hydrocyclone separation for the remediation of...

Hydrocyclone Separation for the Remediationof Contaminated Sediment

Joshua R. WilsonDepartment of Mechanical Engineering

Saginaw Valley State UniversityUniversity Center, MI 48710Email: [email protected]

Thomas A. MahankDepartment of Mechanical Engineering

Saginaw Valley State UniversityUniversity Center, MI 48710Email: [email protected]

Abstract

Hydrocyclones offer the potential for the remediation and volume reduction of the contaminatedsediment found in dredging operations and can be utilized to minimize the mass of sediment shippedto confined disposal facilities. The objective of the undergraduate student-led research project was todevelop and test a hydrocyclone that was designed and built with university and corporate support toseparate particles in a liquid suspension. The results of the laser diffraction analysis, at a mass flowrate of 0.835 kg/s, showed that the D50 of the overflow was 4.546 µm, while that of the underflowwas 12.68 µm. At this flow rate, the hydrocyclone appears capable of separating particles less than24.2 µm from the inlet stream and delivering these particles to the overflow. Considering that fouledsilt sediment ranges in size from 4–62 µm, the hydrocyclone would be suitable for environmentalcleanup of this form of contaminated sediment. The hydrocyclone will be used to study the viabil-ity of larger units for commercial applications. The research supports undergraduate education incontaminant remediation.

Introduction

While established in the mining industry for mineral separation, hydrocyclones offer potential forthe remediation and volume reduction of contaminated sediment found in dredging operations1.Bayo et al.2 demonstrated hydrocyclone technology to be an ideal resource for separating toxicsludge from industrial wastewater. Mansour-Geoffrion et al.3 conducted extensive experiments us-ing hydrocyclones to separate contaminated sludge in wastewater treatment plants where the sludgeconsisted of 15–45% suspended solids. The technology if implemented may have substantially re-duced the mass of sediment shipped to the Saginaw Bay Confined Disposal Facility (CDF) duringthe recent dredging of the Saginaw River north of Independence Bridge in Bay City.

Dredging operations typically involve sediment that is at least partially submerged in water. Theinitial dredged material may have 10–15% solids concentration (high percent water) by weight4.Dewatering of the process stream is therefore essential for cost efficient transport of sediment to

Proceedings of the 2016 ASEE North Central Section ConferenceCopyright c© 2016, American Society for Engineering Education

1

Fig. 1. The fundamentals of hydrocyclone separation.

CDFs. Solids concentrations of 45–80% by weight can be achieved by hydrocyclone dewateringprocesses5. The technology offers low investment and operating costs as well as low maintenance. Itis presently being developed for a range of mineral dewatering applications6;7;8;9 and shows poten-tial for dewatering dredged materials10;11.

Hydrocyclones rely on a centrifugal process to separate the light and heavy solids, as shown in Fig.1. A high speed spiraling vortex accelerates the solids. The heavy solids are driven toward the wallby centrifugal force and descend to the lower outlet as underflow, while the light solids entrained inthe flow ascend to the upper outlet as overflow. The vortex finder prevents the heavy solids near thefeed inlet from entering the overflow stream. The length of the vortex finder can be increased to al-low more time for the heavy solids to be entrained in the underflow stream, increasing the separationefficiency of the unit.

The objective of the undergraduate student-led research project was to develop and test a hydrocy-clone that was designed and built with university and corporate support to separate particles in aliquid suspension.

Hydrocyclone Design and Sizing

The hydrocyclone design was based on the paper by Arterburn12. The D50C is defined as the par-ticle size of which 50% reports to the overflow and 50% to the underflow, as shown in Fig. 1. TheD50C(base) is the particle size that a standard cyclone can achieve operating under base conditions,and is given as

D50C(base) =D50C(application)

C1C2C3(1)

where C1 is the hydrocyclone feed concentration correction, C2 is the pressure drop correction, andC3 is the specific gravity correction, given as

C1 =

(53−V

53

)−1.43

(2)

C2 = 3.27∆P−0.28 (3)


2

TABLE ICHARACTERISTIC DIMENSIONS OF THE HYDROCYCLONE.

Parameter DimensionParticle size, µm 24.3Passing to overflow, % 80.0Multiplier 1.25D50C(application), µm 30.4C1 1.15C2 1.30C3 1.62D50C(base), µm 12.6Cylinder diameter, D, cm 9.56

TABLE IIEXPERIMENTAL PARAMETERS.

Parameter DimensionSolids concentration, % 5Solids density, kg/m3 1630Solids mass, kg 3.02Tank volume, m3 37.9Inlet flow rate, kg/s 0.835–1.13

C3 =

(1.65

GS −GL

)0.5

(4)

and where V is the percent solids by volume of hydrocyclone feed, ∆P is the pressure drop in unitsof kPa, GS is the specific gravity of solids, and GL is the specific gravity of liquid. The diameter andheight of a hydrocyclone chamber (units of centimeters) can be calculated from

D =

[D50C(base)

2.84

]1.515

(5)

H = D (6)

The hydrocyclone was sized to produce an overflow of 80% while passing a particle of 24.3 µm.The D50C(application) was calculated as 30.4 µm using the constants given in Table I. The pressuredrop across the hydrocyclone was determined to be approximately 6 psi from Fig. 9 of Arterburn12,by assuming a 20 GPM flow rate. The flow rate can be reasonably achieved in a laboratory environ-ment by recirculating fluid through a 50 gallon storage tank with a 1.5 hp motor/pump. The specificgravity of the solids was specified as that of silica sand. The characteristic dimensions of the hydro-cyclone and the experimental parameters are shown in Tables I and II, respectively.

