The proposed dryer must satisfy the capability of the boiler and bagasse is required to dry before feeding into the boiler. The essential boiler information (Table 1) has been supplied by the company concerning the target boiler. The information regarding bagasse (Table 2) is sourced or provided from the sugar mill.

Table 1

Table 2

Feeder selection

After consultation with the company about the importance of design aspects and design requirements, a weighted decision matrix (Ouye, Facility Technics Facility Management Consulting) was devised for selecting the best- fitted feeder for the dryer design. The rotary feeder has found as the most appropriate choice for both the inlet and outlet feeders (Table 3). Using the boiler and bagasse information (Table 1-2), the required volumetric feed can be calculated as ~0.111 m³/s.

Table 3

It should be noted that the bagasse is a relatively light density material and at times due to compaction, flows in a sluggish manner. As such, the fill capacity of the bagasse in the rotor pockets is approximately 60-80%. The worst-case scenario (60%) should be considered to identify the best rotary feeder model that could supply the required feed rate of 0.111 m³/s. From the available Meyer rotary feeder types (Company Catalogue, Smoot Inc, 2014), a feeder is available to handle bagasse demand of 36 x 36 inch at 20 RPM (Company Catalogue, Smoot Inc, 2014), capable of transporting 0.113 m³/s when our design requires 0.111 m³/s (Table 4). Considering gear ration of 1:50 between the shaft and the final drive on the motor and 90% efficiency, the required amount of motor power for the selected feeder is 2 hp and the outlet feeder can be specified as the same feeder as the primary feeder (Meyer 36 x 36 inches rotary feeder (Company Catalogue, Smoot Inc, 2014).

Table 4

Dryer

Due to the limited space available at the involved sugar mill, the short residence time required to meet the 60 tph throughput (Table 1-2), and the nature of bagasse (fibrous), a flash dryer (or pneumatic dryer) system (Sosa-Arnao & Nebra, 2009)(Borde & Levy, 2006) would be the most appropriate. The hot gas stream, (in this case, flue gas from the boiler) can be used to evaporate the moisture from the bagasse (solid). Following the drying process, separation via a dry scrubber can be implemented. The appropriateness of the type of flash dryer was determined using a weighted decision matrix (Ouye, Facility Technics Facility Management Consulting) (Table 5). and it was determined that the single-pass circular drying tube is the most appropriate for the design that was later approved by engineers from the company with dimensions.

Table 5

A thin-layer drying model (Vijayaraj, Saravanan, & Renganarayanan, 2007) was used and it was found that the Page model (Eq. 1) is the most appropriate for the modeling of bagasse drying (Gilberd & Sheehan, 2013) (Vijayaraj et al., 2007) (Polanco, Kochergin, & Alvarez, 2013) The equilibrium moisture of bagasse can be calculated by the following equation-

Where, M = material moisture content (%), M _e = equilibrium moisture content in % dry basis, M _o = initial moisture content (%), k = drying rate coefficient (s^-1), n = empirical constant (dimensionless), t = required drying time (s), T = temperature (K), H = relative humidity (%). W, K, K ₁ , K ₂ are all factors that can be solved by using the equations below-

W = -7.7 - 0.1982T + 22.305T⁰⁵

K = - 2778.14 - 2042.09T + 5238.88T⁰⁵

K = - 70.42 - 13.68T + 180.22T⁰⁵

K₂ = - 194.01 - 0.62T + 51.48T⁰⁵

The drying rate coefficient (k), can be determined by multiple regression analyses based upon the Page model constraints (Vijayaraj et al., 2007) (Polanco, Kochergin, & Alvarez, 2013).

Where, H _d = drying air humidity (g of wet air/ kg of dry air), H _t = product thickness (mm). The empirical constant (n) is often calculated by the following equation

However, the investigation has found that this constant can be let to equal n = 0.1485482 for the drying of bagasse (Vijayaraj et al., 2007) (Gilberd & Sheehan, 2013). Again, the dried bagasse is fluidized by drying gas (flue gas in this case) to ensure an even flow of bagasse and to reduce the chance of clumping of particles. The required terminal velocity (V _t ) for the fluidization is selected by an empirical relation V _t = 2.96 D _hed ⁰⁵⁸ (Rasul, Rudolph, & Carsky, 1999), where D _hed is the equivalent diameter (mm). The flue gas compositions are supplied from the target company for the boiler (Table 6).

Table 6

The humidity of the bagasse H = (m_v/m_a) can be found using the molar volume of the flue gas at operating temperature (155 °C), where mv = mass of moisture present, m_a = total dry mass. Trial and error methods were conducted to determine the expected drying parameters at the low drying time. For an arbitrary volume of 10 m³ bagasse, the calculation is made to obtain the mass of flue gases (Table 7) and drying parameters (Table 8). The drying time is within the expected range for a flash drying system.

