47 then equation (3.41) becomes
π2 = π1βπ’πππππ π
π΄ π₯ βπ ππ€
π΄ π₯ β πΌπππππ₯π₯ +ππππ
π΄ π₯ (3.42)
48
Equation (3.36) will provide the pressure at discrete points along the flash tube and this can be used to predict the flow velocity profile of the continuous phase and using the basic thermodynamic equation;
π1π 1 π1 =π2π 2
π2 (3.46)
where , π and π are pressure, volumetric flow rate, and temperature and for a pipe section with uniform cross sectional area:
π’π2= π1π’π1π2
π2π1 (3.47)
Where, π = π΄π’π (A = pipe cross sectional area).
It should be emphasized that absolute values of both pressure and temperature must always be used in these equations. Most data for these values, such as that for minimum conveying air velocity are generally determined experimentally or from operating experience. It is for the purposes of this work, important to take the presence of the particles into account because in accelerating the material at zero velocity at the feed point to some value along the flow line requires momentum exchange between the particles and the continuous phase.
In dilute phase conveying, with particles in suspension in the air, the mechanism of conveying is one of drag force. The velocity of the particles, therefore, will be lower than that of the conveying air. It is a difficult and complex process to measure particle velocity, and apart from research purposes, particle velocity is rarely measured. Once again it is generally only the velocity of the air that is ever referred to in pneumatic conveying.
In a horizontal pipeline the velocity of the particles will typically be about 80% of that of the air. This is usually expressed in terms of a slip ratio, defined in terms of the
49
velocity of the particles divided by the velocity of the air transporting the particles, and in this case it would be 0.8. The value depends upon the particle size, shape and density, and so the value can vary over an extremely wide range. In vertically upward flow in a pipeline a typical value of the slip ratio will be about 0.7.
These values relate to steady flow conditions in pipelines remote from the point at which the material is fed into the pipeline, bends in the pipeline and other possible flow disturbances and shall be used as a ball pack check on the result of analytical methods. At the point at which the material is fed into the pipeline, the material will essentially have zero velocity. The material will then be accelerated by the conveying air to its slip velocity value. This process will require a significant pipeline length and this is referred to as the acceleration length. The actual distance will depend once again on particle size, shape and density.
There is a pressure drop associated with acceleration of the particles in the air stream and it has to be taken into account by some means. It is not only at the material feed point that there is an acceleration pressure drop. It is likely to occur at all bends in the pipeline. In traversing a bend the particles will generally make impact with the
bend wall and so be retarded. The slip velocity at exit from a bend will be lower than that at inlet and so the particles will have to be re-accelerated back to their steady-state value. This additional element of the pressure drop is usually incorporated in the overall loss associated with a bend.
The momentum coupling source term (per unit volume) due to the reverse effect of particles can be expressed as suggested by Hamed M. H (2005):
ππππ = βππ1
2πΆπ·πππ
2
4 πππ₯ π’πβ π’π π’π β π’π (3.48)
50 3.3.5 Energy Equation
In writing the energy equations for a multi phase flow, it is necessary to construct an energy equation for each of the phases or components. First total energy density (per unit mass) ππβ is defined for each component such that
ππβ = ππ +1
2π’πππ’ππ + ππ₯ (3.49)
Then the appropriate statement of the first law of thermodynamics for each phase becomes:
Rate of heat addition to N from outside control volume, ππ + Rate of work done to N by the exterior surroundings, ππ΄π + Heat transfer to N within the control volume, ππΌπ
+ Rate of work done to N by the other component in the control volume, ππΌπ
= Rate of increase of total kinetic energy of N in control volume
+ Net flux of internal energy of N out of the control volume (3.50)
The second term on the RHS of equation (3.50) contains two contributions: (i) minus the rate of work done by the stress acting on the component of N on the surface of the control volume and (ii) the rate of external shaft work, ππ, done on the component N.
In evaluating the first of these, the same modifications to the control volume as we did for the momentum equation are made; specifically small deformations is made to the control volume so that its boundaries lie wholly within the continuous phase.