The hydrocyclone was built at the university from 0.375 inch thick sheets of polycarbonate. Thepolycarbonate was machined in a Bridgeport CNC. The inlet scroll and cone of the hydrocyclonewere shaped by heating the polycarbonate in a convective oven and forming the material over man-drels. The fabricated hydrocyclone test apparatus is shown in Fig. 2.


3

Fig. 2. The hydrocyclone test apparatus.

Procedures

Before beginning the experiments the hydrocyclone, as shown in Fig. 2, was leveled by using the ad-justable pads as necessary. The apparatus was then flushed with clean water for at least five minutesto remove any residual silica and other contaminants from previous experiments. The reservoir wasfilled with ten gallons of water and 0.5 gallons of silica sand to achieve a five percent solids concen-tration by volume. Inside the reservoir of the apparatus, a mixing pump was switched on to agitatethe slurry. Also, the main pump that provided the means of fluid motion through the hydrocycloneitself was switched on. An ultrasonic mass flow meter was attached to the inlet line of the hydrocy-clone and an electronic pressure gauge was attached to the pressure ports of the inlet and outlet ofthe device. In order to achieve steady-state operation, both pumps were allowed to run for 30 min-utes with fluid passing through the hydrocyclone and recycle line for the duration. After steady statewas reached, the appropriate samples were taken from the inlet, underflow, and overflow, dependingon the specific needs of the experiment. For each sample acquisition, the mass flow rate, pressuredrop, and fluid velocity were recorded.

In order to conduct the sand sieve analyses of the samples, the water was first removed from thespecimens using a furnace at 90–120◦C. The samples were left in the furnace for several days untilall of the water had evaporated. Each specimen was subject to a standard foundry sand sieve analy-sis for 15 minutes. At the end of the 15 minute test, each sieve was removed from the testing appara-tus and the contents were weighed to produce a size distribution. More comprehensive distributionswere obtained through a contract with Particle Technology Labs (PTL). Samples were obtained fromthe hydrocyclone at two unique inlet flow rates and sent wet to PTL for processing.


4

Data and Results

The particle size distribution from the sand sieve analysis for a mass flow rate of 0.835 kg/s is shownin Fig. 3. The figure shows an increasing trend in percent by volume of silica sent to the overflowas particle size decreases. It can be seen that the cutoff diameter for the experiment was less than53 µm. The particle size distribution from the sand sieve analysis for a mass flow rate of 1.13 kg/sis shown in Fig. 4. As was shown in Fig. 3, an increasing trend is observed for percent by volumeof silica sand sent to the overflow as particle size decreases. Given the relatively low resolution ofthe sand sieve analyses, the cutoff diameters can be approximated as being less than 53 µm for theparameters of the experiments.

Fig. 3. Particle size distribution obtained from sand sieve analysis for mass flow rate of 0.835 kg/s.

Fig. 4. Particle size distribution obtained from sand sieve analysis for mass flow rate of 1.13 kg/s.


5

The particle size distribution from the PTL laser diffraction analysis at a mass flow rate of 0.835kg/s is shown in Fig. 5. The D50 of the overflow is 4.546 µm, while that of the underflow is 12.68µm. The particle size distribution at a mass flow rate of 1.13 kg/s is shown in Fig. 6. The D50 of theoverflow is 7.960 µm, while that of the underflow is 12.77 µm. The displacement of the D50 valuesfor the overflow and underflow provide a measure of the separation efficiency.

Fig. 5. Particle size distribution obtained from laser diffraction analysis for mass flow rate of 0.835 kg/s.

Fig. 6. Particle size distribution obtained from laser diffraction analysis for mass flow rate of 1.13 kg/s.


6

Discussion

Hydrocyclone separation performance can be explained by a simple force balance on individual par-ticles entering the device during steady-state operation. As the flow rate increases, larger particlesexperience more centripetal force inside the hydrocyclone, are forced to the wall, and eventuallyexit the device through the underflow. Smaller particles are pushed out the overflow due to the pres-sure distribution inside the hydrocyclone and the drag forces the particles experience as they flowthrough the slurry. This phenomenon was quantified during the experiment as the pressure differ-ential across the inlet and overflow outlet. As mass flow rate increased, so did the pressure differ-ential, and a larger volume of small particles experienced high drag forces inside the hydrocycloneand exited the overflow of the device. Figures 3 and 4 exemplify the physical phenomena showingthat as particle size decreased, an increasing volume percentage of particles was sent to the overflow.From the sand sieve analysis, it was shown that the hydrocyclone can successfully separate parti-cles 53 µm in size and smaller from a suspension. Considering that fouled silt sediment ranges insize from 4–62 µm, the hydrocyclone would be suitable for environmental cleanup of this form ofcontaminated sediment. The hydrocyclone will be used to study the viability of larger units for com-mercial applications.

