- Email: [email protected]
Modelling fluidized dense-phase pneumatic conveying of fly ash
Powder Technology 270 (2015) 39–45
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Modelling fluidized dense-phase pneumatic conveying of fly ash G. Setia ⁎, S.S. Mallick Department of Mechanical Engineering, Thapar University, Patiala, Punjab 147004, India
a r t i c l e
i n f o
Article history: Received 8 August 2014 Received in revised form 16 September 2014 Accepted 16 September 2014 Available online 23 September 2014 Keywords: Fluidized dense-phase Conveying characteristics Minimum transport boundary Solid friction factor Volumetric loading ratio
a b s t r a c t This paper presents the results of an ongoing investigation into modelling important design criteria, such as minimum transport condition, straight-pipe pressure drop and solid friction factor for fluidized dense-phase pneumatic conveying of powders. Fly ash (median particle diameter: 19 μm; particle density: 1950 kg/m3; loose-poured bulk density: 950 kg/m3) was conveyed over a wide range of flow conditions (from fluidized dense- to dilute-phase) under different conditions of pipeline diameters and lengths (viz. 43 mm I.D × 24 m length, 54 mm I.D × 24 m length, 69 mm I.D × 24 m length and 69 mm I.D × 70 m length). To define the safe minimum transport boundary, a Froude number based criteria at the pipe inlet has been used (Fri = 7). The Froude number based criterion is aimed to address the requirement of different conveying velocities for different pipe diameters. Straight-pipe pneumatic conveying characteristics obtained from two sets of pressure tapings installed at different locations of pipeline have shown that the trends and relative magnitudes of the pressure drops can be significantly different depending on the location of pressure tapping points, thus indicating a change in flow mechanism along the direction of flow. A new approach of modelling solid friction factor using a volumetric loading ratio term has provided better scale-up accuracy when the model predictions were compared with experimental data. This method of modelling solid friction is aimed to address the partial filling of pipe's cross section by the dune of solids, which appears to be a better representation of the flow conditions, especially for the dense-phase pneumatic conveying of fine powders. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Pneumatic conveying of fly ash is widely used in pulverized coal fired thermal power plants all over the world. Such dry handling and conveying technology make it possible for the fly ash to be subsequently used for other purposes, such as fly ash in dry form can be mixed with cement. Other reasons include: pneumatic conveying is totally enclosed, environmentally friendly, provides increased workplace safety; relatively low capital/maintenance costs (for a well-designed system); layout flexibility; and ease of automation and installation. For countries such as India, where about 60% of total installed capacity of electricity generation is through coal fired power generation route using coals having ash content of as high as 40–50% (for worst coals), a large amount of fly ash has to be pneumatically conveyed everyday from Electro Static Precipitator (ESP) hoppers to remote silos with a total pipeline length of even up to 1 km (depending on the plant layout). This is achieved by either using a combination of negative and positive pressure systems (i.e. by evacuating the ash from individual ESP hoppers to a local surge hopper using a vacuum conveying system and then using a positive pressure conveying system to transport fly ash from the surge hopper to remote silos) or by positive pressure systems only (where blow tanks are placed under each ESP hopper, which directly transfer the ⁎ Corresponding author. Tel.: +91 175 2393370. E-mail address: [email protected] (G. Setia).
http://dx.doi.org/10.1016/j.powtec.2014.09.033 0032-5910/© 2014 Elsevier B.V. All rights reserved.
fly ash to remote silo using compressed air as the transport medium). The positive pressure systems can be of two types: dilute-phase (or suspension flow) and dense-phase (or non-suspension flow). In dilutephase conveying, the carrier gas velocity is sufficiently high to suspend all the particles for conveying. Due to the dispersed and suspended nature of the dilute-phase conveying, the designers and researchers have achieved relatively good success in modelling the relevant particle interactions and mechanisms to develop numerous friction factor and pressure drop models [1]. However, this mode of conveying results in higher gas flows (and reduced solid loading ratio), larger sized compressor (i.e. higher energy consumption), larger pipe, fitting and support dimensions, increased chance of wear of pipe and bends (i.e. increased maintenance requirement) and larger sizing of bag filters. Due to these disadvantages of dilute-phase systems, the plant owners and designers are showing increasing interest in recent times to employ dense-phase pneumatic conveying systems (instead of the conventional dilute-phase mode) to take advantage of the fluidized dune flow capabilities of fly ash under aerated condition. For about 800– 1000 m long pipe (e.g. from intermediate surge hopper to remote silo in a power plant), a well designed fly ash dense-phase pneumatic conveying system would typically operate with a solid loading ratio of 30 (instead of 10–15 for a dilute-phase system) and with an initial conveying velocity 5–7 m/s (instead of about 15 m/s for dilute-phase). This would result in a substantial amount of savings in energy and initial capital costs in terms of smaller compressor and pipe sizes and maintenance
40
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
(reduced frequency of wearing of bends). In spite of such merits and potential to be a better alternative of fly ash conveying, wide spread installation of such system is still limited as designing a reliable fluidized dense-phase system is a difficult task due to the highly turbulent and complex nature of the concentrated dune flow [1,2]. For a reliable design of industrial scale pneumatic conveying systems, the two most important parameters that are to be accurately estimated are: (i) total pipeline pressure drop and (ii) minimum transport criteria. Over-estimation of total pipeline pressure drop would result in unnecessarily higher supply of air flow, energy cost and wearing of pipeline due to excessive conveying velocity. Under-prediction of total pipeline pressure drop would cause reduced material throughput rate. Incorrect estimation of minimum transport condition will result in particle deposition in pipes, leading to pipe blockage [1,3]. In fact, a majority of the practical installations facing unreliable operations and requiring troubleshooting suffer from problems arising out of the above two reasons. Hence, these design parameters (pipeline pressure drop and blockage criteria) must be accurately modelled and scaled-up with high reliability. The pressure drop for the flow of solid and gas flow through a straight horizontal pipe section can be represented using Eq. (1). This equation was originally presented by Barth [4] and believed to be for coarse particles in dilutephase flow. However, various other researchers, such as Pan and Wypych [2], Stegmaier [5], Weber [6], Rizk [7], Pan and Wypych [8] and Jones and Williams [9] have subsequently used this equation to predict the pressure loss for the fluidized dense-phase pneumatic transport of powders, such as fly ash, for horizontal straight pipes. 2 ΔP ¼ λ f þ m λs ðL=DÞρ V =2
ð1Þ
The main task in Eq. (1) is to model the solid friction factor accurately. Whereas more fundamental modelling and scale-up procedures based on powder mechanics have been established for lowvelocity dense-phase slug-flow of granular materials, modelling solid friction for fluidized dense-phase conveying of fine powders has remained a relatively more difficult challenge due to the complex and turbulent nature of the moving bed [1], where it is very difficult to link the particle and bulk properties to the flow mechanisms during actual conveying conditions. Due to these difficulties in modelling, empirical power function based models have been employed by various previous researchers [4–11] over the years to avoid the need of developing fundamental relationships between solid friction factor and the relevant particle and bulk properties. These models have used different dimensionless parameters and have shown only limited success under scale-up evaluation [1]. Hence, there is a requirement to conduct further studies on important design parameters such as straight pipe pressure drop models for solid friction and minimum transport criteria with an aim to provide the industry a reliable scale-up procedures for dense-phase pneumatic conveying of fly ash. 2. Experimental data Conveying trails were performed using the Indian fly ash with different pipeline configurations. Table 1 lists the physical properties of the fly ash. A typical schematic of the test set up used for fly ash conveying is shown in Fig. 1. Kirloskar made electric-powered Model KES 18-7.5 rotary screw compressor was used having the capacity of 3.37 m3/min Table 1 Physical properties of fly ash conveyed. Powder
Median particle diameter, d50 (μm)
Particle density, ρs (kg/m3)
Bulk density, ρbl (kg/m3)
Fly ash
19
1950
950
Fig. 1. Layout of the 54 mm I.D. × 70 m test rig.
