WATER QUALITY RESEARCH PROGRAM

let, INSTRUCTION REPORT W-91 -1

-~ DESIGN AND OPERATION OF AXIAL FLOW PUMPS

FOR RESERVOIR DESTRATIFICATION
___ by

Richard E. Punnett
DEP %RTMENT OF THE ARMY
om~ US Army Engineer District, Huntington
-v Huntington, West Virginia 25701-2070

U 2 21991A

Ol

-~ -July 1991
Final Reporto
kpQrcoved For Public ReleasL. Disltbulicn Unlimited

Prepared for DEPARTMENT OF THE ARMY

I US Army Corps of Engineers
Washington., DC e0314-1000C
Monmtored by Eflvironmentaf and Hydraulics Laboratories
US Army Engineer Waterways Experiment Station
~RA ~ 3909 Halls Ferry Road, Vicksburg, Mississippi 39180 6199

18 20 036
 Form Approved
REPORT DOCUMENTATION PAGE OMB No 0704-0188

Publc repo"tng burden for this c~lecton of information is es lated to average 1 hour oer resiorse, includtng the time #(,, reviewnig nstra Ons. earcnnq e-stnq data sources.
gathering and maintaining the data needed, and coml0eting and revieing the oliecton of nformation Send comments regarding this burden estimate or any other asaect of ths
coecton of informatiOn, ,ncluding suggestions for reducing this burden to Washtnh ton Headquarters Services. Oirectorate for information Ooerations and Reprts. 1215 jetferson
Davis HighwaV. Suite 1204. Arlington. VA 22202-4302. and tO the Office of Management and Budget. Paperwork Reduction Project (0704-0 188). Washington, DC 20503

1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
July 1991 Final report
4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
Design and Operation of Axial Flow Pumps for Reservoir
Destratification

6. AUTHOR(S)

Richard E. Punnett

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) B. PERFORMING ORGANIZATION

REPORT NUMBER

USAED, Huntington, Huntington, WV 25701-2070

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING

AGENCY REPORT NUMBER
US Army Corps of Engineers, Washington, DC 20314-1000
Instruction Report
Laboratories W-9 -i
USAEWES, Environmental and Hydraulics
3909 Halls Ferry Road, Vicksburg, MS 39180-6199
11. SUPPLEMENTARY NOTES
Available from National Technical Information Service, 5285 Port Royal Road,
Springfield, VA 22161

1Za. DISTRIBUTION /AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Approved for public release; distribution unlimited

13. ABSTRACT (Maximum 200 words)

This report presents information on design considerations, construction,
installation, and operation of axial flow pumps for either localized mixing or
lake destratification. When used for localized mixing, the pump displaces the
hypolimnetic water in the withdrawal zone of a low-level intake with epilimne-
tic water. When used for lake destratification, the pump moves the surface
water downward to mix with the bottom water and eliminate thermal stratifica-
tion. The pump consists of a frame, flotation platform, motor, gearbox, drive
shaft, bearings, and a large-diameter propeller. An example of the design pro-
cedure used for destratification of Beech Fork Lake is given.

14. SUBJECT TERMS 15. NUMBER OF PAGES

Axial flow pump Destratification 34
Beech Fork Lake Reservoir water quality 16. PRICE CODE

17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT
OF REPORT OF THIS PAGE OF ABSTRACT
UNCLASSIFIED UNCLASSIFIED I II
NSN 7540-01-28U-5500 Standard Form 298 (Rev 2-89)
bed by ANt Ild 149 II
P1e rcS
2118102
 PREFACE

The work reported herein was conducted as part of the Water Quality
Research Program (WQRP), under the work unit "Hydraulic and Pneumatic Mixers
and Aerators in Principle and Practice." The WQRP is sponsored by the Head-
quarters, US Army Corps of Engineers (HQUSACE), and is assigned to the US Army
Engineer Waterways Experiment Station (WES) under the purview of the Environ-
mental Laboratory (EL). Funding was provided under Department of the Army
Appropriation No. 96X3121, General Investigation. The WQRP is managed under
the Environmental Resources Research and Assistance Programs (ERRAP),
Mr. J. L. Decell, Manager. Mr. Robert C. Gunkel was Assistant Manager, ERRAP,
for the WQRP. Technical Monitors during this study were Mr. David Buelow,
Mr. James Gottesman, and Dr. John Bushman, HQUSACE.
The report was prepared by Dr. Richard E. Punnett, Chief of the Reser-
voir Control Section, Engineering Division, US Army Engineer District,
Huntington, Huntington, WV. Technical review of the report was provided by
Dr. James E. Garton, Professor Emeritus, Oklahoma State University.
The report was prepared under the supervision of Dr. Richard E. Price,
Reservoir Water Quality Branch (RWQB), Hydraulic Structures Division (HS),
Hydraulics Laboratory (HL), WES, and under the general supervision of
Dr. Jeffery P. Holland, Chief, RWQB, and Mr. Glenn A. Pickering, Chief, HS.
Mr. Frank A. Herrmann, Jr., was Chief, HL.
COL Larry B. Fulton, EN, was Commander and Director of WES. Technical
Director was Dr. Robert W. Whalin.

