the flows with the highest mass loading ratio of 2.67, the inter-particle distance increased with increasing forcing amplitude. The effect was more pronounced for the flows with low mass loading ratios. For instance for a mass loading ratio of 0.47, the trend was quasi linear. The inter-particle distance decreased with increasing the mass loading ratio.
Experiments were performed with pine saw dust particles in a laboratory scale entrained flow reactor. The biomass particle flows as well as the carrier air flow were excited by a synthetic jet actuator to investigate the effect on the flow and the inter-particle distance. Velocity measurements were conducted with a PIV system, both for the continuous phase and for the particle phase. The flow was excited with three different excitation amplitudes and the effect of four different mass loading ratios were investigated. The results can be summarized as follows.
Inter−particle distance /mean particle diameter
• The synthetic jet actuator produced a continuous train of vortex rings when the forcing frequency was in the range 10 to 50 Hz. However, the strongest vortex rings and the highest particle dispersion were achieved at 35 Hz. • The biomass particles tend to agglomerate in the unforced laminar flows, but the particle aggregates disappeared in the forced flows, even at high loading ratios (2.67). 15
Mass loading ratio 0.47 1.39
0.75 2.67
10
5
0
Acknowledgment The authors would like to thank financial support by Norbottens Research Council (14-206), Kempe Foundation (SMK-1232), and Swedish Energy Agency for supporting this work.
4. Conclusions
0
• The strong vortex rings generated two different zones: one high momentum zone (between the vortex rings) and one high vorticity zone (the vortex rings themselves). The larger particles with high inertia were mainly concentrated between the vortex rings while the smaller particles were convected radially away from the jet core by the rotational motion of the vortex rings. • The particle loading ratios affected the flow features of the trailing vortex rings. For instance, the distance between the velocity peaks, interpreted as the distance between the centers of two vortex rings, increased with increasing actuator power and decreased with increasing particle mass loading ratio. • In dense particle-air flows, the presence of particles reduced the centerline air velocity while the centerline velocity increased with increasing actuator power. • The particle alignment with respect to the streamwise direction was different for the high momentum (between vortex rings) and high vorticity (inside the vortex rings) flow regions. The particles in the high momentum zone were aligned with the streamwise direction within a range of angles ±30° while the particles at the edge of the vortex rings were aligned within a range of angles between ± 30– 90°, i.e. close to the local direction of gas flow.
0.2
0.4
0.6
0.8
1
1.2
prms, kPa Fig. 10. Inter-particle distance calculated for different loading ratios, as function of Prms. The forcing frequency was 35 Hz.
References [1] M. Asadullah, Barriers of commercial power generation using biomass gasification gas: a review, Renew. Sust. Energ. Rev. 29 (2014) 201–215, http://dx.doi.org/10. 1016/j.rser.2013.08.074. [2] A. Joppich, H. Salman, Wood powder feeding, difficulties and solutions, Biomass Bioenergy 16 (1999) 191–198, http://dx.doi.org/10.1016/S0961-9534(98)00082-8. [3] K. Svoboda, M. Pohořelý, M. Hartman, J. Martinec, Pretreatment and feeding of biomass for pressurized entrained flow gasification, Fuel Process. Technol. 90 (2009) 629–635, http://dx.doi.org/10.1016/j.fuproc.2008.12.005. [4] F. Weiland, H. Hedman, M. Marklund, H. Wiinikka, O. Öhrman, R. Gebart, Pressurized oxygen blown entrained-flow gasification of wood powder, Energy Fuel 27 (2013) 932–941, http://dx.doi.org/10.1021/ef301803s. [5] W.A. Sirignano, Advances in droplet array combustion theory and modeling, Prog. Energy Combust. Sci. 42 (2014) 54–86, http://dx.doi.org/10.1016/j.pecs.2014.01. 002. [6] B. Göktepe, K. Umeki, R. Gebart, Does distance among biomass particles affect soot formation in an entrained flow gasification process? Fuel Process. Technol. 141 (2016) 99–105, http://dx.doi.org/10.1016/j.fuproc.2015.06.038. [7] R.B. Wicker, J.K. Eaton, Structure of a swirling, recirculating coaxial free jet and its effect on particle motion, Int. J. Multiphase Flow 27 (2001) 949–970, http://dx. doi.org/10.1016/S0301-9322(00)00061-6. [8] L.M. Cerecedo, L. Aísa, J.A. García, J.L. Santolaya, Changes in a coflowing jet structure caused by acoustic forcing, Exp. Fluids 36 (2004) 867–878, http://dx.doi.org/10. 1007/s00348-003-0769-8. [9] S. Balachandar, J.K. Eaton, Turbulent dispersed multiphase flow, Annu. Rev. Fluid Mech. 42 (2010) 111–133, http://dx.doi.org/10.1146/annurev.fluid.010908. 165243. [10] D.A. Tamburello, M. Amitay, Active manipulation of a particle-laden jet, Int. J. Multiphase Flow 34 (2008) 829–851, http://dx.doi.org/10.1016/j.ijmultiphaseflow.2008. 02.006. [11] E. Mastorakos, M. Shibasaki, K. Hishida, Mixing enhancement in axisymmetric turbulent isothermal and buoyant jets, Exp. Fluids 20 (1996) 279–290, http://dx.doi. org/10.1007/BF00192673. [12] A. Saber, S. Lundström, G. Hellström, Influence of inertial particles on turbulence characteristics in the near and Outer Wall flow as revealed with high resolution PIV, J. Fluids Eng. 138 (2016) 091303, http://dx.doi.org/10.1115/1.4033369. [13] T. Serra, J. Colomer, B.E. Logan, Efficiency of different shear devices on flocculation, Water Res. 42 (2008) 1113–1121, http://dx.doi.org/10.1016/j.watres.2007.08.027.