DESIGN AND EFFICIENY ASPECTS OF ROTARY DRYERS by Greg Palmer, B.E. Ph.D. and Tony Howes*, B.E. Ph.D. Palmer Technologies Pty Ltd, Brisbane, Australia * Department of Chemical Engineering, University of Queensland, St. Lucia 4072 Australia.
Introduction The drying of products like sand, aggregates, fertilizers and food products is an important step in industrial processes. With an increasing focus on reducing greenhouse gas emissions and energy demand the design of drying units has become critical. In the past rotary dryers have been in some cases thermally very inefficient primarily due to poor design. Fluid bed dryers on the other hand are thermally very efficient due to the interment contact of the gas stream with individual particles and a better understanding of the design principles. Thus, the science of fluid bed dryer design means these units are relatively easy to design even though thermal energy demand between each type of unit is approximately the same. The problem is that the same level of engineering knowledge has not been available for rotary dryers and as consequence these drying units are generally over designed and thermally inefficient. This paper discusses the difference between a rotary dryer and a fluid bed dryer used to dry slag1. It is important to understand the difference between the two pieces of equipment and the design aspects required in the rotary dryers. One of the difficulties with the design of rotary dryers is determining the amount of material falling through the gas stream at any moment in time. Work carried out by Wang, Cameron and Lister (1990??) ……….. Due to the complexity in calculating the percentage hold up in the gas stream and the percentage hold up in the lifter various lifter designs have been tried over the years on a trail and error basis, many with poor results. Because of complexity of estimating some parameters most rotary dryers are over designed and as a consequence the final product can be over dried and heated wasting thermal energy and higher than necessary equipment costs. Work carried out by Palmer Technologies and The University of Queensland has been aimed at understand the aspects of drying in a rotary dryer. This work enabled computer models to be developed and validated against numerous industrial dryers in the sand and cement industries. The results from drying slag using a fluid bed dryer and a rotary dryer are first compared followed by the design criteria for rotary dryers are discussed in this paper. Drying in General The three types of continuous dryers used throughout industry are, flash dryers, fluid bed dryers and rotary dryers. With flash dryers the heat transfer from the gas to the suspended solids is high, and drying is rapid, with a drying time in the order of 3 to 4 seconds. Fluid bed dryers also have a short drying time though the residence time is more variable and particle size dependent. A fluid bed dryer will have an average residence time of approximately 30 to 60 seconds. The third type of dryer, rotary dryers, has a low heat transfer rate in comparison.
1
a by-product from the steel industry
The residence time in a rotary unit varies from about 5 to 25 minutes depending on the product to be dried.
Figure 1 Variation of bed pressure drop versus fluidizing velocity
Figure 1 is a typical velocity versus pressure drop curve associated with fluid bed units. To achieve fluidization of a packed bed the air volume drag force must be equal to the net weight of solids in the bed. Figure 1 shows the increase in pressure with gas velocity and the point at which fluidization occurs. Fluid bed dryers take advantage of the solid gas contact and the fact that solid particles are discrete in the gas stream. The downside with fluid bed units is the high electrical energy required to maintain the pressure drop across the bed. Most dryers can operate with a gas temperature well above the boiling point of water but some products are limited for quality reasons, eg breadcrumbs. In this case dryers operating at low gas temperatures, ie below the boiling point of the fluid to be evaporated, are mass transfer dominated. When the units are operated above the fluid's boiling point then heat transfer dominates the drying. The effect of gas velocity, assuming negligible radiation and conduction, on the drying rate (constant drying rate, Nc) is proportional to the gas mass velocity, G, to the power of 0.8 (ie, G0.8). The effect of gas temperature is directly proportional to the energy transfer and drying rate (Nc), thus increasing the gas temperature will increase the drying rate. The effect of gas humidity on the drying rate is inversely proportional, thus as the gas humidity increases the drying rate decreases (for constant gas temperature). In the drying of a materials like slags and aggregates, which can be highly porous, the removal of moisture is heat transfer dominated and the transport of free moisture to the surface is controlled by capillarity. As long as the transport of water to the surface of the particle keeps the surface wet the drying rate remains constant. As drying continues the water layer recedes into the particle and the drying rate starts to fall. A point is reached where the interfacial tension in the capillary breaks and the pore fills with air. This state is called the pendular state. At this point the drying rate decreases rapidly and the vaporization rate is independent of the
fluidizing air velocity. This sorption effect is particularly important is the energy requirement to dry the product can be significantly higher than the heat of vapourization. In this paper two units have been evaluated a rotary dryer and a fluid bed dryer both drying slag. Energy Balance over a Fluid Bed Dryer As already mentioned a fluid bed dryer operates on the principal of the upward drag force on the packed bed equaling the weight of material in the bed. At this point the bed will start to become fluid and intimate contact is achieved between the bed and the hot gas. To measure the efficiency of the fluid bed dryer all the input streams are measured and a mass energy balance was calculated. In this case the energy mass balance was carried on a Dorr Oliver fluid bed dryer. The unit has a fluid bed combustion zone and the hot gas from coal combustion passes through the wet slag bed. Coal is used because of it very low thermal energy costs. The energy balance figures are shown in Table 1. The figures show the unit has a specific energy consumption of approximately 600 MJ/t which is considered quite good for slag which can have a moisture content as high as 16%. Also the energy balance shows approximately 65% of the thermal input energy goes in evaporation of the water and the specific air requirements are 0.33kgair/kgwet feed. Table 1 Mass energy balance fluid bed dryer Heat Input Fluidizing air Fuel Conv air Overbed air Slag (wet) TOTAL
Heat Output MJ/h 717 19,810 35.17 135.35 1573 22,270
% 3.219 88.952 0.158 0.608 7.063 100.00
Exit gas (wet) Slag (dry) Rad & Conv. water (evap)
MJ/h 3,689 3,306 440 14,769
% 16.614 14.887 1.982 66.517
TOTAL
22,204
100
Reference 00C and 1 atm.