This report should be cited ar follows:

Punnett, Richard E. 1991. "Design and Operation of Axial Flow Pumps

for Reservoir Destratification," Instruction Report W-91-1, US Army
Engineer Waterways Experiment Station, Vicksburg, MS.

Accession For
NTIS GPA&I
DTI1 T.J0 F]
sow j i , , 01
oL_--_--_-

Dit Specia.l
 CONTENTS
Page
PREFACE.................................................................... 1
CONVERSION FACTORS, NON-SI TO SI (METRIC)
UNITS OF MEASUREMENT..................................................... 3
PART I: INTRODUCTION................................................... 4
PART 11: DESIGN OF AXIAL FLOW PUMPS..................................... 7
Theory of Design..................................................... 7
Sizing of Pumps..................................................... 9
PART III: CONSTRUCTION................................................... 12
Generalized Parts and Construction.................................. 12
Drive Train......................................................... 14
PART IV: INSTALLATION................................................... 16
Launching and Site Location......................................... 16
Anchoring........................................................... 16
PART V: OPERATION AND MAINTENANCE...................................... 18
PART VI: SUMMARY........................................................ 20
REFERENCES................................................................ 21
APPENDIX A: BEECH FORK LAKE PUMP DESIGN................................. Al
APPENDIX B: SPECIFICATIONS FOR AXIAL FLOW PUMPS......................... Bl

2
 CONVERSION FACTORS, NON-SI TO SI (METRIC)
UNITS OF MEASUREMENT

Non-SI units of measurement used in this report can be converted to SI

(metric) units as follows:

Multiply By To Obtain

acres 4,046.873 square meters

cubic feet 0.02831685 cubic meters
degrees (angle) 0.01745329 radians

feet 0.3048 meters

gallons (US liquid) 3.785412 liters

horsepower (550 foot-pounds

(force) per second) 745.6999 watts
inches 2.54 centimeters

3
 DESIGN AND OPERATION OF AXIAL FLOW PUMPS
FOR RESERVOIR DESTRATIFICATION

PART I: INTRODUCTION

1. As lake surfaces in temperate regions warm during spring and summer,

lakes become thermally stratified. The resulting stratification can dramati-
cally affect the water quality in eutrophic lakes. Thermal stratification is

not of itself undesirable; however, without the benefit of surface to bottom

mixing, dissolved oxygen (DO) is often depleted in the lower layers of the

lake, and the water quality deteriorates. In the absence of DO, high concen-
trations of hydrogen sulfide, iron, manganese, and ammonia nitrogen often
persist. As temperatures moderate in the fall, the thermal differences in the
lake are reduced, and the lake eventually becomes isothermal (destratified).
When isothermal conditions exist, the lake water quality parameters are im-

proved as the lake naturally mixes and DO is increased throughout the water
column.

2. Lake destratification by man-made means has been demonstrated to be

effective in improving water quality (Quintero and Garton 1973; Steichen, Gar-
ton, and Rice 1974; Strecher 1976; Punnett 1978, 1988; Garton and Punnett

1980; Robinson, Carton, and Punnett 1982; Price and Sneed 1989). The two
principal methods of mixing to cause destratification are diffused-air pumping
and mechanical pumping. Diffused-air pumping is used for direct aeration of
the bottom waters and for inducing lake mixing by entraining bottom waters in
the rising bubbles. Mechanical pumping may be performed to pump the oxygen-

rich surface waters downward to mix with the lower lake levels. Both methods
can be used for either localized or whole lake destratification. Localized
destratification by mechanical mixing can be used to improve the outflow water

quality from a lake by displacing the bottom waters with surface water in the
vicinity of the intake structure, which withdraws predominantly from the
hypolimnion (Givens 1978; Moon, McLaughlin, and Moretti 1979; Busnaina,

