EXPERIMENTAL DETERMINATION OF THE APPARENT VISCOSITY OF SLURRIES: SAND-WATER SYSTEM, USING THE ROTATIONAL VISCOMETER METHOD

BED TIBOR* Transilvania University of Brasov, B-dul Eroilor nr. 29, 500036, Brasov, Romania

Abstract: Measurement and characterization of viscous behavior of non-Newtonian fluids serve multiple purposes. They help understanding a fluid's response to deformations; knowledge essential in developing critical specifications for handling, transportation, application, and in determining the optimum conditions for the efficient operation of processing equipment. This information pertains to design and/or select appropriate equipment for mixing, dispersing, storing, pumping, and processing. In this paper the measurement of the apparent viscosity of slurries: sand-water system, was made by a simple technique employing the rotational viscometer method.

Keywords: viscosity, slurries, rotational viscometer method

1. INTRODUCTION Slurry and non-Newtonian flows occurs in a wide range of industries, for example mining, wastewater and food processing. These flows are divided into two separate types called non-Newtonian/non-settling and nonNewtonian/settling. Non-settling slurries entail very ne particles which can form stable homogeneous mixtures exhibiting increased apparent viscosity. These slurries usually have low wearing properties but require very careful consideration when selecting the correct pump and drive as they often do not behave in the manner of a normal liquid. When ne solids are present in the slurry in sufcient quantity to cause this change in behavior away from a normal liquid, they are referred to as non-Newtonian. Settling slurries are formed by coarser particles and tend to form an unstable mixture. Therefore, particular attention must be given to ow and power calculations. These coarser particles tend to have higher wearing properties and form the majority of slurry applications. This type of slurry is also referred to as heterogeneous [1]. A variety of industrial operations (mineral, metal and material processing) require the transportation of slurries and suspensions. Measurement and characterization of viscous behavior of non-Newtonian fluids serve multiple purposes. They help understanding a fluid's response to deformations; knowledge essential in developing critical specifications for handling, transportation, application, and in determining the optimum conditions for the efficient operation of processing equipment. This information pertains to design and/or select appropriate equipment for mixing, dispersing, storing, pumping, and processing. Viscosity, in particular non-Newtonian viscous behavior, is an important material property that contributes to a fluid's performance, and often is the main source of problems in handling, processing, and application. The relationship between shear stress and shear rate for a Newtonian fluid can be shown as a straight line passing through the origin, where the slope of the line gives the liquids viscosity. For a non-Newtonian/non-settling slurry this is not the case. The relationship might not be linear or may not pass through the origin. The viscosity
*

email bedo.tibor@unitbv.ro

of these types of slurries is dependent on the shear rate commensurate with the engineering conditions which apply at the time. This viscosity is sometimes called apparent viscosity. The figure 1 shows the different types of viscosity [2].

Fig. 1. Types of viscosity The classical viscometry is based on generation of a deformation or motion in the slurry and observation of the resultant stresses or vice versa. In the study of fluids flow it is important to know the velocity gradient of deformation or the shear rate , because the shear stress is proportional with the deformation rate or the deformation rate is dependent of shear stress and time. The relation between shear stress and the velocity gradient (shear rate) for an isotropic fluid is given by:

ij = ij &ii + 2 &ij
i

(1)

where and are material parameters, ii is the volumetric shear rate and ij is the form deformation rate. Lot of important industrial materials is non-Newtonian fluid. From such materials the dynamic viscosity coefficient depend on shear stress. The correlation between ij and ij is:

a =

ij &ij

(2)

where a represent the apparent viscosity which is dependent on structural transformations given up in fluid by the shear stress. Apparent viscosity a slurrys resistance to deformation when subjected to shear stress. This term is applied to differentiate from absolute viscosity which is used in conjunction with Newtonian fluids. Standard rheometric methods of estimating the apparent viscosity of slurries are based either on capillary flow (Oswald and Ubbelhode viscometer) or the rotational flow (rotary viscometer). The standard viscometers as mentioned above have their limitations in measuring the apparent viscosity of slurries that are not sufficiently dilute. In rotational methods the test fluid is continuously sheared between two surfaces, one or both of which are rotating. These devices have the advantage of being able to shear the sample for an unlimited period of time, permitting transient behavior to be monitored or an equilibrium state to be achieved, under controlled rheometric conditions. Rotational methods can also incorporate oscillatory and normal stress tests for characterizing the viscoelastic properties of samples. In general, rotational methods are better suited for the measurement of concentrated suspensions, gels, and pastes, but are generally less precise as compared to capillary methods. Rotational measurements fall into one of two categories: stress-controlled or rate-controlled. In stress-controlled measurements, a constant torque is applied to the measuring tool in order to generate rotation, and the resulting rotation speed is then determined. If a well-defined tool geometry is used, the rotation speed can be converted into a corresponding shear rate. In rate-controlled measurements, a constant rotation speed is maintained and the resulting torque generated by the sample is determined using a suitable stress-sensing device, such as a torsion spring or strain gauge. Some commercial instruments have the capability of operating in either stress-controlled