of free air delivery and maximum delivery pressure of 750 kPa. Air flow control valve was installed in the compressed air line upstream of the blow tank to vary the conveying air flow rates over a wide range of air flows. For the measurement of air flow rates, a vortex flow metre was installed in the compressed air line. Bottom discharge type blow-tank was used as feeder having 0.2 m3 capacity of water fill volume. It was mounted with solenoid operated dome type material inlet, outlet and vent valves. The blow tank was provided with air supply system with the help of orifice plate at the air inlet to blow tank. A receiver bin of 0.65 m3 capacity was installed on top of blow tank. The receiver bin was fitted with bag filters having pulse jet type cleaning mechanism. The blow tank and receiver bin were supported by shear beam type load cells. Different mild steel pipelines of 43 mm I.D × 24 m length, 54 mm I.D × 24 m length, 69 mm I.D × 24 m length and 69 mm I.D × 70 m length were used as the test rigs. The test loops included a 3 m vertical lift and 4 × 90o bends having 1 m radius of curvature. Various static pressure measurement points were installed along the pipeline and bends, where P5 was used to measure the total pipeline pressure drop and all other transducers (i.e. P6 to P9) were installed to measure static pressures at the respective points, which were used to model solid friction factor of straight horizontal pipe. P6 and P7 were to measure static pressure in the initial part of the test loop, whereas P8 and P9 were used to measure static pressure for the later part of pipeline. All the static pressure transducers in the solid– gas flow line were strategically placed sufficient distance away from the influence of any change in flow pattern caused by the bends. Specification of static pressure transducers: manufacturer: Endress & Hauser, model: Cerabar PMC131, pressure range: 0–2 bar, maximum pressure: 3.5 bar (absolute), current signal: 4–20 mA. All other required instruments such as PRV (pressure reducing valve), flow metre, NRV (non-return valve), blow valve, pressure gauge and load cells (shear beam type) were suitably placed. Calibration of the pressure transducer, load cells and flow metre was performed using a standardized calibration procedure [1]. To record the electrical output signals from the load cells, pressure transducers and flow metre, a portable PC compatible data logger was used. The data logger had 16 different channels with 14 bit resolution. Every pipeline was installed with two sets of 300 mm long sight-glasses made of borosilicate materials for flow visualization (using high speed digital camera).
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
A unique data-base of total pipeline, straight pipe pressure loss and minimum conveying condition have been generated by conveying for the Indian fly ash for a wide range of flow conditions through different pipelines. Observation through the sight glass showed significant amount of non-suspension flow (a dense highly turbulent non-suspension layer of powders flowing along the bottom of pipe with particles travelling on top in suspension flow), thus confirming fluidized densephase convey ability of the Indian fly ash. At higher air flows, the fly ash was conveyed in dilute-phase. With gradual reduction of air flow rate, continuous dune flow was observed. Further reduction of air velocity resulted in pulse-type dunning or discontinuous dunes. Further reduction of air flows provided high pressure fluctuations (i.e. unstable conveying) and ultimately complete pipe blockage. The unstable conveying regime is characterized by a gradual build up of material in the pipeline, where the amount of material which is pushed in the line, does not completely return to the receiver bin (which was found from the records of load cells installed under blow tank and receiver bin). In the present study, this unstable-phase conveying is considered as the initiation of blockage condition as repeated attempts of conveying with such a gradual product build-up condition would completely block the pipeline in few cycles of operation in industry. It was found that experimentally it would be difficult to be very precise about the air flow rate for which line blockage would occur due to the practical limitation of setting the air flow control valve exactly for the blockage condition. Therefore a series of tests were carried out near the blockage boundary to define a zone of air flow rates for which line blockage would take place. The solid flow rates were varied using different prepressurizations of the blow tank aeration. Certain tests were performed multiple times to ensure repeatability of test data. 2.1. Pneumatic conveying characteristics and minimum transport criteria Pneumatic conveying characteristics (PCC) for a given product are the representation of variation of pressure drop for different solid and air flow rates (e.g. dense- to dilute-phase). PCC constitute an essential requirement for the reliable design of pneumatic transportation systems by providing valuable information, such as the expected mode of flow, minimum conveying air velocity requirement to prevent blockage and the optimal operating conditions. Total pipeline PCC for fly ash is based on the steady-state pressure drop data obtained from the pressure transducer P5 (Fig. 1) and are shown in Figs. 2 to 5 for the different pipelines. Experimental blockage boundaries are shown in the PCC. Different predictions of blockage boundary (or the minimum transport line) using different models for minimum transport conditions have been superimposed on the above figures. The lines joining the initiation of unstable zone have been designated as blockage boundary or minimum transport condition in this paper. All the PCC (Figs. 2 to 5) show
Fig. 2. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 54 mm I.D. × 70 m long pipe.
41
Fig. 3. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 54 mm I.D. × 24 m long pipe.
that pressure drop rises slowly with an increase in air flow rate up to about 0.025 kg/s of air flow. Pressure drops sharply increase with further rise in air flow rates. However, total pipeline PCC are not conclusive about the pressure drop characteristics (change in pressure drop with increase in air flow rate) for straight pipes, as the total pipeline PCC are formed by the superimposition of pressure drop characteristics of the bends to that of straight pipes. Hence, a relatively large contribution of the bend loss characteristics could have considerably influenced the overall trends of the total pipeline PCC. The PCC (Figs. 2 to 5) do not show any clearly defined Pressure Minimum Curve (PMC), which indicates flow transition (from dense- to dilute-phase flow) [1]. For the whole range of test rigs, the Froude number at the feed point (bottom of blow tank) varied from 3.5 to 6.4. The following contains a list of existing models to predict the blockage boundary. Weber [6] provided the following correlation to predict the minimum conveying velocity to avoid line blockage. Weber [6], however, did not comment on the reliability and applicability of the model (i.e. he did not provide details of pipe and product on which the model should be applied): For U t ≤3m=s 0:25 0:1 Fri ¼ ½7 þ ð8=3ÞUt m ððds =DÞÞ
ð2Þ
For U t ≥3m=s 0:25 0:1 Fri ¼ 15 m ðds =DÞ :
ð3Þ
Fig. 4. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 43 mm I.D. × 24 m long pipe.
42
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
of Mallick and Wypych [13] was based mostly on the Australian fly ash data, which had a narrower size distribution (more finer) than the Indian fly ash having a wider range of size distribution having a considerable share of coarser particles. 2.2. Pressure drop characteristics for straight pipe sections
Fig. 5. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 69 mm I.D. × 24 m long pipe.