Table 7

Table 8

According to the company, the estimated flue gas exiting from the boiler is m _fg = 227,672 kg/hr assuming _pfg = 1 kg/m³ of gas flow. The bagasse flow rate is m _b = 60 tph assuming _pb = 150 kg/m³ of bagasse flow. Calculating a total flow balance with the summation of the flue gas and bagasse flow, we can determine the desired diameter and length of the dryer (Table 9).

Table 9

Parametric Study for Dryer

Due to the size restriction of a 6 x 6 x 25 m footprint for the bagasse drying system, the drying tube was optimized for both performance and design requirements emplaced by the company. The fluid mixture velocity influences the dryer length (Fig. 1(a)). As the dryer length should be kept to a minimum, it is advantageous to keep the fluid mixture velocity to a minimum; thus, the values calculated for the arbitrary volume of the bagasse (V = 6.742 m/s, and L =4.61 m) were kept fixed and used.

Fig. 1

At fixed dryer length (L = 4.61 m), it is necessary to reduce the dryer diameter (D) to limit the flow rate of flue gas into the dryer. The excess flue gas is directed to the exhaust stack as per the current operation. However, the optimum flow rate of the flue gas can be estimated at a particular dryer diameter using the figures presented in Fig. 1(b).

From this analysis, it has been determined that a flow rate of flue gas of 6 m³/s, results in a dryer diameter of 1 m (Fig. 1(b)). This dimension allows for a minimal footprint while being comparable to the required inlet for the cyclone, thus reducing losses, and reducing the chance of bagasse becoming stuck and igniting during operation (Table 10).

Table 10

Mass balance

Assuming the desired outlet bagasse moisture at 30% (as per the expected goal), an overall mass balance can be achieved as below-

As the dry components of bagasse and flue gas are equal before and after the reaction, ignoring these parameters the modified mass balance equation is-

This mass balance can be broken into its respective mass flowrates and moisture content as shown below-

Calculated summary from the from the above equations, presented in Table 11 and Table 12.

Table 11

Table 12

Bagasse Exit Temperature

To calculate the temperature of bagasse particles at the exit of the drying tube, a lumped capacitance method is assumed with flue gas temperature of 150 °C, bagasse properties are constant, bagasse particles are spherical and have no effect on neighboring particles. Using a free body diagram of the idealized case (Fig. 2) and using flue gas and bagasse properties (Table 13), the final bagasse temperature (T_f) can be determined using the following relation-

Table 13

Fig. 2:

Using the empirical correlations for heat transfer coefficient available for gas-particle flows [6], we can determine the Reynolds number (Re = 387.16) and Nusselt number (Nu = 58.07) to obtain the value h _fg = 1164.52 W/m²K as well as T _f = 67.96 °C. ^g

Heat Loss through Drying Tube

In order to determine the insulation, thickness the Fig. 3 and Fig. 4 was used and result summarized in Table 14.

Fig. 3

Fig. 4

Table 14 summarizes the additional parameters used, and the correlations or equations used to find them. The heat transfer throughout the drying tube occurs via three mechanisms, (i) Convective (internal flow), (ii) Conductive (through drying tube, and insulation) and (iii) Convective (external flow) (Table 15).

Table 14

Table 15

An analysis is first conducted to find the outer surface temperature of the drying tube without insulation. The overall heat transfer rate is calculated by Eq. 9-10.

By establishing the overall heat transfer rate, we can conduct a simple nodal analysis, working from inside the drying tube to the outside.

This analysis allows the determination of the inner (96.06 °C) and outer (95.96 °C) surface temperatures.

Knowing the outer temperature of the drying tube, we can now specify an appropriate insulative material. A500 Insulative paint (Product catalogue, Insulpaint Australia Pty Ltd, 2004) from Insulpaint Australia was chosen to be applied to the outside of the drying tube. A parametric study to find an appropriate thickness of insulation was conducted (Fig. 5). Including insulation, R _tot becomes,

Fig. 5

It is recommended that a paint coating of between 2-5 mm be applied. Assuming a coating of 4 mm, Table 16 summarizes the calculated surface temperatures for completeness.

Table 16

To calculate the exit temperature of the flue gas, the overall energy balance has been considered as shown in Fig. 6. It should be noted that the “one bagasse particle” is a total representation of all the individual particles in the dryer. The overall balance is described below-

Fig. 6

From Table 16 the overall heat transfer rate out of the dryer wall is 8034.93 W with a 4 mm insulative paint applied. Therefore the Eq. 12 can be solved for all other parameters.

Heat Transfer into Bagasse

In works published by Mujumdar (Mujumdar, 2014), the heat transfer rate (per volume) between and gas and solid phase can be calculated via the following Eq. 13.

where, Є _s = solid volume fraction in the mixture, d _s = diameter of bagasse particle (m), h _gs = heat transfer coefficient (W/m²K), T _g = temperature of the gas, and T_s = average temperature of the bagasse. The solid volume fraction is calculated as Є _s = Q _b /(Q _b +Q _g ) = 0.111/ (0.111+6.0) = 0.01816. Using the Baeyens et al. (Maurice, Mezhericher, Levy, & Borde, 2015) correlation for the Nusselt’s number, the heat transfer coefficient can be established and the heat transfer rate calculated (Table 17).