or rate-controlled modes. The device measures fluid viscosity at fixed rotation speeds by driving a measurement tool ("spindle"), immersed in the test fluid (figure 2). The calculated shear rate depends on the rotation speed, the tool geometry, and i the size and shape of the sample container. For Newtonian fluids the torque is proportional to the product of viscosity and rotational speed, but this proportionality is lost in the case of a non-Newtonian fluid. Because these instruments are robust and fairly simple to use, they have found wide application in industry, but they offer limited capabilities and precision for research-oriented applications [3].

Fig. 2. The device for fluid viscosity measurement 2. EXPERIMENTAL RESEARCH Notwithstanding these limitations, the rotary viscometers are considered more efficient. In this paper the measurement of the apparent viscosity of slurries: sand-water system was made by a simple technique [4] employing the experimental set-up presented in figure 3. Typically, the flow of suspensions and slurries in an agitated vessel depends on three parameters: the vessel geometry and the aspect ratio of the vessel; the type of impeller, the number and location of blades in the impeller, the diameter of the impeller, the clearance of the impeller from the vessel bottom, etc.; the rheology of the slurry system.

Fig. 3. Experimental set-up The first two parameters decide several phenomena, such as, the macro fluid dynamics, local hydrodynamics near the agitator, local energy dissipation rate, bulk fluid mixing, etc. For characterising the flow patterns of a slurry in a stirred viscometer, the Reynolds number (NRe) and power number (NP) have been defined approximately as:

N Re

l nr D 2 =
P
3 l nr D 5

(3)

NP =

(4)

where l is the density of the fluid medium, [g/cm3], nr the speed of the impeller, [rps], D the impeller diameter, [cm], the fluid viscosity, [cP], P the power input to the agitator, [W]. In stirred vessels, the flow of a Newtonian fluid is turbulent if NRe>2104 and laminar if NRe<10. Thus, unlike pipe flow, a large transition regime between 10<NRe<2104, is observed in a stirred vessel. The power number NP represent the equivalent to the friction factor for pipes. NP vs. NRe (NP/NRe) characterises the slurry hydrodynamics in stirred vessels and is analogous to the friction factor vs. NRe in pipe flow. For the rheological fluids the product of power number (NP) and the Reynolds number (NRe) becomes a constant:

N P N Re = Ci
where the magnitude of the constant Ci depends on the impeller-vessel configuration.

(5)

The rheology of the slurry needs to be ascertained in order to determine the apparent viscosity. Adopting a suitable mathematical model can do this. Table 1 provides three models in common use for characterising slurry rheology: the power law model, the Prandtl-Eyring model and the Ellis model [5]. Table 1. Simple models for slurry rheology Equation Parameters = k n-1 = k1arcsh ( / k2) = / (k1 + k2 n-1) k-consistency k1,k2 coefficient of material k1,k2 coefficient of material

No. 1 2 3

Name of model Power law Prandtl-Eyring Ellis

Remarks n = 1 Newtonian fluids n < 1 shear thinning n > 1 shear thickening

k2 = 0 Newton law k1 = 0 power law

Since stirred vessel viscometer does not involve very high or very low shear rates, it is usually preferable to employ a two parameters power law model for describing the slurry rheology owing to its simplicity. Unfortunately the presented models are empirical expressions and their use is limited to the concrete conditions mentioned by the model. 2.1. Experimental results Employing a simple technique for measurement of apparent viscosity of slurries: sand-water system [4], in the present paper is shown the variation of apparent viscosity with sand concentration. For the experiment was used silica sand with medium granulation = 0,23 mm and the uniformity degree GU=52%. Sieve analysis of used sand is shown in table 2 and figure 4. Experiments were carried out using various weight % of sand in water: 6,25% (100 g sand + 1500 g water); 11,76% (200 g sand + 1500 g water); 16,66% (300 g sand + 1500 g water); 21,05% (400 g sand + 1500 g water); 25% (500 g sand + 1500 g water); 28,57% (600 g sand + 1500 g water). Table 2. Sive analysis of silica sand used Mesh number Particle size [mm] Wt. [%] 1,0 over 1,0 0,63 0,63 1,0 0,4 0,40 0,40 0,63 5,6 0,315 0,315 0,40 17,52 0,25 0,25 0,315 17,82 0,20 0,20 0,25 17,72 0,16 0,16 0,20 15,82 0,10 0,10 0,16 17,72 0,063 0,063 0,10 3,4 tray 0,02 0,063 4,0