Martinussen [12] used test facility having horizontal pipeline of 53 mm I.D and 15 m length. By applying the fluid analogy, he developed the following model to determine the minimum conveying velocity, as given by Eq. (4). According to Martinussen [12], this model could provide better predictions for fine materials than for the coarser ones. 3 2 Vi ¼ KDgðρbl =ρ f Þ 1−m ðρ f =ρbl Þ
ð4Þ
where, K (geometrical factor) = П/4 at the filling level of D/2. Mills [3] conveyed cement through a 52.5 mm I.D. × 95 m test rig and proposed a constant velocity based criteria to represent the minimum conveying conditions in which he suggested the inlet velocity equal to 3 m/s as the minimum transport criteria for fine powders. He also proposed a power function type relationship between minimum inlet velocity and solid loading ratio (i.e. lower minimum conveying velocity is required for larger solid loading ratio). Mallick and Wypych [13] investigated into minimum conveying criteria using the experimental data from twelve different powders conveyed over a wide range of pipe lengths and diameters. It was found that the representation of minimum conveying velocity just as a function of solid loading ratio may not be adequate for reliable predictions under diameter scale-up conditions. Based on the experimental data of ESP dust and fly ash, they found that with increase in pipe diameter, the requirement of minimum conveying air velocity rises. To incorporate the effect of pipe diameter on minimum conveying velocity, they proposed a Froude number based criteria at the feed point (Fri = 6) as the minimum transport condition for powders having a loose-poured bulk density of ρbl ≤ 1000 kg/m3 and that are is suitable for fluidized dense-phase conveying. The above models (Weber [6] to Mallick and Wypych [13]) have been validated for their scale-up accuracy by using them to predict the blockage boundary for the given Indian fly ash. The results are shown in Figs. 2 to 5. Results show that the experimental minimum transport boundary vary in the range of Fri = 5 to 7. Mills' criteria of Vi = 3 consistently provided under-prediction for all cases as the actual conveying velocity requirements are higher than this. The Martinussen [12] model provided over-prediction for all the pipelines. In fact, this model resulted in over-predictions that were found to be out of the range of the PCC in Figs. 2 and 5, hence, predictions using this model have not been shown in these figures. Similarly, the Weber [6] model has also provided over-prediction in all cases, although the margin of overprediction is less than that predicted by Martinussen [12]. The Mallick and Wypych [13] model (Fri = 6) seems to be closest to the experimental observation. Considering the actual minimum transport boundary for the fly ash varied from Fri = 5–7, the safe criteria to avoid line blockage with the Indian fly ash is proposed here as Fri = 7. This criteria of minimum transport condition requires slightly higher velocities than that proposed earlier by Mallick and Wypych [13], as the Fri = 6 criteria
Using the pressure drop data between tapping points installed on straight pipe lengths P6–P7 and P8–P9 on the 54 mm I.D. × 70 m long pipeline, pneumatic conveying characteristics for straight pipes have been developed for different solid and air flow rates. These are shown in Fig. 6. It can be seen from Fig. 6 that depending on the location of tapping points, pressure drop characteristics get different. The pressure drop per unit length is more at the beginning of pipe compared to that at the end. This indicates a change in flow mechanism along the length of pipe. This could be due to the mixing of air with solid (fly ash) along the flow, which would cause the material dune to get aerated and become more fluidic, hence easier to transport. It is found that below the air flow rate of 0.02 kg/s, the pressure drop values almost do not change with increase in air flow. However with further rise in air flow rates, the pressure drop values have increased rather sharply with an increase in air flow rate, thus indicating a gradual change from non-suspension to suspension mode of flow. Experimentally it was found (view of flow through sight-glass) that the transition from non-suspension to suspension mode was rather gradual with no clear location of pressure minimum curve on the straightpipe PCC. 3. Modelling and validation of solid friction factor models In order to model solid friction factor for straight pipes, the following two formats have been used in the present study: a b c Model 1 : λs ¼ K m ð FrÞ ðρ f =ρs Þ
ð5Þ
and, a
b
Model 2 : λs ¼ KðVLRÞ ðw f0 =VÞ
ð6Þ
where, volumetric loading ratio (VLR) is given as: VLR ¼ fðms =ρs Þ=ðm f =ρ f Þg:
ð7Þ
Several previous researchers have used m* and Fr in power function format to model solid friction factor, such as Mallick [1] and Pan and Wypych [8]. However, these models do not take care of the effect
Fig. 6. Comparison between straight-pipe PCC, fly ash, 54 mm I.D. × 70 m long pipe.
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
of reduction of gas density along the direction of flow. Hence, to account for the same, a dimensionless air to particle density ratio term has been introduced in the format of modelling solid friction factor (Model 1). A different approach to modelling has been provided by Model 2, which uses a volumetric loading ratio term. In this approach, it is considered that the solid loading ratio term (which is a ratio of the masses) does not properly account for important flow criteria, such as the pipe volume fraction that is actually available for the flow of air, as a considerable portion of the pipe section gets filled by products in dense-phase. Hence, the effective area for air flow is considerably reduced. To account for this partial filling up of pipe, the volumetric loading ratio term is proposed here (Eq. (7)). Rautiainen et al. [14] mentioned that solid volume fraction is an important contributor to the pressure drop of solid–gas flows as the occupied volume of solid is expected to influence the available gas path, which would contribute to additional energy drop. Ahmed and Elgobashi [15] proposed that the solid–gas flows will introduce two-way coupling and influence the turbulence and pressure drop of flow. Huber and Sommerfield [16] suggested that small particles would follow the eddies of the gas-phase and as a result, there would not be much relative velocities amongst the particles within an eddy. This would reduce the inter-particle collision effect. The above suggests that the amount of pipe volume occupied by the solids is an important feature towards representing the flow mechanism of dense-phase solid–gas flows. Therefore, the volumetric loading ratio term has been introduced in the present study to model solid friction factor. The wf0/V term attempts to indicate dense- to dilute-phase transition phenomenon. Larger sized and/or higher particle density products will have increased free settling velocity. It can be thought intuitively that particles having larger value of free settling velocity (i.e. higher tendency of settling down) will require higher velocity to keep them in suspension. Using steady-state pressure drop data between P6 and P7 tapping points over a range of solid and air flow rates, the following models are devolved using both the formats using regression analysis in Microsoft Excel: h i −0:696 −1:392 −0:458 2 Model 1 format : λs ¼ 169:96 m ð FrÞ ðρs =ρ f Þ R ¼ 0:985
ð8Þ
−0:316
Model 2 format : λs ¼ 0:643ðVLRÞ
0:777
ðw fo =VÞ
h i 2 R ¼ 0:930 : ð9Þ
Figs. 7 and 8 show plots of y = λs (Fr)a(ρs/ρf)b versus x = m* and y = λs (wfo/V)c versus x = VLR through the experimental data for Model 1 and Model 2, respectively, where ‘a’, ‘b’ and ‘c’ are the adjustable exponents. R2 values from the trends are also provided in figures. High values of R2 were obtained using the method of least squares to develop the model that indicates good agreement of straight pipe data
Fig. 7. Development of Model 1 — trends of y = λs (Fr)a(ρs/ρf)b versus x = m*.
43
Fig. 8. Development of Model 2 — trends of y = λs (wfo/V)c versus x = VLR.
sets with the model estimates. Terminal settling velocity of solids has been calculated as 0.0235 m/s using the following expression (Eq. (10)) and using appropriate drag coefficient as per [17]: w fo ¼
4gdðρs −ρ f Þ 0:5 : 3CD ρ f
ð10Þ
The above solid friction factor models have been evaluated by using them to predict the total pipeline drop for the two pipelines: 54 mm I.D. × 70 m long and 69 mm I.D. × 24 m for different solid and air flow rates and by comparing the predicted pneumatic conveying characteristics against the experimental plots. To predict the losses in bends, Chambers and Marcus [10] model was used (Eq. (11)). The results are shown in Figs. 9 and 10. 2 ΔPb ¼ NB 1 þ m ρV =2
ð11Þ
Fig. 9 shows that Model 1 provides under-prediction for 3.5 t/h of solid flow rate, whereas it provides over-prediction for the 1.5 t/h line. The model somewhat over-predicts in the lower velocity range and begins to under-predict with rise in air velocity. Model 2 provides good predictions for the higher tonnage (3.5 t/h), but somewhat overpredicts for the lower tonnages. Fig. 10 shows that the Model 1 format provides considerable amount of under-predictions for diameter scale-up conditions for all tonnages, whereas Model 2 provides good predictions for the 5.5 t/h and better predictions than Model 1 for the lower tonnages (the level of underpredictions have reduced with Model 2). The estimated pressure drop values are found to be increasing with increase in air flow rates. This matches the experimental trends for air flow rates above 0.045 kg/s. Below this air flow rate, the experimental trends show a change (decrease) in slope (in highly dense-phase region that is closer to the
Fig. 9. Experimental versus predicted PCC, fly ash, 54 mm I.D. × 70 m long pipe.
