The performance of a fluidized particulate feeding system with multiple discharge sections

Abstract

A particulate feeding system based on air flluidization principles was developed at the Thermal and Fluids Engineering Group at EESC-USP. The system is basically formed of a cylindrical chamber comprising two levels of fluid bed distributor plates. The performance of the system was addressed. Repeatability experiments were performed in order to check out for the possibility of solids feeding control through a sphere valve which controls the air feed rate. Silica sand particles of three different size were used, with mean diameters of 152, 287 and 484 µm. Experiments showed that the total particulate discharge depends on the primary air feeding rate, the distance between the exit of the feeding hopper and the primary air distributor plate, and the particulate diameter, but independs on the secondary air feeding rate. The system was observed to be of difficult control and showed poor repeatability. However, the system was found to be an excellent device for distributing the flow of solids throughout several discharge sections. For the operational conditions considered and air volumetric flowrates above 3×10-3 m³/s the deviation on the particulate mass flowrate through a given discharge section related to the average value through the four sections of the injector was always below 5 %.

Fluidization; feeding system; particulate feeding; multipoint feeding; particulate injector

The performance of a fluidized particulate feeding system with multiple discharge sections

J. P. Tureso^I; F. E. Milioli^II; G. Lombardi^III

^IE-mail: [e-mail] ^IIE-mail: [e-mail] ^IIINúcleo de Engenharia Térmica e Fluidos - NETeF, Escola de Engenharia de São Carlos, Universidade de São Paulo, Av. Trabalhador São-carlense, 400 Centro, 13566-590 São Carlos, SP. Brazil. E-mail: [e-mail]

ABSTRACT

A particulate feeding system based on air flluidization principles was developed at the Thermal and Fluids Engineering Group at EESC-USP. The system is basically formed of a cylindrical chamber comprising two levels of fluid bed distributor plates. The performance of the system was addressed. Repeatability experiments were performed in order to check out for the possibility of solids feeding control through a sphere valve which controls the air feed rate. Silica sand particles of three different size were used, with mean diameters of 152, 287 and 484 µm. Experiments showed that the total particulate discharge depends on the primary air feeding rate, the distance between the exit of the feeding hopper and the primary air distributor plate, and the particulate diameter, but independs on the secondary air feeding rate. The system was observed to be of difficult control and showed poor repeatability. However, the system was found to be an excellent device for distributing the flow of solids throughout several discharge sections. For the operational conditions considered and air volumetric flowrates above 3×10^-3 m³/s the deviation on the particulate mass flowrate through a given discharge section related to the average value through the four sections of the injector was always below 5 %.

Keyworks: Fluidization; feeding system; particulate feeding; multipoint feeding; particulate injector

Introduction

A common issue in industry regards the transportation of huge amounts of solid particulates, which must be taken from deposits and conveyed for processing. A number of mechanical systems such as belt and screw conveyers and vibrating feeders are usually used to fulfil this task. Those systems present some disadvantages such as the wearing of moving parts, excessive size, limited velocities of operation, and requirement for periodical maintenance. Alternatives to mechanical systems are the pneumatic conveying systems. There is a growing use of such systems in industry owing to a more secure operation and lower costs. They are easier to operate when conveying high temperature and/or high pressure solids. Also, the pneumatic conveying systems allow a good control of the solids feeding rate, which is a factor of great importance in many applications. In general, the introduction of reactants in process reactors must be controlled. In fluidized bed reactors it is desirable to provide a feeding which is uniform and well distributed over the bed cross section. This provides a bed with uniform reaction efficiency all over its volume. In fluidized bed coal combustors using limestone for sulphur absorption a non-uniform feeding may cause the reactive particles to concentrate, and thereby considerable loss by elutriation and unburned CO and hydrocarbons (Borges, 1987).