Lilley, and Moretti 1981; Robinson, Garton, and Punnett 1982; Price and Sneed
1989). Laboratory tests indicated that a maximum of about 80 percent surface
water pumped downward can be released downstream from a bottom intake using

localized mixing.
3. For a particular application, the best method can be selected after

4
evaluating the following considerations: cost, operation and maintenance,
lake depth, degree of stratification, and areal extent of destratification.
Mechanical pumping systems are often cheaper than diffused-air pumping sys-
tems. Mechanical pumping is generally inexpensive in daily operation and can
be easily maintained. However, mechanical pumping is often not feasible in
deep lakes that have a strong stratification pattern because of buoyant forces
that inhibit the downward penetration of the lighter surface water.
4.A specific type of mechanical pump that has been applied success-
fully to mechanical pumping is the axial flow pump, often referred to as the
Garton pump. The axial flow pump consists basically of a frame, flotation
platform, motor, gearbox, drive shaft, and propeller (Figure 1). The pump is

designed for moving large volumes of water with a low power input. A propel-
ler, such as a cooling tower fan (Figure 2), is suspended below the water
surface and rotated to pump surface water downward. Even though a low-head,
low-velocity jet is produced, the large-diameter propellers (up to 5.2 m) pump
3 The design of such a system is
a large volume of water (up to 4.5 m /sec).
discussed in the next section.

5
 Figure 1C arton pumnp

Figure 2.Pump propeller

6
 PART II: DESIGN OF AXIAL FLOW PUMPS

Theory of Design

5. As surface water is pumped downward into the hypolimnion, a plume of

warm (light-density) water is formed within the lower layers of cold (heavy-
density) water. Buoyant forces acting upon the plume impede the downward
velocity of the plume until a relatively stable mixing depth is established.
The plume must penetrate to the desired depth in order to be effective. The

desired depth would be the lake bottom if lake destratification is the objec-

tive or to the intake invert if localized destratification is the objective.

Several equations have been developed by Punnett to predict the depth of plume
penetration for axial flow pumps (Punnett 1984), the best nondimensional form
of which is

= 176 V + 0.756 . (1)

D g(Ap/p 0 ) D

where
Hv = length of plume, m

D = pump diameter, m
V = initial jet velocity, m/sec
g = gravitational constant (9.81 m/sec2 )

Ap = difference in density between surface and

3
desired depth of penetration, kg/m
p. - average density of pumped water
He = length from pump to thermocline, m

6. In Equation 1, the first term on the right side accounts for plume

penetration into dissimilar density strata. The second term on the right side
accounts for penetration within the epilimnion where little buoyant resistance
is encountered. The depth of the thermocline (for the above equation) was

considered to be from the pump propeller to the depth at which the first major

increase in density (or temperature) occurred. In a case where no apparent

thermocline exists but there is a thermal gradient, the midpoint between tne
pump and the desired depth of penetration should be used. Other predictive
equations have been developed (Holland 1984, Punnett 1984) for surface pumps;

7
however, Equation 1 was derived from field tests specifically designed to
determine the best penetration equation for axial flow pumps.
7. The values for density can be obtained from a lake temperature pro-
file and water density tables. In the absence of chemical-density gradients,
the density of water ( p , in kilograms per cubic meter) can be calculated
using Equation 2 (Ford 1983), and the average density of pumped water (po) can
be calculated using Equation 3:

P = 1,000 - (T - 3.98)2 (T + 283) (2)

(503.57) (T + 67.26)

where T is the temperature of water (0C),

and

B + P1 + P2 )
(PP (3)
3

where
p. = density of water at the surface
p, = density of water at 1 m below the surface
P2 = density of water at 2 m below the surface

8. Early work (Robinson, Garton, and Punnett 1982; Steichen, Garton,

and Rice 1974) concluded that the following fan laws provide an effective
means of predicting the pump performance in water from the available data
based on air tests for a constant diameter:

Q1 _ N1 (4)
Q 2 N2

8
(5)
Nz2PI

-m - mmmam
--- H ram m IN ~ i n
where
Q - pump flow rate
N - rotative Pneed, rpm
P - blade input power, kw

Note:
1 - air
2 - water

3
= 1.22 x 10-
P2

9. Using Equations 4 and 5 and a manufacturer's propeller performance

curve developed for a specific propeller in air, the performance in watel 2an
be determined. The power ir Equation 5 represents the power required by the
propeller; motor sizing should also include calculations of power losses in
the bearings and gearbox. An example application of the above equations is
given in Appendix A.