44
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
plants that handle fly ash having wide size distribution (material characteristics similar to the conveyed fly ash of this paper). Further scope of work will include incorporation of specific material characteristics to model minimum transport criteria and solid friction factor. List of symbols and abbreviations
Fig. 10. Experimental versus predicted PCC, fly ash, 69 mm I.D. × 24 m pipe.
blockage boundary), whereas, the predicted trends (model predictions) appear to be monotonously linear. Therefore, further work is required for better predictability of solid friction in very dense-phase region. 4. Conclusion From the experimental data of conveying the Indian fly ash through four different pipelines, Fri = 7 has been found to be a reliable criteria to define safe minimum transport boundary. The use of Froude number seems to address the issue of different conveying velocity requirements for different pipe diameters and hence, a Froude number based approach appears to be more stable than specifying minimum conveying velocity. Several existing models for minimum transport criteria have been evaluated by comparing the experimental blockage boundary with those obtained through model predictions that are found to generally provide some under or over-predictions. The fly ash sample has been found to be a good candidate for fluidized densephase type conveying. Comparing the straight-pipe conveying characteristics obtained using two sets of tapping points located at different sections of the pipeline that has provided pressure drop characteristics (for different locations) reveals a possible change in flow mechanism along the direction of flow. Two different methods have been applied to model solid friction factor. One approach is based on introducing a dimensionless gas to particle density ratio term to account for the expansion of gases along the flow. The other method considered a volumetric loading ratio term that addresses partial filling of the pipe's cross section by the solid during actual conveying conditions. The models so developed were evaluated for scale-up accuracy by using them to predict the total pipeline pressure drops. The volumetric loading ratio approach of modelling has provided reliable predictions. This also indicates that the volumetric loading ratio based modelling approach will have greater potential for reliable prediction for over 100 m length pipes than the conventional Froude number based models. Future scope of work includes further validation of these models for longer pipes (100 m or more). This paper provides a test design procedure addressing important issues such as prediction of blockage boundary, dense-to-dilute transition criteria and solid friction factor. The product conveyed was a fly ash sample obtained from a specific thermal power plant. Whilst, strictly speaking, all fly ashes (and bulk solids) are different from each other based on particle and bulk densities, size distribution, boiler/furnace condition, etc. (and hence, truly speaking, all experimental research works in areas of bulk solid handling provide information/models specific to the tested product only), but these results/models still form in important guideline to the practicing or design engineer to select the plant design parameters within certain acceptable limits of variations of particle and bulk properties. Hence, the results presented in this paper (e.g. Fr = 7 and solid friction factor models) can be expected to contribute to the design for pneumatic conveying of fly ash systems in
B bend loss factor coefficient of drag CD D internal diameter of pipe [m] particle diameter [m] ds median particle diameter [μm] d50 Fr = V/(gD)0.5 Froude number of flow Froude number at pipe inlet Fri g acceleration due to gravity [m/s2] K constant of power function L length of pipe or test section [m] mass flow rate of air [kg/s] mf mass flow rate of solids [kg/s] ms m* = ms/mf solid loading ratio VLR = {(ms/ρs)/(mf/ρf)} volumetric loading ratio N number of bends ΔP pressure drop through a straight horizontal pipe or pipe section [Pa] single particle terminal velocity, m/s Ut V superficial air/gas velocity [m/s] superficial air/gas velocity at the inlet of a pipe section [m/s] Vi free settling velocity of an isolated particle wf0 ρ density of air [kg/m3] particle density [kg/m3] ρs loose-poured bulk density [kg/m3] ρbl air/gas only friction factor λf solid friction factor through straight pipe λs
Subscripts b bend bl bulk f fluid (air) i inlet condition s solids
Abbreviations BD bottom discharge I.D. internal diameter of pipe PCC pneumatic conveying characteristics
Acknowledgement The authors would like to thank the Department of Science and Technology, Science and Engineering Research Board, Ministry of Science and Technology (Government of India) for the financial assistance provided under the Young Scientist Scheme No: SR/FTP/ETA-15/2011. References [1] S.S. Mallick, PhD Dissertation: Modelling Dense-phase Pneumatic Conveying of Powders, University of Wollongong, 2010. [2] R. Pan, P.W. Wypych, Scale up procedures for pneumatic conveying design, Powder Handl. Process. 4 (2) (1992) 167–172. [3] D. Mills, Pneumatic Conveying Design Guide, 2nd ed. Elsevier/ButterworthHeinemann, 2004. [4] W. Barth, Strömungsvorgänge beim transport von festteilchen und flüssigkeitsteilchen in gasen, Chem. – Ing. – Tech. 30 (3) (1958) 171–180.
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45 [5] W. Stegmaier, Zur berechnung der horinentalen pneumatischen forderung feinkorniger feststoffe - for the calculation of horizontal pneumatic conveying of fine grained solids, Fordern Heben 28 (1978) 363–366. [6] M. Weber, Principles of hydraulic and pneumatic conveying on pipes, Bulk Solids Handl. 1 (1) (1981) 57–63. [7] F. Rizk, Pneumatic transport in dilute and dense phase, Bulk Solids Handl. 2 (2) (1982) 235–241. [8] R. Pan, P.W. Wypych, Dilute and dense phase pneumatic conveying of fly ash, The Proceedings of 6th International Conference on Bulk Materials Storage and Transportation, Wollongong, NSW, Australia, 1998, pp. 183–189. [9] M.G. Jones, K.C. Williams, Solids friction factors for fluidized dense phase conveying, Part. Sci. Technol. 21 (2003) 45–56. [10] A.J. Chambers, R.D. Marcus, Pneumatic conveying calculations, The Proceedings of 2nd International Conference on Bulk Materials Storage and Transportation, Wollongong, Australia, July 7–9 1986, pp. 49–52. [11] S. Keys, A.J. Chambers, Scaling pneumatic conveying characteristics for pipeline pressure drop, The Proceedings of National Conference on Bulk Materials Handling, Preprints, Capricorn International Resort, Yeppoon, Queensland, September 22–25 1993, pp. 205–213.
45
[12] S.E. Martinussen, PhD Dissertation: The Influence of the Physical Characteristics of Particulate materials on their Conveyability in Pneumatic Transport Systems, Technological Department, Telemark University College, Porsgruun, Norway, 1996. [13] S.S. Mallick, P.W. Wypych, Minimum transport boundaries for pneumatic conveying of powders, Powder Technol. 194 (2009) 181–186. [14] Aimo Rautiainen, Graeme Stewart, Visa Poikolainen, Pertti Sarkoma, An experimental study of vertical pneumatic conveying, Powder Technol. 104 (1999) 139–150. [15] A.M. Ahmed, S. Elghobashi, On the mechanisms of modifying the structure of turbulent homogeneous shear flows by dispersed particles, Phys. Fluids 12 (2000) 2906–2920. [16] N. Huber, M. Sommerfeld, Modelling and numerical calculation of dilute phase pneumatic conveying in pipe systems, Powder Technol. 99 (1998) 90–101. [17] G.E. Klinzing, F. Rizk, R.D. Marcus, L.S. Leung, Pneumatic Conveying of Solids — A Theoretical and Practical Approach, Publ. Springer, 2010.