Then using continuous phase stress tensor,ππππ, as defined earlier the expression for ππ΄π becomes:
ππ΄π = ππ+ π
ππ₯π π’ππππππ (3.51)
51 And
ππ΄π = ππ (3.52) Also the last two terms of equation (3.3.60) can be written as
π
ππ‘ πππΌππππβ + π
ππ₯π πππΌπππβπ’ππ (3.53) Then the energy equation can be written as:
π
ππ‘ πππΌπππβ + π
ππ₯π πππΌπππβπ’ππ = ππβ ππ + ππΌπ+ ππΌπ+ πΏπ π
ππ₯π π’ππππππ (3.54) Note that the two terms involving internal exchange of energy between the phases may be combined into an energy interaction term given by
πππππππ¦ ,π = ππΌπ+ ππΌπ (3.55) It then follows that
ππΌπ
π
= 0
and
ππΌπ
π
= 0
and
πππππππ¦ ,π
π
= 0
Moreover, the work done terms, ππΌπ, may clearly be related to the interaction forces, ππππ. In a two phase flow with one dispersed phase:
ππΌπ = βππΌπ , ππΌπ = βππΌπ = βπ’π ππΉπ π , πππππππ¦ ,π = βπππππππ¦ ,π (3.56) When the left hand side of equation (3.54) are expanded and use is made of continuity equations and momentum equation , it results in the thermodynamic form of the energy equation. Using expression (3.54) and the relation
ππ = πππππ+ ππππ π‘πππ‘ (3.57)
52
Between the internal energy, ππ , the specific heat capacity at constant volume, πππ, and the temperature, ππ, of the continuous phase, the energy equation can be written as
πππΌππΆπ£π πππ
ππ‘ + π’ππππ
ππ₯π = πΏπππππ ππ’ππ
ππ₯π + ππ+ ππ + ππΌπ+ ππππ π’π β π’π β ππββ π’π2 ππππ π (3.58)
In equation (3.58) it has been assumed that the specific heat,ππ£π, is constant and uniform. Finally the one βdimensional duct flow equation for energy balance is:
π
ππ‘ πππΌπππβ +1
π΄
π
ππ₯ π΄πππΌπππβπ’π = ππ + ππ+ ππΌπ + ππΌπ + πΏπ π
ππ₯ ππ’π (3.59) In simplifying the last term on the RHS of equation (3.59) notice that for continuous phase, πΏπ = 1 while for the dispersed phase, πΏπ = 0. And that for the flow situation under consideration, there no shaft work done on the gas component therefore
π
ππ‘ πππΌπππβ +1 π΄
π
ππ₯ π΄πππΌπππβπ’π = ππ+ ππΌπ+ ππΌπ+ π
ππ₯ ππ’π
Since there is no energy accumulation on the control volume, this further simplifies to
1 π΄
π
ππ₯ π΄πππΌπππβπ’π = ππ+ ππΌπ + ππΌπ + π
ππ₯ ππ’π (3.60) Now
π
ππ₯ π΄πππΌπππβπ’π
=ππ΄
ππ₯ πππΌπππβπ’π +πππ
ππ₯ π΄πΌπππβπ’π +ππΌπ
ππ₯ π΄ππππβπ’π +πππβ
ππ₯ π΄πππΌππ’π +ππ’π
ππ₯ π΄πππΌπππβ but
ππ΄
ππ₯ = 0 πππ
ππ₯ = 0 ππΌπ
ππ₯ = 0
So LHS of equation (3.60)
53
1 π΄
π
ππ₯ π΄πππΌπππβπ’π =πππ
β
ππ₯ πππΌππ’π +ππ’π
ππ₯ πππΌπππβ equation (3.60) becomes:
πππΌπ π’ππππ
β
ππ₯ + ππβ ππ’π
ππ₯ = ππ + ππΌπ+ ππΌπ+ π
ππ₯ ππ’π (3.61)
π
ππ₯ ππ’π = π’πππ
ππ₯+ πππ’π
ππ₯ let
ππ’π
ππ₯ = ππππ π πππΌπ = π πππ
ππ₯ = βπΌπππππ₯ βπ ππ€
π΄ +ππππ
π΄ β π’πππππ π = π
Since the values of ππ’π
ππ₯ and πππ