Sizing of Pumps

10. For use as a design factor, Equation I can be used to solve for the
design velocity assuming a pump diameter. For large-diameter propellers
normally manufactured for operation in air, the maximum design velocity should
be less than about 1.0 m/sec because the propeller hib is designed for
stresses related to high-speed, low-resistance operation. A good target
velocity is about 0.5 to 0.8 m/sec. A propeller designed for operating in
water, such as a ship's propeller, can be operated at higher velocities than
the target velocity.
11. The pump diameter and number of pumps required are determined by
considering the flow rate needed to be pumped. The flow rate ( Q , in meters
per second) is calculated by

Q = 0.785D 2V (6)

12. Because the pump velocity has important design limitations, and
because only discrete pump diameters are available, there is no

9
straightforward approach to solving for the velocity and diameter simultane-
ously. However, an iterative process will quickly lead to a solution. Pro-
peller diameters of 1.22, 1.83, and 2.44 m are generally available from manu-

facturers of cooling tower fans. Assuming a propeller diameter and using

Equation 1 to determine the required velocity, Equation 6 can then be used to

determine the flow rate. If the required flow rate is greater than what the
2.44-m pump will yield, multiple pumps generally will be required. Use of
multiple pumps also allows for more operating options and, if required, per-
mits pump maintenance without complete shutdown. Aircraft propellers have
been used for pumps having a propeller diameter of 5.2 m.

13. If localized destratification is desired for the purpose of improv-

ing the outflow water quality, the pumping flow rate is a critical design

parameter as well as the depth of penetration. Too little flow will not pro-
duce the desired results; too much flow may cause a greater mixing action,
which can give less than maximum benefits. Some site-specific tests have
shown that the maximum benefit is achieved by pumping about half of the
release rate (Robinson, Garton, and Punnett 1982).

14. If lake destratification is desired, an evaluation of the required

flow rate is difficult. Factors such as wind action, basin morphometry, size
and shape of the lake, volume of hypolimnion, degree of stratification, time
of year, and pump location have major influences on the required pumping rate.

The pumping rate in early summer is much greater than that required in late
summer because the larger influx of heat attempts to restratify the impound-

ment. As pumping occurs, the warm water being forced downward mixes with the

cold bottom waters. The mixed water, which has an intermediate temperature
and density, rises to the depth of approximately the thermocline (or to a
depth of equal density). As pumping continues, the zone of mixed water

spreads horizontally and begins to widen vertically in a "lens" fashion. The

rate of spreading may be limited by the pumping rate or by the hydrodynamic

and buoyant forces associated with mixing water of dissimilar densities.

Pumping a flow rate that is too low will limit the amount of water available
for mixing. Pumping a flow rate that is too high may set up a recirculating

cell around the pump, resulting in excess operating costs. Ideally, the
pumping rate would equal the maximum rate at which the mixed buoyant pumping

plume would spread throughout the lake.

15 . In liam's Lake (40 ha, surface) , Okl ahoma, a 1.83-m-diam pump

10
completely destratified the lake within a 1-week period (Quintero and Garton
1973; Steichen, Garton, and Rice 1974; Strecher 1976). The l.l-kW pump was
placed near the middle of the lake, and its flow rate was about equal to pump-
ing the volume of the hypolimnion once every 4 days. However, successful
destratification was achieved using smaller pumps that pumped the volume of
the hypolimnion once every 8 days. In Beech Fork Lake (291 ha), West
Virginia, the same ratio (volume of the hypolimnion pumped in 8 days) of pump-
ing did not have the same success (Punnett 1988). Beech Fork Lake has a
bifurcated shape, and the pumps were placed within 30 m of the dam. Unlike
the Ham's Lake project, the pumps were not located in the middle of Beech Fork
Lake, which meant the mixed lens of water could not spread in a full radial
fashion. Unpublished observations by the author indicate that the pumping
rate was too great for the rate at which the mixed water would spread through-
out the lake; thus, much of the pumped water was being recirculated in a
localized cell around the pumps. Although Beech Fork Lake was not completely
destratified until the heat input moderated in late August, the thermocline
was lowered throughout the lake and the temperature difference within the lake
was reduced to about 30 C for most of the summer.