The particulate injection system studied in this work was designed for injecting particulate into pneumatic conveying lines. The device presents four particulate discharge sections, providing good uniformity when feeding process reactors such as fluidized bed combustors. The design of the multiple point particulate injector is based on a previous single point feeding system (Lombardi et al., 1990, Lombardi et al., 1994, Pécora, 1995). The basic multiple point system includes a primary level air distributor plate in form of tray over which a repose cone of particles fed from a hopper is formed. When fluidized, the cone is destroyed and the particles fall into four annular radially placed isolated compartments holding at the bottom secondary level distributor plates that provide solids discharge. The superficial velocities of the air flowing through the primary distributor plate are typical of the bubbling fluidized bed regime. The air feeding rate actually determines the total particulate feeding rate of the system. The solids feeding rate is, therefore, dependent on fluidization parameters. In this work the air flow that causes particulate discharge is expressed in terms of the minimum fluidizing velocity of the particulate, U_mf. The literature presents good correlations for this parameter (Geldart, 1986). The correlation of Wen and Yu (1966) is used in this work:

where:

The main objective of this work is to establish the performance of a multiple point fluidized particulate injector under different operational conditions, having in view the uniformity of particulate distribution among the discharge sections. The literature does not present previous works related to this type of particulate injection system.

Description of the Experimental Set Up

A schematic diagram of the experimental set up is shown in Figure 1. It is basically comprised of a particulate feeding hopper, a multiple point fluidized particulate injector, a particulate reposition hopper, and an exhaust system. A sketch of the particle injector is shown in Figure 2. It is a cylindrical chamber comprising an upper opening for receiving particulate, a primary level air distributor plate, four secondary level air distributor plates, and four discharge sections. The primary level plate is an unbounded air distributor plate in form of tray that supports the particulate repose cone as shown in Figure 3. The secondary level plates, which are also air distributor plates, are placed at the bottom of four isolated compartments that receive particulate from the primary level plate. Each one of such compartments is connected to a particulate discharge section. The discharge sections are connected to pneumatic conveying lines, which are connected to cyclones for particulate retrieval. The primary level plate presents 16 tower type injectors, while the secondary level plates present 32 bubblecap type injectors. The design procedure for those injectors is presented in Lombardi et al. (1997). The cylindrical chamber is axially movable allowing the distance between the exit of the feeding hopper and the primary air distributor plate to be changed. At the top of the reposition hopper there is a by pass plate which allows either the flow through the cyclones for particle sample collection, or delivers the flow to the hopper. The injector chamber is still equipped with three visors and two lamps to allow visualization. A pipe connects the injector chamber to the reposition hopper to provide pressure equalization, so that the discharge of solids is independent of pressure. Further details of the experimental set up can be found in Tureso, (1998).

Description of the Experiments

Two sets of experiments were carried out to establish the total particulate mass flowrate, the uniformity of particulate discharge distribution among the discharge sections, and the effect of the particulate granulometry over the process. A uniformity criterion is defined as the deviation on the particulate mass flowrate through a given discharge section related to the average value through the four sections of the injector. Another set of experiments was carried out to establish the effect over the process of the distance between the exit of the feeding hopper and the primary air distributor plate (h_c in Figure 3). The three sets of experiments were considered to evaluate the effect of the primary air superficial velocity over the process. Various runs in each set of experiments were carried out to determine repeatability. Table 1 shows the operational parameters considered. Silica sand was used in all experiments. Table 2 shows de size distribution of the particulates.

The particle Sauter mean diameter was determined as (Geldart, 1986):

where:

Results

As shown in Table 1 the first set of experiments comprising runs 1 to 6 was carried out for the 287 µm mean size particulate, for h_c = 15 mm. Figures 4 to 9 show the results. Figure 4 shows the total particulate mass flowrate as a function of the primary air volumetric flowrate. The standard deviation of the residuals between the experimental data and the fitted least squares line results 30,26 g/s. Assuming a normal distribution for the residuals and a confidence interval of 95 %, the deviations between the predictions and the fitting result between ± 14 % and ± 387 % (i.e. ± 1.96s), increasing as the air volumetric flowrate decreases. These large deviations reflect the poor repeatability of the system. Figures 5 to 8 show the particulate mass flowrate at each discharge section as a function of the primary air volumetric flowrate. The deviations between the experimental data and the least squares line fitting are of the same order as above.