The research grant provided funding for laser diffraction analysis to determine the separation ef-ficiency of the device more accurately. From the results of the PTL laser diffraction analysis at amass flow rate of 0.835 kg/s, the D50 of the overflow was determined to be 4.546 µm, while that ofthe underflow was 12.68 µm. At a mass flow rate of 1.13 kg/s, the D50 of the overflow climbed to7.960 µm, while that of the underflow remained relatively constant at 12.77 µm. The separation ef-ficiency is superior at the lower flow rate, as the D50 values are further apart. Furthermore, at a massflow rate of 0.835 kg/s, the hydrocyclone appears capable of separating particles less than 24.2 µmfrom the inlet stream and delivering them to the overflow, as shown in Fig. 5.

Conclusions

The study met the objective of the undergraduate student-led research project—to develop and test ahydrocyclone that was designed and built with university and corporate sponsorship to separate par-ticles in a liquid suspension. A thorough analysis of the hydrocyclone would not have been possiblewithout the continuous size distributions gained from the contracted laser diffraction analyses. Therewere several possible sources for error in the experiment. The most likely was the sand sieve anal-yses, as potential errors were magnified by the fact that the analyses were conducted using a largefoundry sieve with samples of relatively fine silica sand. The laser diffraction analysis results werefar more accurate. Additionally, the density of the slurry was approximated as the density of waterfor calculation purposes, but the presence of five percent silica sand by volume undoubtedly had aneffect on the density. The research project represents a step forward in the development of commer-cial opportunities for cleanup efforts throughout the Saginaw Bay Watershed.

Acknowledgments

The authors are grateful for the generous support provided by Saginaw Valley State University in theform of the Undergraduate Student-Led Research Grant.


7

Bibliography[1] R. E. Randall, A. C. Drake, and W. A. Cenac, “Improvements for dredging and dredged material handling,” in

Proceedings of the Western Dredging Association (WEDA XXXI) Technical Conference and Texas A&M University(TAMU 42) Dredging Seminar, Nashville, TN, 2011, pp. 104–122.

[2] J. Bayo, J. Lopez-Castellanos, R. Martınez-Garcıa, A. Alcolea, and C. Lardın, “Hydrocyclone as a cleaning devicefor anaerobic sludge digesters in a wastewater treatment plant,” Journal of Cleaner Production, vol. 87, pp. 550–557, 2015.

[3] M. Mansour-Geoffrion, P. L. Dold, D. Lamarre, A. Gadbois, S. Deleris, and Y. Comeau, “Characterizing hydrocy-clone performance for grit removal from wastewater treatment activated sludge plants,” Minerals Engineering, vol.23, no. 4, pp. 359–364, 2010.

[4] E. S. Gorleski and D. A. Gaffney, “Dewatering and amending dredged material for beneficial use,” in Innovationsin Grouting and Soil Improvement, chapter 16, pp. 1–9. 2005.

[5] T. J. Olin-Estes and M. R. Palermo, “Recovery of dredged material for beneficial use: the future role of physicalseparation processes,” Journal of Hazardous Materials, vol. 85, pp. 39–51, 2001.

[6] B. Firth, “Hydrocyclones in dewatering circuits,” Minerals Engineering, vol. 16, no. 2, pp. 115–120, 2003.

[7] G. V. Franks, P. D. Yates, N. W. A. Lambert, and G. J. Jameson, “Aggregate size and density after shearing, impli-cations for dewatering fine tailings with hydrocyclones,” International Journal of Mineral Processing, vol. 77, no.1, pp. 46–52, 2005.

[8] E. D. Galvez, R. Cruz, P. A. Robles, and L. A. Cisternas, “Optimization of dewatering systems for mineral process-ing,” Minerals Engineering, vol. 63, pp. 110–117, 2014.

[9] C. Wang, D. Harbottle, Q. Liu, and Z. Xu, “Current state of fine mineral tailings treatment: A critical review ontheory and practice,” Minerals Engineering, vol. 58, pp. 113–131, 2014.

[10] K. H. Park, J. H. Lee, B. H. Bae, Y. H. Kim, and Y. K. Choung, “Use of hydrocyclone and flotation column forreducing the volume of contaminated dredged material,” Water Science & Technology, vol. 53, no. 7, pp. 151–157,2006.

[11] H. P. Pandit, N. M. Shakya, H. Stole, and N. K. Garg, “Hydraulic and sediment removal performance of a modifiedhydrocyclone,” Minerals Engineering, vol. 22, no. 4, pp. 412–414, 2009.

[12] R. Arterburn, “The sizing and selection of hydrocyclones,” in Design and Installation of Comminution Circuits,pp. 592–607. Edited by A. L. Mular and G. J. Jergensen, Littleton, CO, SME, 1982.


8

hydrocyclone separation for the remediation of...

Documents