ππ₯ are coefficients their values can be replaced with a and b for convenience
π
ππ₯ ππ’π = π’ππ + ππ Equation (3.61) can be rewritten as
πππΌπ π’ππππβ
ππ₯ + ππβππππ π
πππΌπ = ππ+ ππΌπ+ ππΌπ + π’ππ + ππ rearranging
πππΌππ’ππππβ
ππ₯ + ππβππππ π = ππ + ππΌπ + ππΌπ+ π’ππ + ππ and
πππβ
ππ₯ + ππβ ππππ π
πππΌππ’π = 1
πππΌππ’π ππ+ ππΌπ+ ππΌπ + π’ππ + ππ (3.62) Again, since the entire LHS is a constant, it is denoted with m for convenience
1
πππΌππ’π ππ+ πππππππ¦ + π’ππ + ππ = π
54 Equation (3.62) becomes
πππβ
ππ₯ + ππβ ππππ π
πππΌππ’π = π
πππβ
ππ₯ = π β ππβ ππππ π πππΌππ’π
πππβ
ππ₯ = βππππΌππ’π
ππππ π +ππβ β ππππ π πππΌππ’π
πππβ ππ₯ ππββπ π π πΌπ π’ π
ππππ π
= β ππππ π
πππΌππ’π (3.63)
Equation (3.63) is variable separable and integrating gives ππ ππββππππΌππ’π
ππππ π = β ππππ π
πππΌππ’ππ₯ + πΆ ππβ = π πππΌππ’π
ππππ π + πΆπβ
ππππ π π π πΌπ π’ ππ₯
(3.64) Using the following initial conditions
π₯ = 0 ; ππβ = ππ1β Then
πΆ = ππ1β βππππΌππ’π ππππ π ππβ = ππππΌππ’π
ππππ π + ππ1β βππππΌππ’π ππππ π πβ
ππππ π πππΌππ’ππ₯
ππβ = πππππ+π’π2 2 + ππ₯ ππ2= ππππΌππ’π
ππππ π + ππ1β βππππΌππ’π ππππ π πβ
ππππ π πππΌππ’ππ₯
βπ’π2
2 β ππ₯ πππ
ππ2 = ππππΌππ’π
ππππ π + πππππ1+π’π1
2
2 + ππ₯ βπ πππΌππ’π
ππππ π πβ
ππππ π π ππΌ ππ’ ππ₯
βπ’π2
2 β ππ₯ πππ (3.65)
55 3.3.6 Constitutive Relationships (Energy)
The energy coupling source term for the total energy equation involves convective heat transfer and the work due to particle drag as suggested by Hamed M. H (2005) is expressed as:
πππππππ¦ = βππππ΄ππππ2 ππβ ππ + πππππ’π (3.66) The dispersed phase is introduced into the dispersing phase at a point along the flow path; the feed point which is always upstream of the flash tube. At this point the dispersed phase temperature is much smaller than the dispersing phase temperature.
Heat transfer between the phases tends to reduce the difference in temperature.
Therefore it is necessary to characterize the rate of equilibration of the particle and fluid temperatures by defining a temperature relaxation time, π‘π. This temperature relaxation time can be obtained by equating the rate of heat transfer from the continuous phase to the particle with the rate of increase of heat stored in the particle.
The heat transfer to the particle can occur as a result of conduction, convection or radiation and there are practical flows in which each of these mechanisms are important but for the situation at hand the radiation component shall be neglected.
If the relative motion between the particles and the fluid is sufficiently small, the only contributing mechanism is conduction and it is limited by the thermal conductivity, ππ of the gas since the thermal conductivity of the particle is usually much higher.