Figure 9

The second set of experiments comprising runs 7 to 10 was carried out for the 484 µm mean size particulate, also for h_c = 15 mm. Figures 10 to 15 show the results. Figure 10 shows the total particulate mass flowrate as a function of the primary air volumetric flowrate. The standard deviation of the residuals between the experimental data and the fitted least squares line results 30,79 g/s. Assuming a normal distribution for the residuals and a confidence interval of 95 %, the deviations between the predictions and the fitting result between ± 16 % and ± 546 % (i.e. ± 1.96s), increasing as the air volumetric flowrate decreases. Again, these large deviations reflect the poor repeatability of the system. Figures 11 to 14 show the particulate mass flowrate at each discharge section as a function of the primary air volumetric flowrate. The deviations between the experimental data and the least squares line fitting are of the same order as above. Figure 15 shows the higher deviation on the particulate mass flowrate through a given discharge section related to the average value through the four sections. Maximum deviations between 0,2 % and 14,5 % (in modulus) are observed, increasing as the air volumetric flowrate decreases. For primary air volumetric flowrates above 3× 10^-3 m³/s the maximum deviation from the average results 2 %.

Effect of h_c

As shown in Table 1 the third set of experiments comprising runs 11 to 15 was carried out for the 152 µm mean size particulate, for the distance between the exit of the feeding hopper and the primary air distributor plate (h_c) changing from 7 mm to 15 mm. Figures 16 to 21 show the results. Figure 16 shows the total particulate mass flowrate as a function of both the primary air volumetric flowrate and h_c. The total particulate mass flowrate increases as h_c is increased, showing a linear dependence on the primary air volumetric flowrate for all h_c. The curve gradients result higher for higher h_c, indicating that the system becomes more sensitive to the primary air volumetric flowrate for growing h_c. Figures 17 to 20 show the particulate mass flowrate at each discharge section as a function of the primary air volumetric flowrate. The behaviour of the curves is similar to that for the total particulate mass flowrate. The uniformity of solids distribution among the discharge sections is shown in Figure 21 for various h_c. The maximum deviations on the particulate mass flowrate through a given discharge section related to the average value among the four sections result between 1.4 % and 27.6 % (in modulus), increasing as the air volumetric flowrate decreases. For primary air volumetric flowrates above 3× 10^-3 m³/s the maximum deviation from the average results 5 %. The deviations are smaller for higher h_c over the whole range of primary air volumetric flowrates, showing a more uniform feeding condition for larger particulate repose cones.

Effect of Granulometry

Figures 4, 10 and 16 show that for high primary air volumetric flowrate the total particulate mass flowrate is directly proportional to granulometry. For low volumetric air flowrates the solids feeding becomes quite unstable making it difficult a comparative evaluation between the different granulometries. It is seen that the finer the particulate the smaller the gradient of curve, indicating lower sensitivity regarding the primary air flowrate. As seen in Figures 9, 15 and 21, the higher deviations on the particulate mass flowrate through a given discharge section related to the average among the four sections occur at lower primary air flowrates, and it is lower for the coarser particulate. The distribution of particulate among the discharge sections result more uniform for the coarser particulate.