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Modelling fluidized dense-phase pneumatic conveying of fly ash G. Setia ⁎, S.S. Mallick Department of Mechanical Engineering, Thapar University, Patiala, Punjab 147004, India
a r t i c l e
i n f o
Article history: Received 8 August 2014 Received in revised form 16 September 2014 Accepted 16 September 2014 Available online 23 September 2014 Keywords: Fluidized dense-phase Conveying characteristics Minimum transport boundary Solid friction factor Volumetric loading ratio
a b s t r a c t This paper presents the results of an ongoing investigation into modelling important design criteria, such as minimum transport condition, straight-pipe pressure drop and solid friction factor for fluidized dense-phase pneumatic conveying of powders. Fly ash (median particle diameter: 19 μm; particle density: 1950 kg/m3; loose-poured bulk density: 950 kg/m3) was conveyed over a wide range of flow conditions (from fluidized dense- to dilute-phase) under different conditions of pipeline diameters and lengths (viz. 43 mm I.D × 24 m length, 54 mm I.D × 24 m length, 69 mm I.D × 24 m length and 69 mm I.D × 70 m length). To define the safe minimum transport boundary, a Froude number based criteria at the pipe inlet has been used (Fri = 7). The Froude number based criterion is aimed to address the requirement of different conveying velocities for different pipe diameters. Straight-pipe pneumatic conveying characteristics obtained from two sets of pressure tapings installed at different locations of pipeline have shown that the trends and relative magnitudes of the pressure drops can be significantly different depending on the location of pressure tapping points, thus indicating a change in flow mechanism along the direction of flow. A new approach of modelling solid friction factor using a volumetric loading ratio term has provided better scale-up accuracy when the model predictions were compared with experimental data. This method of modelling solid friction is aimed to address the partial filling of pipe's cross section by the dune of solids, which appears to be a better representation of the flow conditions, especially for the dense-phase pneumatic conveying of fine powders. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Pneumatic conveying of fly ash is widely used in pulverized coal fired thermal power plants all over the world. Such dry handling and conveying technology make it possible for the fly ash to be subsequently used for other purposes, such as fly ash in dry form can be mixed with cement. Other reasons include: pneumatic conveying is totally enclosed, environmentally friendly, provides increased workplace safety; relatively low capital/maintenance costs (for a well-designed system); layout flexibility; and ease of automation and installation. For countries such as India, where about 60% of total installed capacity of electricity generation is through coal fired power generation route using coals having ash content of as high as 40–50% (for worst coals), a large amount of fly ash has to be pneumatically conveyed everyday from Electro Static Precipitator (ESP) hoppers to remote silos with a total pipeline length of even up to 1 km (depending on the plant layout). This is achieved by either using a combination of negative and positive pressure systems (i.e. by evacuating the ash from individual ESP hoppers to a local surge hopper using a vacuum conveying system and then using a positive pressure conveying system to transport fly ash from the surge hopper to remote silos) or by positive pressure systems only (where blow tanks are placed under each ESP hopper, which directly transfer the ⁎ Corresponding author. Tel.: +91 175 2393370. E-mail address: [email protected] (G. Setia).
http://dx.doi.org/10.1016/j.powtec.2014.09.033 0032-5910/© 2014 Elsevier B.V. All rights reserved.
fly ash to remote silo using compressed air as the transport medium). The positive pressure systems can be of two types: dilute-phase (or suspension flow) and dense-phase (or non-suspension flow). In dilutephase conveying, the carrier gas velocity is sufficiently high to suspend all the particles for conveying. Due to the dispersed and suspended nature of the dilute-phase conveying, the designers and researchers have achieved relatively good success in modelling the relevant particle interactions and mechanisms to develop numerous friction factor and pressure drop models [1]. However, this mode of conveying results in higher gas flows (and reduced solid loading ratio), larger sized compressor (i.e. higher energy consumption), larger pipe, fitting and support dimensions, increased chance of wear of pipe and bends (i.e. increased maintenance requirement) and larger sizing of bag filters. Due to these disadvantages of dilute-phase systems, the plant owners and designers are showing increasing interest in recent times to employ dense-phase pneumatic conveying systems (instead of the conventional dilute-phase mode) to take advantage of the fluidized dune flow capabilities of fly ash under aerated condition. For about 800– 1000 m long pipe (e.g. from intermediate surge hopper to remote silo in a power plant), a well designed fly ash dense-phase pneumatic conveying system would typically operate with a solid loading ratio of 30 (instead of 10–15 for a dilute-phase system) and with an initial conveying velocity 5–7 m/s (instead of about 15 m/s for dilute-phase). This would result in a substantial amount of savings in energy and initial capital costs in terms of smaller compressor and pipe sizes and maintenance
40
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
(reduced frequency of wearing of bends). In spite of such merits and potential to be a better alternative of fly ash conveying, wide spread installation of such system is still limited as designing a reliable fluidized dense-phase system is a difficult task due to the highly turbulent and complex nature of the concentrated dune flow [1,2]. For a reliable design of industrial scale pneumatic conveying systems, the two most important parameters that are to be accurately estimated are: (i) total pipeline pressure drop and (ii) minimum transport criteria. Over-estimation of total pipeline pressure drop would result in unnecessarily higher supply of air flow, energy cost and wearing of pipeline due to excessive conveying velocity. Under-prediction of total pipeline pressure drop would cause reduced material throughput rate. Incorrect estimation of minimum transport condition will result in particle deposition in pipes, leading to pipe blockage [1,3]. In fact, a majority of the practical installations facing unreliable operations and requiring troubleshooting suffer from problems arising out of the above two reasons. Hence, these design parameters (pipeline pressure drop and blockage criteria) must be accurately modelled and scaled-up with high reliability. The pressure drop for the flow of solid and gas flow through a straight horizontal pipe section can be represented using Eq. (1). This equation was originally presented by Barth [4] and believed to be for coarse particles in dilutephase flow. However, various other researchers, such as Pan and Wypych [2], Stegmaier [5], Weber [6], Rizk [7], Pan and Wypych [8] and Jones and Williams [9] have subsequently used this equation to predict the pressure loss for the fluidized dense-phase pneumatic transport of powders, such as fly ash, for horizontal straight pipes. 2 ΔP ¼ λ f þ m λs ðL=DÞρ V =2
ð1Þ
The main task in Eq. (1) is to model the solid friction factor accurately. Whereas more fundamental modelling and scale-up procedures based on powder mechanics have been established for lowvelocity dense-phase slug-flow of granular materials, modelling solid friction for fluidized dense-phase conveying of fine powders has remained a relatively more difficult challenge due to the complex and turbulent nature of the moving bed [1], where it is very difficult to link the particle and bulk properties to the flow mechanisms during actual conveying conditions. Due to these difficulties in modelling, empirical power function based models have been employed by various previous researchers [4–11] over the years to avoid the need of developing fundamental relationships between solid friction factor and the relevant particle and bulk properties. These models have used different dimensionless parameters and have shown only limited success under scale-up evaluation [1]. Hence, there is a requirement to conduct further studies on important design parameters such as straight pipe pressure drop models for solid friction and minimum transport criteria with an aim to provide the industry a reliable scale-up procedures for dense-phase pneumatic conveying of fly ash. 2. Experimental data Conveying trails were performed using the Indian fly ash with different pipeline configurations. Table 1 lists the physical properties of the fly ash. A typical schematic of the test set up used for fly ash conveying is shown in Fig. 1. Kirloskar made electric-powered Model KES 18-7.5 rotary screw compressor was used having the capacity of 3.37 m3/min Table 1 Physical properties of fly ash conveyed. Powder
Median particle diameter, d50 (μm)
Particle density, ρs (kg/m3)
Bulk density, ρbl (kg/m3)
Fly ash
19
1950
950
Fig. 1. Layout of the 54 mm I.D. × 70 m test rig.