11
 PART III: CONSTRUCTION

Generalized Parts and Construction

16. The construction of an axial flow pump requires only basic shop
functions, but machining of the drive shaft ends is sometimes required. The
pumps can be constructed using a welded metal frame, flotation platform,

motor, right-angle gearbox, drive shaft with couplings, bearings, and a

propeller. The axial flow pump has been constructed in-house as well as

commercially. Many pumps have been constructed by college students in a

university shop. From a generalized parts list and a hand sketch, a machine
shop in Point Pleasant, WV, constructed the main components for four axial
flow pumps (l.83-m-diam with a single hermetic steel tank for flotation) for
about $24,000 in 1986 (Figures 3-5). The pump sizing calculations, actual
design specifications, generalized parts list, and sketch used by the machine

shop are given in Appendix A. E. C. Baker & Sons, Inc., Sigel, IL, markets
axial flow pumps that are completely equipped for about $11,000; delivery and
installation are available at additional cost.
17. The flotation platform can be made either with foam blocks or with
hermetic containers. If hermetic containers are used, some flotation material
should be put inside, in case the containers develop a leak. After estimating
the weight of the pump, the platform should be designed to float the pump with

Figure
i i i
3.
i
Two assembled pumps 01

12
 Figure 4. Two pumps in operation

Figure 5. Pump gearbox, motor, and switch

13
about 0.5 m freeboard. If flotation blocks are used, protection from
waterfowl may be necessary.

Drive Train

18. Both diesel and gasoline engines have been used to power the pumps
where electrical connections were not feasible; however, electrical motors are

much easier to operate. The use of fuel not only dramatically increases
operational expenses and hardships, but can deteriorate some flotation mate-

rials and create environmental concerns.

19. If an electric motor is used, the electrical connections consist of

individual switch boxes for each motor as well as a central starter switch
station for a cluster of pumps. The motor selection should include adverse-
environment casing (specified TEFC, total enclosed fan cooled) and be con-

sistent with the available power source. In case of power failure, a manual
starter switch can prevent multiple pumps from automatically starting simulta-
neously. Delay circuits are available which allow the pumps to restart

sequentially, thus avoiding a possible overloading of the power lines. Gener-

ally, the power line runs from the source onshore, along the lake bottom, and

up to the pumps.
20. The gearbox should be designed for continuous operation under ad-

verse environmental conditions. A high-quality, heat-resistant gear oil

should be used. The reduction ratio of the gearbox is dependent upon the
input speed of the motor and the required drive shaft speed of the propeller.
The gearbox should be mounted on an elevated frame above the platform surface
so that the coupling joining the drive shaft is serviceable from the top of

the flotation platform. This is a critical consideration if a gearbox

requires replacement while the pump is anchored in the lake.

21. A suitable drive shaft material is cold-rolled steel. Stainless
steel shafts are unnecessarily expensive. A stress analysis should be made to
determine the necessary diameter. A 3.8-cm-diam shaft has been used success-
fully for a 1.83-m propeller, and a 5.l-cm-diam shaft has been used success-
fully for a 2.44-m propeller. The shaft length should allow the propeller to

be suspended about 1.5 to 2.0 m below the water surface for propellers less
than 2.5 m. The ends of the shaft may require machining to attach the coupler

and propeller. Key slots are often required. If a rigid coupler is used, the

14
gearbox can generally handle both the weight of a suspended propeller and the
upward thrust forces when operating. If a flexible coupling is used, shaft
bearings (or bushings) will need to handle the vertical forces as well as
stabilize the shaft horizontally.
22. The actual blade configuration (shape and number) of the propeller
seems to be of little consequence to the penetration performance of the pump.
Aerovent, Inc., Piqua, OH, manufactures six-bladed cooling tower fans that
perform well as propellers for an axial flow pump (Figure 2). The propeller
pitch is adjustable so that the flow rate can be set for a given rate of
rotation. The blades are also reversible. Care should be given to ensure a
proper blade setting; the blade angle should be greater at the hub than at the
blade tips. An improperly installed blade produces unpredictable pumping
results. A propeller shroud, designed to improve pumping efficiency by
guiding the flow and reducing entrance losses, is optional.
23. Once all the material and parts are gathered, approximately 80 man-
hours (engineer and/or technician) is required for construction of the major
components. If a pump is built on contract and is no greater than 2.44 m wide
(highway limitation), the major components should be assembled by the contrac-
tor and delivered as a complete unit. If a pump is larger than 2.44 m, deliv-
ery of the components and site assembly may be best.
24. Where public access could be a problem, fences have been installed
around the flotation platform. To prevent debris from entering the propel-
lers, fencing has been installed below the flotation platform. Warning signs,
indicating high voltage, have been used. Yellow flashing lights have been
installed where a potential nighttime boating hazard existed. High-
performance epoxy base paints have been successful in preventing corrosion of
metal parts. Wheels have been installed on some pumps to facilitate loading
from a boat launch Area.