Then the rate of heat transfer to the particle of radius, R will be given approximately by
2ππ ππ(ππ β ππ ) (3.67)
where ππ πππ ππ are respectively temperatures of the gas phase and the particle.
Since the situation being modelled involves conveyance and drying, the relative motion that is slip velocity can only be low to the extent that it guarantees conveyance
56
and in this situation conveyance takes precedence over heat transfer which drives drying.
In determining the convective heat transfer coefficient the empirical approach was used. The usual drawback of using the empirical approach is that it requires a large number of experiments to obtain the required data. This challenge is overcome by the use of dimensionless numbers. To formulate this approach, first the required dimensionless numbers are identified: Reynolds number, Re, Nusselt number, Nu, and Prantl number, Pr
To add the component of heat transfer by convection caused by relative motion is done by defining the Nusselt number, Nu, as twice the ratio of the rate of heat transfer with convection to that without convection. Then the rate of heat transfer becomes Nu times the above result for conduction.
The convective heat transfer coefficient, h, was calculated from Nusselt number, Nu, which is expressed as a function of Reynold number, Rep and Prantl number, Pr, which are defined as:
π π = 2 π’π β π’π π /π£π (3.68) ππ =ππ π£ππΆππ
ππ (3.69) Various empirical correlations that can be used to calculate the heat transfer coefficient has been proposed and are listed below.
ο· Frantz correlation (Radford R. D., 1997)
The correlation was used by Radford to calculate the heat transfer coefficient in pneumatic conveying dryer.
ππ’ = 0.015π ππ1.6ππ0.667 (3.70)
ο· De Brandt correlation (Fyhr C. and Rasmuson A., 1997 )
57
The correlation was developed for pneumatic drier,
ππ’ = 0.16π ππ1.6ππ0.667 (3.71)
ο· De Brand correlation (Debrand S., 1974 ) The correlation was developed for a pneumatic dryer,
ππ’ = 0.035π ππ1.15ππ0.333 (3.72)
ο· Bayeans et al. Correlation (Baeyens et al, 1995) The correlation was developed for large scale pneumatic conveyor
ππ’ = 0.15π ππ (3.73)
ο· Modified Ranz-Marshall correlation (Levy and Borde, 1999) The correlation was developed for simple droplet/wet particle and it takes into account the resistance of the liquid vapour around the particle to the heat transfer by Spalding number, B.
ππ’ =2+0.6π ππ
0.5ππ0.333
1+π΅ 0.7 (3.74) π΅ =πΆππ€π£ ππβππ
π»ππ (3.75)
ο· Modified Weber correlation (Kemp et al, 1994) An additional term proportional to Rep0.8
was added to Ranz-Marshall correlation to account for turbulent flow.
ππ’ = 2 + 0.5π ππ0.5+ 0.06π ππ0.8 ππ0.333 (3.76)
ο· Ranz and Marshall correlation
ππ’ = 2 + 0.6π π1/2ππ1/3 (3.77)
The correlation above reduces to pure conduction result, Nu = 2, when the second term on the right hand is small. Assuming that the particle temperature has a roughly uniform value of ππ , it follows that
ππΌπ = 2ππ ππππ’ ππβ ππ ππ = ππ πΌπ ππ π·ππ
π·π‘ (3.78)
58
where the material derivative D/Dt, follows the particle. This provides the equation that must be solved for ππ , namely
π·ππ π·π‘ =ππ’
2 ππβππ
π‘π (3.79)
where,
π‘π = ππ ππ π 2/3ππ (3.80)
ο· Singh and Heldman correlation (Singh and Heldman, 2001)
For a flow past a single sphere, when the single sphere may be heated or cooled, the following equation will apply:
ππ’ = 2 + 0.60π π12ππ13 for 1 < π π < 70000 and 0.6 < ππ < 400 (3.81) where the characteristic dimension, ππ, is the outside diameter of the sphere.
The correlation suggested by Singh and Heldman (2001) shall be used for this work.
59