Table 17

Thus, by substituting the known values into Eq. 13, we can solve the heat transfer rate (per volume) into a singular bagasse particle as q _gs = q _bag = 6,892,047.57 W/m³. Knowing the volume of one particle the heat transfer rate into one particle is q _bag = 0.02322 W, the total heat transfer rate can be found as q _{bag, tot} = 765541 W.

Heat Transfer Rate In and Out

To calculate the energy at the inlet and outlet of the drying tube, an analysis of the flue gas enthalpy was conducted, assuming constant pressure throughout. Expressions for the enthalpy at the inlet and outlet are presented below (Eq. 14-15). Table 18 summarizes the parameters used to solve the enthalpy analysis.

Table 18

Thus, Eq. 14 and Eq. 15 gives q_in = 5593055 W and q_out = 8876280 + 10224.108 T _{fg out} respectively. By subbing all components into Eq. 15, we can solve the outlet temperature T _{fg out} = 396.8 K = 123.8 °C. Table 19 below summarizes the parameters determined via the heat and mass analysis as described above.

Table 19

Cyclone separator

In the case of separating the dried bagasse from the moisture-rich flue gas exiting the dryer the most common form of the dry scrubber is a cyclone separator (Strumiłło, L. Jones, & Zyłła, 2014)(Moor, 2007)(Bashir, 2015)(Pynadathu, Thomas, & Arjun, 2014). The most common type of dryer is shown in Fig. 7. Over the last 50 years, the design and optimization of cyclones have been extensively investigated by researchers such as Stairmand, Swift 1, Lapple, Swift 2 and Peterson/ Whitby (Leith, 1990). Each design determines the various component lengths and diameters as a function of the body diameter (D ) as outlined below in ref (Leith, 1990).

Fig. 7

From this review of the literature and consultation with the company (required throughput rate of 60 tph), a high throughput cyclone is required. The scheme chosen is the Swift scheme (Leith, 1990) which is a function of the input flow rate, inlet velocity and combined density (Table 20).

Table 20

The main diameter is calculated and found to be D = 1.39 m () according to the ref. (Leith, 1990). In reference to Fig. 8, the following equations describe the functional relationships between the major diameter and the other functional variables (Bashir, 2015)(Leith, 1990).

Fig. 8

Table 21

The pressure drops = 13.38 cm of water ( ) across the cyclone is a function of the velocity heads at the inlet and outlet N _H = 3.804 (N_h = (K_VH)(ab)/De²), the density of the gas ( kg/m³), and the velocity of the inlet (V _i m/s). Using the Shepard and Lapple empirical relationship for velocity heads at the inlet and outlet of the cyclone, the empirical constant has a value K _VH = 16 for tangential inlets and K _VH = 7.5 for cyclones with an inlet vane [10].

A forced draught fan is installed upstream of the drying tube to ensure the desired flue gas flowrate is achieved. To ensure the correct specification of the fan, the losses in the system must be considered. This can be achieved by applying a control volume (Fig. 9) and solving Bernoulli’s equation (Eq. 16).

Fig. 9

where, P - absolute pressure (Pa), V - velocity (m/s), z - elevation (m), h _L ,_1-2 - head loss (m). The h _L ,_1-2 can also be described as h _L , _1-2 = h _dryer + h _cyclone . Table 22 below contains the parameters, equations, and assumptions made to solve for the required fan head.

Table 22

The head loss through the dryer can be calculated via Eq. 17

while the head loss due to the pressure drop in the cyclone can be found via h_dryer = ΔP/pg. With the given information, Eq. 16 can be solved (Table 23).

Table 23

To be able to source parts and spares nationally a fan will be selected from Aerotech (Company catalogue, Aerotech Fans Pty. Ltd.). It has been determined that a centrifugal pump is most suited for the bagasse drying system. From these options, it is apparent that the MAVX405 (Table 24) (Company catalogue, Aerotech Fans Pty. Ltd.) is the only option that satisfies our criteria. It has the required flow rate, while also having a large enough static pressure head for our requirements.

Table 24

After consultation with the company about the importance of design aspects and design requirements, a weighted decision matrix (Ouye, Facility Technics Facility Management Consulting) was devised (Table 25) and determined that the single-pass circular drying tube is the most appropriate and accepted by company engineers.

Table 25

After determining all dryer design parameters and parts specifications, this paper proposes an improved design for the bagasse drying system for the boiler of the cogeneration plant (Fig. 10). A scale model of the overall structure of the proposed design has drawn (Fig. 10), which included all the parts (feeders, fan, drying tube, and cyclone). All parts of the proposed design (Fig. 10) are supported by a structure that fits within the specified area of the mill without hindering any accessibility problem. The proposed design can be constructed in-place to dry the bagasse. Since all the parts and specifications for this design are derived for industrial scale, it will be easy to modify the proposed design with necessary bagasse and boiler information for any kind of industrial bagasse drying system.

Figure 10