15
 PART IV: INSTALLATION

Launching and Site Location

25. With the possible exception of fencing, the pumps should be complete

and ready for operation before installation since most tasks are more diffi-
cult while the pumps are floating. The most efficient method of installation
is to place the assembled pump in the lake using a crane. If wheels have been
installed on the base of the pump frame, a cable to control the rate of

descent down the ramp should be attached as low as possible on the frame (near
the wheels). Because of the deep draft of the pump, the ramp depth should be
checked for suitability. Loading from a boat launch area will require person-
nel who are ready and able to skin dive. After a pump is in the lake, it can
be easily pushed by a small boat (in barge fashion) to the pumping site.

Generally, three people are able to install one pump in about 2 hr.
26. The location of the pump site is an important issue that involves
not only lake mixing considerations, but also common logistics such as the

availability of power, potential boating hazards, likelihood of vandalism, and

access for maintenance. The best location, for lake mixing concerns only, is
over the deepest site that is centrally located. In lakes with extensive

dendritic patterns or where the hypolimnion may be partitioned by submerged

topography, several locations may need to be evaluated. For localized mixing,
the pump should be placed just upstream of the intake port. Pump location and
configuration (Price and Sneed 1989) for a three-pump, localized destratifica-
tion operation at J. Percy Priest Reservoir indicated the location in front of
the intake can have a significant influence on system efficiency.

Anchoring

27. Generally, pumps are anchored in position, but some have been
secured to an existing structure. The two main concerns are the ability to

accommodate changes in surface elevations and to compensate for induced

torque. In the case of anchoring, some slack should be left in the anchor

lines to accommodate expected lake rises. The torque induced by the turning
pLopelier will cause slack lines to partially wrap around the pump (or cluster

of pumps) until all lines are taut. This is a useful side effect since the

16
pump(s) will "unwind" as the lake level increases and "rewind" as the lake

lowers, thereby maintaining position. In the Beech Fork application (Punnett

1988), a cluster of four 1.83-m-diam pumps was held in position by four an-
chors and 0.635-cm steel cables. The anchors were made by cutting 55-gal*
drums in half and filling with concrete. A short piece of heavy-gauge chain
was set in the concrete so that the cable was easily attached. The anchors
were positioned off the corners of the square cluster of pumps far enough away

so that the taut cables formed a 45-deg angle with the lake surface.

28. The total weight of the anchors should not exceed the floating
capacity of the platform in case of extreme lake rises. If it appears that
the anchors may be lifted because of an imminent extreme rise, the pump(s)
should be shut off. An auxiliary anchor with a longer cable could hold the

pumps near the site if it is important to maintain the relative position.

* A table of factors for converting non-SI units of measurement to SI

(metric) units is presented on page 3.

17
 PART V: OPERATION AND MAINTENANCE

29. For localized destratification, the pumping activity should commence

with declining water quality conditions in the lower strata of the lake and
coincide with the release schedule. One of the benefits of localized mixing
is that the pumps need only to be operated during a period of hypolimnetic

releases (e.g., hydropower generation or summer flood control).

30. For lake destratification programs, significanu stratification can

be prevented in lakes if pumping begins early, thereby possibly avoiding

anoxic conditions in the hypolimnion. In the late spring and early summer,
the lake surface temperatures may increase rapidly, and the full pumping
capacity should be used. Although the system is designed to pump the volume
of the hypolimnion in a given time, full pumping capacity is not needed until

the maximum heat loading to the lake occurs. If start-up of the system is
delayed or solar input is much more than design condition, complete destrati-
fication may not be achieved.

31. The effect of pumping is to warm bottom waters to surface tempera-

tures. There may be a slight cooling of surface temperature, but it does not
appear to be significant. As the heat input moderates (about mid-August),

only a minimal amount of pumping is required to maintain isothermal condi-

tions. The pumping schedule depends primarily upon weather conditions. By
early September, pumping may not be needed even though the lake would normally
stay stratified until November.
32. The success of the pumping effort can be easily determined using
temperature and DO profiles. Profiles of temperature and DO, taken immedi-

ately beside the pump(s), will reveal whether the pump jet is penetrating to

the proper depth. The temperature in the pumped plume will be relatively
constant (at surface values) until the depth of penetration is reached. At
that point, readings become erratic; readings below the penetration depth
become stable at a colder temperature. Changes in release water temperature

and DO, as monitored downstream, will quickly indicate the success of a

localized pumping effort.