The electrical energy on the other hand is approximately XXkWh/t. Energy Balance over a Rotary Dryer To measure the efficiency of the rotary all the input streams were measured and a mass energy balance was calculated. In this case the energy mass balance was carried on an Armstrong Holland rotary dryer. Measurements on both units were measured using a vane anemometer or a pitot tube. A mass energy balance results on the dryer are presented in Table I. The energy balance gives a breakdown of energy on the output streams. As can be seen approximately 71 percent of the energy goes into evaporating the water. The specific air requirement for this type of unit is approximately 0.22kgair/kgwet feed.
Table 2 Mass energy balance on rotary dryer Heat Input
Heat Output MJ/h 136 14,874 0.02 62.70 677 15,750
Secondary air Fuel False air Primary air Feed (wet) TOTAL
% 0.865 94.439 0.000 0.398 4.297 100
Exit gas Product (dry) Rad & Conv. water (evap)
MJ/h 2,180 1,973 300 11,212
% 13.918 12.596 1.915 71.571
TOTAL
15,666
100
The electrical energy was measured at approximately 0.13kWh/t. Thus it can be seen that advantages exists with rotary dryers if the a better understanding of the critical parameters associated with the design of rotary units can be a achieved. Drying in Rotating Drums The numerical models developed are based on well known heat transfer equations used to calculate the heat transfer coefficients under conduction, convection and radiation conditions. The equations are empirical relations for Nusselt, Reynolds and Prandtl numbers and care must be taken when extrapolation and with scale up. The model used to predict the drying rates have been correlated against actual dryer conditions. A number of assumptions with respect to particle size, curtain density and surface area are required. The models have been used to predict the drying of flyash through a different rotary dryer originally used to dry slag. The aim was to increase the throughput, which was initially less than 1tph. The predicted and actual moisture results under the initial conditions are presented in Figure 2 below. A good correlation between actual and predicted has been achieved. Flyash Drying Curves (low throughput) 14.00%
Moisture (%)
12.00% 10.00% 8.00%
Axial sample
6.00%
Predicted
4.00% 2.00% 0.00% 0
5
10
15
Axial distance from feed end (m)
Figure 2 Initial drying curves (actual and predicted) at the low throughput
Flyash Drying Curves (high throughput) 20.00%
Moisture (%)
18.00% 16.00% 14.00% 12.00%
Axial sam ple 1
10.00%
Axial sam ple 2
8.00% 6.00%
Predicted
4.00% 2.00% 0.00% 0
2
4
6
8
10
Axial distance from feed end (m)
Figure 3 Predicted drying curves at the increase throughput The capacity of the dryer was increased to near the desired value and the model rerun the results are shown in Figure 3 above. Again a good correlation was achieved but the drying of the flyash to less then 7% is not possible. The issue is then how best to improve or increase the drying rate. A change in the rotational speed will increase the residence time in the drum but at the expense of solid to gas contact time. Under this condition heat transfer is pushed from convection towards conduction. The model is then used to predict the necessary drum speed and angle to increase throughput by a factor of 5. The curve shown has a drum slope of 1.50 and the rotational speed of 6rpm. The predicted drying curve for changes to the dryer slope and speed is shown in Figure 4 below. Predicted drying curve "Flyash" (5tph feed, 6 rpm , slope 1.5 0)
Moisture kg water/kg feed (%)
18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% 0
2
4
6
8
10
Drum distance (m )
Figure 4 Predicted drying curve for increased drum speed and decreased slope.
The internal lifter design has been done for the dryer at the high throughput rate. A photograph
Figure 5 View from discharge end with new lifters installed. of the drum internals is shown in Figure 5. In this case the lifters have been designed to suit the material characteristics, which change as the material goes from wet to dry, and throughput. The lifters are designed by software developed by Palmer Technologies and is discussed in a separate paper. However, care must be taken when extrapolating the data.
DESIGN OF ROTARY DRYERS The function of a rotary is to pass falling particles through a hot gas stream and in so doing dry the material as it passes through the rotating drum. Drying of a particle can be analyzed using heat transfer empirical equations relating the Reynolds and Prandtl numbers. In general terms if the velocity of the gas stream is increased then the heat transfer coefficient is increased and consequently the drying of the particle is improved. However, there is a practical limit to this in that the increasing the gas velocity will increase the amount of dust entrained and removed from the dryer. Quality problems may arise from this and consequently there is an upper limit to the velocity inside the drum. The aim in designing internal fitments for rotary dyers is to maximize the amount of material and the time spent by a particle in the gas stream.
CONCLUSIONS It is has been shown that fluid bed dryers have a similar thermal efficient as a rotary dryer with approximately 70% of the thermal energy going to evaporation of the water. However, the specific electrical energy consumption on a rotary unit can be approximately XX% less due to the higher pressure drop associated with fluidization.