of free air delivery and maximum delivery pressure of 750 kPa. Air flow control valve was installed in the compressed air line upstream of the blow tank to vary the conveying air flow rates over a wide range of air flows. For the measurement of air flow rates, a vortex flow metre was installed in the compressed air line. Bottom discharge type blow-tank was used as feeder having 0.2 m3 capacity of water fill volume. It was mounted with solenoid operated dome type material inlet, outlet and vent valves. The blow tank was provided with air supply system with the help of orifice plate at the air inlet to blow tank. A receiver bin of 0.65 m3 capacity was installed on top of blow tank. The receiver bin was fitted with bag filters having pulse jet type cleaning mechanism. The blow tank and receiver bin were supported by shear beam type load cells. Different mild steel pipelines of 43 mm I.D × 24 m length, 54 mm I.D × 24 m length, 69 mm I.D × 24 m length and 69 mm I.D × 70 m length were used as the test rigs. The test loops included a 3 m vertical lift and 4 × 90o bends having 1 m radius of curvature. Various static pressure measurement points were installed along the pipeline and bends, where P5 was used to measure the total pipeline pressure drop and all other transducers (i.e. P6 to P9) were installed to measure static pressures at the respective points, which were used to model solid friction factor of straight horizontal pipe. P6 and P7 were to measure static pressure in the initial part of the test loop, whereas P8 and P9 were used to measure static pressure for the later part of pipeline. All the static pressure transducers in the solid– gas flow line were strategically placed sufficient distance away from the influence of any change in flow pattern caused by the bends. Specification of static pressure transducers: manufacturer: Endress & Hauser, model: Cerabar PMC131, pressure range: 0–2 bar, maximum pressure: 3.5 bar (absolute), current signal: 4–20 mA. All other required instruments such as PRV (pressure reducing valve), flow metre, NRV (non-return valve), blow valve, pressure gauge and load cells (shear beam type) were suitably placed. Calibration of the pressure transducer, load cells and flow metre was performed using a standardized calibration procedure [1]. To record the electrical output signals from the load cells, pressure transducers and flow metre, a portable PC compatible data logger was used. The data logger had 16 different channels with 14 bit resolution. Every pipeline was installed with two sets of 300 mm long sight-glasses made of borosilicate materials for flow visualization (using high speed digital camera).
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
A unique data-base of total pipeline, straight pipe pressure loss and minimum conveying condition have been generated by conveying for the Indian fly ash for a wide range of flow conditions through different pipelines. Observation through the sight glass showed significant amount of non-suspension flow (a dense highly turbulent non-suspension layer of powders flowing along the bottom of pipe with particles travelling on top in suspension flow), thus confirming fluidized densephase convey ability of the Indian fly ash. At higher air flows, the fly ash was conveyed in dilute-phase. With gradual reduction of air flow rate, continuous dune flow was observed. Further reduction of air velocity resulted in pulse-type dunning or discontinuous dunes. Further reduction of air flows provided high pressure fluctuations (i.e. unstable conveying) and ultimately complete pipe blockage. The unstable conveying regime is characterized by a gradual build up of material in the pipeline, where the amount of material which is pushed in the line, does not completely return to the receiver bin (which was found from the records of load cells installed under blow tank and receiver bin). In the present study, this unstable-phase conveying is considered as the initiation of blockage condition as repeated attempts of conveying with such a gradual product build-up condition would completely block the pipeline in few cycles of operation in industry. It was found that experimentally it would be difficult to be very precise about the air flow rate for which line blockage would occur due to the practical limitation of setting the air flow control valve exactly for the blockage condition. Therefore a series of tests were carried out near the blockage boundary to define a zone of air flow rates for which line blockage would take place. The solid flow rates were varied using different prepressurizations of the blow tank aeration. Certain tests were performed multiple times to ensure repeatability of test data. 2.1. Pneumatic conveying characteristics and minimum transport criteria Pneumatic conveying characteristics (PCC) for a given product are the representation of variation of pressure drop for different solid and air flow rates (e.g. dense- to dilute-phase). PCC constitute an essential requirement for the reliable design of pneumatic transportation systems by providing valuable information, such as the expected mode of flow, minimum conveying air velocity requirement to prevent blockage and the optimal operating conditions. Total pipeline PCC for fly ash is based on the steady-state pressure drop data obtained from the pressure transducer P5 (Fig. 1) and are shown in Figs. 2 to 5 for the different pipelines. Experimental blockage boundaries are shown in the PCC. Different predictions of blockage boundary (or the minimum transport line) using different models for minimum transport conditions have been superimposed on the above figures. The lines joining the initiation of unstable zone have been designated as blockage boundary or minimum transport condition in this paper. All the PCC (Figs. 2 to 5) show
Fig. 2. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 54 mm I.D. × 70 m long pipe.
41
Fig. 3. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 54 mm I.D. × 24 m long pipe.
that pressure drop rises slowly with an increase in air flow rate up to about 0.025 kg/s of air flow. Pressure drops sharply increase with further rise in air flow rates. However, total pipeline PCC are not conclusive about the pressure drop characteristics (change in pressure drop with increase in air flow rate) for straight pipes, as the total pipeline PCC are formed by the superimposition of pressure drop characteristics of the bends to that of straight pipes. Hence, a relatively large contribution of the bend loss characteristics could have considerably influenced the overall trends of the total pipeline PCC. The PCC (Figs. 2 to 5) do not show any clearly defined Pressure Minimum Curve (PMC), which indicates flow transition (from dense- to dilute-phase flow) [1]. For the whole range of test rigs, the Froude number at the feed point (bottom of blow tank) varied from 3.5 to 6.4. The following contains a list of existing models to predict the blockage boundary. Weber [6] provided the following correlation to predict the minimum conveying velocity to avoid line blockage. Weber [6], however, did not comment on the reliability and applicability of the model (i.e. he did not provide details of pipe and product on which the model should be applied): For U t ≤3m=s 0:25 0:1 Fri ¼ ½7 þ ð8=3ÞUt m ððds =DÞÞ
ð2Þ
For U t ≥3m=s 0:25 0:1 Fri ¼ 15 m ðds =DÞ :
ð3Þ
Fig. 4. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 43 mm I.D. × 24 m long pipe.
42
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
of Mallick and Wypych [13] was based mostly on the Australian fly ash data, which had a narrower size distribution (more finer) than the Indian fly ash having a wider range of size distribution having a considerable share of coarser particles. 2.2. Pressure drop characteristics for straight pipe sections
Fig. 5. Pneumatic conveying characteristics and evaluation of minimum transport boundary, fly ash, 69 mm I.D. × 24 m long pipe.