33. In the case of whole lake destratification where pumping begins

early (e.g., April), weekly profiles are usually sufficient for monitoring

progress. If pumping commences after stratification has been established,

profiling every other day would be important until nearly isothermal

18
conditions prevail; then, weekly profiles become sufficient. To identify all
the changes caused by lake destratification, an expensive and intensive pro-
gram is needed; however, to assess the success of the pumping effort, tempera-
ture and DO are usually sufficient indicators. A program to identify all
changes would include a study, both inlake and downstream, of benthos,
plankton, chemistry, and fish.
34. For localized destratification, three sampling stations may be suf-
ficient to monitor the pumping program: one station upstream of the dam to
monitor lake profile conditions, one station within the pump plume, and one
station immediately downstream of the dam. For lake destratification, several
stations should be considered in addition to those identified for localized
destratification. The shape and bottom contour of the lake are important
considerations in determining sampling station locations. Stations located in
the thalweg will usually give the best indication of destratification success.
35. Maintenance needs of the axial flow pump, with an electric motor,
are relatively minor. An axial flow pump driven by a fuel engine will be
essentially as reliable as the engine. The gearbox requires an occasional
oil-level check, about every 2 months of operation. An oil change after a
specified operation time, such as once a year under continuous operation, is
recommended. The gear oil should be of high quality and heat resistant; in
some cases, oils have "baked" and allowed gear failure. In gearboxes, a brass
gear might require replacement if worn excessively. The gears should be
checked at least once a year. Pumps have been left in the water for 3 years
without problems. A good maintenance plan would require pump removal every
other year for inspection, cleaning, and repainting. The expected useful life
of a pump with proper maintenance would be limited by the life of the gear box
and motor, assuming rust problems do not develop. Based on previous destrati-
fication projects, 5 to 10 years of operation can be anticipated. If fencing
was used below the water surface, replacement of the fencing each year may be
necessary due to corrosion. If a pump is removed from the lake for winter
storage, the anchor cables can be attached to a single buoy and left in place.

19
 PART VI: SUMMARY

36. This report discusses design and construction methods for axial flow
pumps used for localized mixing and lake destratification. The theory of

design along with computation procedures for sizing of pumps is given. Con-
struction methods, including materials and parts, are discussed. The propel-

lers used in the design are cooling tower fans with variable pitch to achieve
a desired flow rate. Installation information includes location of the pumps
on the reservoir and anchoring techniques. A section on operation and mainte-

nance gives techniques for monitoring the success of a localized mixing or

lake destratification project. Maintenance needs, although minimal, are based

on operation of the pump during the stratified season. The design computa-
tions used for construction of the Beech Fork Lake destratification system are
given in Appendix A. Specifications for axial flow pumps are presented in

Appendix B.

20
 REFERENCES

Busnaina, A. A., Lilley, D. G., and Moretti, P. M. 1981. "Prediction of

Local Destratification of Lakes," Journal of the Hydraulics Division, ASCE.
Vol 107, No. HY3.
Ford, D. E. 1983. "An Assessment of Reservoir Density Currents and Inflow
Processes," Technical Report E-83-7, US Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Garton, J. E., and Punnett, R. E. 1980 (May). "Quality Improvement of
Releases from Reservoirs," Proceedings, Surface Water Impoundments Symposium,
Minneapolis, MN, Paper No. 5-7.
Givens, M. R. 1978. "Hydraulic Modeling of Local Destratification of Lakes
Using Propeller Pumps," M. S. thesis, Oklahoma State University, Stillwater,
OK.
Holland, J. P. 1984. "Parametric Investigation of Localized Mixing in Reser-
voirs," Technical Report E-84-7, US Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Moon, J. J., McLaughlin, D. K., and More ti, P. M. 1979 (Sept). "Enhancement
of Reservoir Release Water Quality by Localized Mixing," Final Report,
Phase 1, Contract DACW39-78-C-0045, US Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Price, R. E., and Sneed, R. B. 1989 (Aug). "Localized Mixing for Improvement
of Reservoir Releases from J. Percy Priest Reservoir," Proceedings, ASCE Con-
ference on Hydraulic Engineering, New Orleans, LA.