Effect of the Primary Air Superficial Velocity

The analysis were performed assuming a primary air superficial velocity established for a cross sectional area with the diameter of the primary air distributor plate, having in view the experimental results for h_c=15 mm. The experiments for the 152 µm, the 287 µm and the 484 µm particulates were performed for superficial velocities in the range 4.08U_mf to 37.44U_mf, 1.06U_mf to 10.49U_mf, and 0.41U_mf to 3.67U_mf respectively. In each case the lower limit concerns the minimum superficial velocity for what the repose cone was broken. For the first two cases the minimum superficial velocities required for breaking the repose cones were above the respective minimum fluidizing velocities. For the third case the repose cone was broken at superficial velocities below the minimum fluidizing velocity. For the 152 µm and the 287 µm particulates at low primary air volumetric flowrates there was not enough perturbation to break the repose cone even for superficial velocities above U_mf. It seems in this case that the possible fluidization is restricted to regions very close to the distributor plate. As the air stream goes away from the distributor plate it flows through the repose cone over a growing cross sectional areas, causing defluidization. Otherwise, for the 484 µm particulate the repose cone is enough disturbed to flow even at superficial velocities below U_mf. It seems in this case that the disturbances imposed by the air jets at the distributor plate propagate through the particulate causing the repose cone to break. Further results and discussion can be found in Tureso (1998).

Conclusions and Final Remarks

Literature does not present previous studies on particulate feeding systems such as the one studied in this work. The main parts of the system are the tray like unbounded primary air distributor plate that provides solids feeding, and the secondary air distributor plate which provides its distribution throughout four discharge sections.

For good uniformity of particulate distribution among the discharge sections the multiple point fluidized particulate injector must be operated with primary air volumetric flowrates above 3×10^-3 m³/s, for all granulometries considered. In such cases the deviation on the particulate mass flowrate through a given discharge section related to the average value through the four sections of the injector was always below 5 %.

At first, the operation at the above high air flowrates does not prevent the system of operating at lower particulate feeding rates providing additional discharge sections are added. Such a situation requires further experiment.

The experiments suggest that two possible mechanisms may account for the breaking of the repose cone and particulate feeding. For the finer particulate the process appears to be controlled by the fluidization of the particles at the bottom of the primary air distributor plate, while for the coarser particulate the action of air jets seems to control.

The multiple point fluidized particulate injector showed to be of difficult control, requiring improvements if the total particulate flowrate is to be accurately controlled. Otherwise, the system showed good behaviour regarding the uniformity of particulate distribution among the discharge sections.

Paper accepted September, 2002. Technical Editor: Aristeu da Silveira Neto

• Borges, M.G.S., 1987, "Dense Fluidized Bed Pneumatic Particulate Dosator-Injector," (in Portuguese), Master Dissertation, School of Engineering of São Carlos, University of São Paulo, São Carlos, SP.
• Geldart, D., 1986, "Single Particles, Fixed and Quiescent Beds," in Gas Fluidization Technology, ed. D. Geldart, John Wiley, New York.
• Lombardi, G., Pagliuso, J.D., Pécora, A.A.B., Goldstein Jr., L., 1994, "New Particulate Solids Pneumatic Feeding Device With Mass Flowrate Control," Powder Technology, Vol.79, pp.265-268.
• Lombardi, G., Pagliuso, J.D., Pécora, A.A.B., Goldstein Jr., L., 1990, "Development of a Particulate Solids Pneumatic Feeding Device," (in Portuguese), 3rd Brazilian Congress of Engineering and Thermal Sciences, Itapema, SC.
• Lombardi, G.; Pagliuso, J.D.; Goldstein Jr, L. (1997). Performance of a tuyere gas distributor. Powder Technology. n.94. pp.5-14.
• Pécora, A.A.B., 1995, "Fluid Dynamic Study of Circulating Fluidized Beds with Secondary Air Injection," (in Portuguese), Doctorate Theses, State University of Campinas, Campinas, SP.
• Tureso, J.P., 1998, "Study of a Fluidized Particulate Feeding System with Multiple Discharge Sections," (in Portuguese), Master Dissertation, School of Engineering of São Carlos, University of São Paulo, São Carlos, SP.
• Wen, C.Y., Yu, Y.H., 1966, "A Generalized Method for Predicting the Minimum Fluidization Velocity," AIChE Journal, Vol.12, n.3, pp.610-612.

Publication Dates

• Publication in this collection
  18 Mar 2004
• Date of issue
  Mar 2003

History

• Received
  Sept 2002