Martinussen [12] used test facility having horizontal pipeline of 53 mm I.D and 15 m length. By applying the fluid analogy, he developed the following model to determine the minimum conveying velocity, as given by Eq. (4). According to Martinussen [12], this model could provide better predictions for fine materials than for the coarser ones. 3 2 Vi ¼ KDgðρbl =ρ f Þ 1−m ðρ f =ρbl Þ
ð4Þ
where, K (geometrical factor) = П/4 at the filling level of D/2. Mills [3] conveyed cement through a 52.5 mm I.D. × 95 m test rig and proposed a constant velocity based criteria to represent the minimum conveying conditions in which he suggested the inlet velocity equal to 3 m/s as the minimum transport criteria for fine powders. He also proposed a power function type relationship between minimum inlet velocity and solid loading ratio (i.e. lower minimum conveying velocity is required for larger solid loading ratio). Mallick and Wypych [13] investigated into minimum conveying criteria using the experimental data from twelve different powders conveyed over a wide range of pipe lengths and diameters. It was found that the representation of minimum conveying velocity just as a function of solid loading ratio may not be adequate for reliable predictions under diameter scale-up conditions. Based on the experimental data of ESP dust and fly ash, they found that with increase in pipe diameter, the requirement of minimum conveying air velocity rises. To incorporate the effect of pipe diameter on minimum conveying velocity, they proposed a Froude number based criteria at the feed point (Fri = 6) as the minimum transport condition for powders having a loose-poured bulk density of ρbl ≤ 1000 kg/m3 and that are is suitable for fluidized dense-phase conveying. The above models (Weber [6] to Mallick and Wypych [13]) have been validated for their scale-up accuracy by using them to predict the blockage boundary for the given Indian fly ash. The results are shown in Figs. 2 to 5. Results show that the experimental minimum transport boundary vary in the range of Fri = 5 to 7. Mills' criteria of Vi = 3 consistently provided under-prediction for all cases as the actual conveying velocity requirements are higher than this. The Martinussen [12] model provided over-prediction for all the pipelines. In fact, this model resulted in over-predictions that were found to be out of the range of the PCC in Figs. 2 and 5, hence, predictions using this model have not been shown in these figures. Similarly, the Weber [6] model has also provided over-prediction in all cases, although the margin of overprediction is less than that predicted by Martinussen [12]. The Mallick and Wypych [13] model (Fri = 6) seems to be closest to the experimental observation. Considering the actual minimum transport boundary for the fly ash varied from Fri = 5–7, the safe criteria to avoid line blockage with the Indian fly ash is proposed here as Fri = 7. This criteria of minimum transport condition requires slightly higher velocities than that proposed earlier by Mallick and Wypych [13], as the Fri = 6 criteria
Using the pressure drop data between tapping points installed on straight pipe lengths P6–P7 and P8–P9 on the 54 mm I.D. × 70 m long pipeline, pneumatic conveying characteristics for straight pipes have been developed for different solid and air flow rates. These are shown in Fig. 6. It can be seen from Fig. 6 that depending on the location of tapping points, pressure drop characteristics get different. The pressure drop per unit length is more at the beginning of pipe compared to that at the end. This indicates a change in flow mechanism along the length of pipe. This could be due to the mixing of air with solid (fly ash) along the flow, which would cause the material dune to get aerated and become more fluidic, hence easier to transport. It is found that below the air flow rate of 0.02 kg/s, the pressure drop values almost do not change with increase in air flow. However with further rise in air flow rates, the pressure drop values have increased rather sharply with an increase in air flow rate, thus indicating a gradual change from non-suspension to suspension mode of flow. Experimentally it was found (view of flow through sight-glass) that the transition from non-suspension to suspension mode was rather gradual with no clear location of pressure minimum curve on the straightpipe PCC. 3. Modelling and validation of solid friction factor models In order to model solid friction factor for straight pipes, the following two formats have been used in the present study: a b c Model 1 : λs ¼ K m ð FrÞ ðρ f =ρs Þ
ð5Þ
and, a
b
Model 2 : λs ¼ KðVLRÞ ðw f0 =VÞ
ð6Þ
where, volumetric loading ratio (VLR) is given as: VLR ¼ fðms =ρs Þ=ðm f =ρ f Þg:
ð7Þ
Several previous researchers have used m* and Fr in power function format to model solid friction factor, such as Mallick [1] and Pan and Wypych [8]. However, these models do not take care of the effect
Fig. 6. Comparison between straight-pipe PCC, fly ash, 54 mm I.D. × 70 m long pipe.
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
of reduction of gas density along the direction of flow. Hence, to account for the same, a dimensionless air to particle density ratio term has been introduced in the format of modelling solid friction factor (Model 1). A different approach to modelling has been provided by Model 2, which uses a volumetric loading ratio term. In this approach, it is considered that the solid loading ratio term (which is a ratio of the masses) does not properly account for important flow criteria, such as the pipe volume fraction that is actually available for the flow of air, as a considerable portion of the pipe section gets filled by products in dense-phase. Hence, the effective area for air flow is considerably reduced. To account for this partial filling up of pipe, the volumetric loading ratio term is proposed here (Eq. (7)). Rautiainen et al. [14] mentioned that solid volume fraction is an important contributor to the pressure drop of solid–gas flows as the occupied volume of solid is expected to influence the available gas path, which would contribute to additional energy drop. Ahmed and Elgobashi [15] proposed that the solid–gas flows will introduce two-way coupling and influence the turbulence and pressure drop of flow. Huber and Sommerfield [16] suggested that small particles would follow the eddies of the gas-phase and as a result, there would not be much relative velocities amongst the particles within an eddy. This would reduce the inter-particle collision effect. The above suggests that the amount of pipe volume occupied by the solids is an important feature towards representing the flow mechanism of dense-phase solid–gas flows. Therefore, the volumetric loading ratio term has been introduced in the present study to model solid friction factor. The wf0/V term attempts to indicate dense- to dilute-phase transition phenomenon. Larger sized and/or higher particle density products will have increased free settling velocity. It can be thought intuitively that particles having larger value of free settling velocity (i.e. higher tendency of settling down) will require higher velocity to keep them in suspension. Using steady-state pressure drop data between P6 and P7 tapping points over a range of solid and air flow rates, the following models are devolved using both the formats using regression analysis in Microsoft Excel: h i −0:696 −1:392 −0:458 2 Model 1 format : λs ¼ 169:96 m ð FrÞ ðρs =ρ f Þ R ¼ 0:985
ð8Þ
−0:316
Model 2 format : λs ¼ 0:643ðVLRÞ
0:777
ðw fo =VÞ
h i 2 R ¼ 0:930 : ð9Þ
Figs. 7 and 8 show plots of y = λs (Fr)a(ρs/ρf)b versus x = m* and y = λs (wfo/V)c versus x = VLR through the experimental data for Model 1 and Model 2, respectively, where ‘a’, ‘b’ and ‘c’ are the adjustable exponents. R2 values from the trends are also provided in figures. High values of R2 were obtained using the method of least squares to develop the model that indicates good agreement of straight pipe data
Fig. 7. Development of Model 1 — trends of y = λs (Fr)a(ρs/ρf)b versus x = m*.
43
Fig. 8. Development of Model 2 — trends of y = λs (wfo/V)c versus x = VLR.
sets with the model estimates. Terminal settling velocity of solids has been calculated as 0.0235 m/s using the following expression (Eq. (10)) and using appropriate drag coefficient as per [17]: w fo ¼
4gdðρs −ρ f Þ 0:5 : 3CD ρ f
ð10Þ
The above solid friction factor models have been evaluated by using them to predict the total pipeline drop for the two pipelines: 54 mm I.D. × 70 m long and 69 mm I.D. × 24 m for different solid and air flow rates and by comparing the predicted pneumatic conveying characteristics against the experimental plots. To predict the losses in bends, Chambers and Marcus [10] model was used (Eq. (11)). The results are shown in Figs. 9 and 10. 2 ΔPb ¼ NB 1 þ m ρV =2
ð11Þ
Fig. 9 shows that Model 1 provides under-prediction for 3.5 t/h of solid flow rate, whereas it provides over-prediction for the 1.5 t/h line. The model somewhat over-predicts in the lower velocity range and begins to under-predict with rise in air velocity. Model 2 provides good predictions for the higher tonnage (3.5 t/h), but somewhat overpredicts for the lower tonnages. Fig. 10 shows that the Model 1 format provides considerable amount of under-predictions for diameter scale-up conditions for all tonnages, whereas Model 2 provides good predictions for the 5.5 t/h and better predictions than Model 1 for the lower tonnages (the level of underpredictions have reduced with Model 2). The estimated pressure drop values are found to be increasing with increase in air flow rates. This matches the experimental trends for air flow rates above 0.045 kg/s. Below this air flow rate, the experimental trends show a change (decrease) in slope (in highly dense-phase region that is closer to the
Fig. 9. Experimental versus predicted PCC, fly ash, 54 mm I.D. × 70 m long pipe.