No. 1 2 3 4 5 6 7 8 9 10

Fig. 4. Medium silica sand M(50) = 0,23 mm I have used in these experiments a cylindrical vessel with one blade straight propeller. The size of the tank is 150 mm diameter and 170 mm height. Height of the liquid column in unagitated condition was kept constant at 85 mm from the bottom of vessel. The diameter of the impeller is D=110 mm, the breadth of blade is 35 mm and the distance of the blade from bottom of vessel is 20 mm. The evaluation of apparent viscosity of sand-water slurries was made as follows: Using pure water at room temperature (Newtonian liquid) power data were generated at different stirrer speeds. Using the above power data NP vs. NRe curve was drawn (figure 5). For water, the density was taken as 1 g/cm3 and viscosity as 1 cP. Its easy to see that the correlation is linear and the product of power number and the Reynolds number is approximately constant. The Ci constant was calculated for each stirrer speeds. The value of Ci constant is between 0,035107 and 0,040107.
7 6 Power number 5 4 3 2 1 0 67760 76230 84700 93170 Reynolds number 101640 110110

Fig. 5. NP vs. NRe curve The power input to the stirrer was calculated for each slurry concentrations and stirrer speed. Using the power curve drawn for the Newtonian liquid and the power data calculated previously it was determined the apparent viscosity of sand water slurries as follows:

N Re(slurry) =

s nr D 2 a( s )

(6)

where s is the density of slurry (calculated by law of mixture) [g/cm3], nr revolution of stirrer per second, D diameter of the impeller [cm], a(s) apparent viscosity of the slurry system which is required to be determined [P]. 3. CONCLUSIONS Table 3 and figure 6 shows the values of viscosity of sand-water slurry for different concentrations and stirrer speed.

Table 3. Experimental results No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Slurry 11,76 % sand nr [rps] 5,6 6,3 7,0 8,4 5,6 6,3 7,0 8,4 5,6 6,3 7,0 8,4 5,6 6,3 7,0 8,4 5,6 6,3 7,0 8,4 U [V] 2 2,25 2,5 3 2 2,25 2,5 3 2 2,25 2,5 3 2 2,25 2,5 3 2 2,25 2,5 3 I [A] 1 1,1 1,17 1,3 1,05 1,15 1,25 1,4 1,1 1,2 1,3 1,45 1,15 1,25 1,35 1,5 1,15 1,25 1,4 1,55 Ci 399135,54 367308,92 364387,02 366265,84 399135,54 367308,92 364387,02 366265,84 399135,54 367308,92 364387,02 366265,84 399135,54 367308,92 364387,02 366265,84 399135,54 367308,92 364387,02 366265,84

a(s)

[Poise]

16,66 % sand

21,05 % sand

25 % sand

28,57 % sand

0,0117647 0,0125 0,0120618 0,0111111 0,0123529 0,0130682 0,0128866 0,0119658 0,0129411 0,01363638 0,01340208 0,01239316 0,01352942 0,01420456 0,01391755 0,0128205 0,01352942 0,01420456 0,01443301 0,01324786

Apparent viscosity [Poise]

0.08 28.57% sand 0.06 0.04 0.02 0 5.6 6.3 n [rps] 7 8.4 25% sand 21.05% sand 16.66 % sand 11.76% sand

Fig. 6. Correlation between the apparent viscosity of slurry [Poise] and the impeller speed [rps] It may be noticed: a simple technique for measurement of apparent viscosity of slurries (the author used sand-water system) - it was necessary only the measurement of power input to the stirrer; generally at high rotation speed of the stirrer the apparent viscosity of slurry decreased though it is expected that the viscosity should increase. This discrepancy can be explained on the basis of particle distribution around the blade, piping and air-entrainment condition and the design of the blade; the apparent viscosity are in the range of 0,011 0,014 when the slurry concentration varied between 11,76% and 28,57% and the stirrer rps between 5,6 and 8,4; over a certain value of stirrer speed the apparent viscosity of slurry decrease with the increase of the stirrer speed (figure 6), which is constant with the characteristics of the pseudoplastic fluids. REFERENCES [1] http://www.scribd.com/doc/52052493/6/CHARACTERISTICS-OF-A-SLURRY (5.03.2011) [2] http://en.wikipedia.org/wiki/File:Viscous_regimes_chart.png (15.03.2011) [3] http://ciks.cbt.nist.gov/~garbocz/SP946/node14.htm (15.03.2011) [4] Biswas, P.K., Godiwalla, K.M., Sanyal, D., Dev, S.C., A simple technique for measurement of apparent viscosity of slurries: sand-water system, Materials & Design 23 (2002), p 511 519 [5] Tudose, R.Z., Volintiru, T., Asandei, N., Lungu, M., Meric, E., Ivan, Gh., Introducere n reologie, Editura Tehnic, Bucureti, 1982