44
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45
plants that handle fly ash having wide size distribution (material characteristics similar to the conveyed fly ash of this paper). Further scope of work will include incorporation of specific material characteristics to model minimum transport criteria and solid friction factor. List of symbols and abbreviations
Fig. 10. Experimental versus predicted PCC, fly ash, 69 mm I.D. × 24 m pipe.
blockage boundary), whereas, the predicted trends (model predictions) appear to be monotonously linear. Therefore, further work is required for better predictability of solid friction in very dense-phase region. 4. Conclusion From the experimental data of conveying the Indian fly ash through four different pipelines, Fri = 7 has been found to be a reliable criteria to define safe minimum transport boundary. The use of Froude number seems to address the issue of different conveying velocity requirements for different pipe diameters and hence, a Froude number based approach appears to be more stable than specifying minimum conveying velocity. Several existing models for minimum transport criteria have been evaluated by comparing the experimental blockage boundary with those obtained through model predictions that are found to generally provide some under or over-predictions. The fly ash sample has been found to be a good candidate for fluidized densephase type conveying. Comparing the straight-pipe conveying characteristics obtained using two sets of tapping points located at different sections of the pipeline that has provided pressure drop characteristics (for different locations) reveals a possible change in flow mechanism along the direction of flow. Two different methods have been applied to model solid friction factor. One approach is based on introducing a dimensionless gas to particle density ratio term to account for the expansion of gases along the flow. The other method considered a volumetric loading ratio term that addresses partial filling of the pipe's cross section by the solid during actual conveying conditions. The models so developed were evaluated for scale-up accuracy by using them to predict the total pipeline pressure drops. The volumetric loading ratio approach of modelling has provided reliable predictions. This also indicates that the volumetric loading ratio based modelling approach will have greater potential for reliable prediction for over 100 m length pipes than the conventional Froude number based models. Future scope of work includes further validation of these models for longer pipes (100 m or more). This paper provides a test design procedure addressing important issues such as prediction of blockage boundary, dense-to-dilute transition criteria and solid friction factor. The product conveyed was a fly ash sample obtained from a specific thermal power plant. Whilst, strictly speaking, all fly ashes (and bulk solids) are different from each other based on particle and bulk densities, size distribution, boiler/furnace condition, etc. (and hence, truly speaking, all experimental research works in areas of bulk solid handling provide information/models specific to the tested product only), but these results/models still form in important guideline to the practicing or design engineer to select the plant design parameters within certain acceptable limits of variations of particle and bulk properties. Hence, the results presented in this paper (e.g. Fr = 7 and solid friction factor models) can be expected to contribute to the design for pneumatic conveying of fly ash systems in
B bend loss factor coefficient of drag CD D internal diameter of pipe [m] particle diameter [m] ds median particle diameter [μm] d50 Fr = V/(gD)0.5 Froude number of flow Froude number at pipe inlet Fri g acceleration due to gravity [m/s2] K constant of power function L length of pipe or test section [m] mass flow rate of air [kg/s] mf mass flow rate of solids [kg/s] ms m* = ms/mf solid loading ratio VLR = {(ms/ρs)/(mf/ρf)} volumetric loading ratio N number of bends ΔP pressure drop through a straight horizontal pipe or pipe section [Pa] single particle terminal velocity, m/s Ut V superficial air/gas velocity [m/s] superficial air/gas velocity at the inlet of a pipe section [m/s] Vi free settling velocity of an isolated particle wf0 ρ density of air [kg/m3] particle density [kg/m3] ρs loose-poured bulk density [kg/m3] ρbl air/gas only friction factor λf solid friction factor through straight pipe λs
Subscripts b bend bl bulk f fluid (air) i inlet condition s solids
Abbreviations BD bottom discharge I.D. internal diameter of pipe PCC pneumatic conveying characteristics
Acknowledgement The authors would like to thank the Department of Science and Technology, Science and Engineering Research Board, Ministry of Science and Technology (Government of India) for the financial assistance provided under the Young Scientist Scheme No: SR/FTP/ETA-15/2011. References [1] S.S. Mallick, PhD Dissertation: Modelling Dense-phase Pneumatic Conveying of Powders, University of Wollongong, 2010. [2] R. Pan, P.W. Wypych, Scale up procedures for pneumatic conveying design, Powder Handl. Process. 4 (2) (1992) 167–172. [3] D. Mills, Pneumatic Conveying Design Guide, 2nd ed. Elsevier/ButterworthHeinemann, 2004. [4] W. Barth, Strömungsvorgänge beim transport von festteilchen und flüssigkeitsteilchen in gasen, Chem. – Ing. – Tech. 30 (3) (1958) 171–180.
G. Setia, S.S. Mallick / Powder Technology 270 (2015) 39–45 [5] W. Stegmaier, Zur berechnung der horinentalen pneumatischen forderung feinkorniger feststoffe - for the calculation of horizontal pneumatic conveying of fine grained solids, Fordern Heben 28 (1978) 363–366. [6] M. Weber, Principles of hydraulic and pneumatic conveying on pipes, Bulk Solids Handl. 1 (1) (1981) 57–63. [7] F. Rizk, Pneumatic transport in dilute and dense phase, Bulk Solids Handl. 2 (2) (1982) 235–241. [8] R. Pan, P.W. Wypych, Dilute and dense phase pneumatic conveying of fly ash, The Proceedings of 6th International Conference on Bulk Materials Storage and Transportation, Wollongong, NSW, Australia, 1998, pp. 183–189. [9] M.G. Jones, K.C. Williams, Solids friction factors for fluidized dense phase conveying, Part. Sci. Technol. 21 (2003) 45–56. [10] A.J. Chambers, R.D. Marcus, Pneumatic conveying calculations, The Proceedings of 2nd International Conference on Bulk Materials Storage and Transportation, Wollongong, Australia, July 7–9 1986, pp. 49–52. [11] S. Keys, A.J. Chambers, Scaling pneumatic conveying characteristics for pipeline pressure drop, The Proceedings of National Conference on Bulk Materials Handling, Preprints, Capricorn International Resort, Yeppoon, Queensland, September 22–25 1993, pp. 205–213.
45
[12] S.E. Martinussen, PhD Dissertation: The Influence of the Physical Characteristics of Particulate materials on their Conveyability in Pneumatic Transport Systems, Technological Department, Telemark University College, Porsgruun, Norway, 1996. [13] S.S. Mallick, P.W. Wypych, Minimum transport boundaries for pneumatic conveying of powders, Powder Technol. 194 (2009) 181–186. [14] Aimo Rautiainen, Graeme Stewart, Visa Poikolainen, Pertti Sarkoma, An experimental study of vertical pneumatic conveying, Powder Technol. 104 (1999) 139–150. [15] A.M. Ahmed, S. Elghobashi, On the mechanisms of modifying the structure of turbulent homogeneous shear flows by dispersed particles, Phys. Fluids 12 (2000) 2906–2920. [16] N. Huber, M. Sommerfeld, Modelling and numerical calculation of dilute phase pneumatic conveying in pipe systems, Powder Technol. 99 (1998) 90–101. [17] G.E. Klinzing, F. Rizk, R.D. Marcus, L.S. Leung, Pneumatic Conveying of Solids — A Theoretical and Practical Approach, Publ. Springer, 2010.