Punnett, R. E. 1978 (Jul). "Destratification of Lake Arbuckle with a Cluster

of Low Energy Pumps," M. S. thesis, Oklahoma State University, Stillwater, OK.

. 1984 (Dec). "Predicting the Depth of Penetration of a Buoyant

Plume in a Stratified Lake," Ph. D. thesis, Oklahoma State University,
Stillwater, OK.
__ 1988. "Destratification of Beech Fork Lake," Proceedings. Water
Ouality '88 Seminar, Charleston, South Carolina, US Army Corps of Engineers,
Committee on Water Quality, Washington, DC.
Quintero, J. E., and Garton, J. E. 1973. "A Low Energy Lake Destratifier,"
Transactions, American Society of Agricultural Engineers, Vol 16, No. 5.
Robinson, K. M., Garton, J. E., and Punnett, R. E. 1982. "Localized Destrat-
ification at Lake Texoma," Journal of the Environmental Engineering Division,
ASCE Vol 108, No. EE4.
Steichen, J. M., Garton, J. E., and Rice, C. E. 1974. "The Effect of Lake
Destratification on Water Quality Parameters," Paper No. 74-5008, American
Society of Agricultural Engineers, St. Joseph, MI.
Strecher, R. G. 1976. "Design, Construction, and Evaluation of a Prototype
Low-Energy Lake Destratifier," M. S. thesis, Oklahoma State University,
Stillwater, OK.

21
 APPENDIX A: BEECH FORK LAKE PUMP DESIGN

Introduction

1. Beech Fork Lake is located on a tributary of Twelvepole Creek in

northwestern West Virginia. This 750-acre lake, with a maximum depth of
35 ft, is used for recreation and fish and wildlife enhancement. Thermal
stratification during the summer results in a shallow epilimnion (about 2 m)
and anoxic conditions in the hypolimnion. This limits habitat available to

the fishery, as well as biological productivity. To increase the habitat

available to the fishery, a destratification project was initiated in 1987.

Destratification Objective

2. The major objective of the destratification was to increase the

depth of the epilimnion and thereby increase the available habitat for the
fishery. This was accomplished using four Carton-type pumps to mix the lake.

These pumps were operated to pump epilimnetic water through the thermocline
into the hypolimnion. The epilimnetic water mixed with the hypolimnetic water
to produce a volume of water with a temperature near the thermocline tempera-
ture. Thus, the mixed water moved throughout the lake as a layer in the
thermocline region. As pumping continued, this layer increased in length and

thickness until both the warm and cold water were mixed and the lake was
isothermal.

Pump Design

3. The design of the Carton pumps was accomplished using the design
equations and guidance provided in the main text. Pertinent data from Beech

Fork Lake are given in Table Al.

Al
 Table Al
Pertinent Data, Beech Fork Lake, West Virginia

Parameter Value
Surface area 293 ha
Maximum depth 10.7 m
Extreme thermal conditions
Surface 240 C
I m below surface 220 C
2 m below surface 210 C
10.7 m (bottom) 140 C
Minimum depth to top of thermocline 2.1 m
Volume of anoxic hypolimnion 6.0 x 106 m 3

4. Using Equations 2 and 3 of the main text, the densities associated

with the thermal stratification are

3
P24 = 997.3 kg/m
3
P2Z = 997.8 kg/m
3
P21 = 998.0 kg/m
3
P14 = 999.3 kg/m

=
Po (997.3 + 997.8 + 998.0) = 997.7 kg/M 3
3

Ap _ (999.3 - 997.7) - 0.0016

P0 997.7

5. For whole lake destratificatior, the pump plume should penetrate

to the lake bottom at the pump site. Assuming a propeller diameter of 1.83 m,
Equation 1 is solved for the velocity:

10.7 m _ 0.176V 2 + 0.756(2.1 m)

9.81(0.00167 1.83 m

A2
Therefore,

V - 0.67 m/sec

6. Using Equation 6 to find the associated flow rate yields

Q = 0.785 (1.83)2 (0.67) = 1.76 m 3/sec

From the manufacturer's propeller performance curve (as shown in Figure Al),
for a 6-ft-diam propeller having six reversible blades, a blade pitch of
22 deg, and a rotation rate of 50 rpm, the flow rate is about 3,600 cfm at a
low head (static pressure). After conversion, the flow rate is about
3
1.70 m /sec. For this project, an electric motor with an input rotation rate
of about 1,750 rpm was used. To produce 50-rpm output to generate 1.70 m 3/sec
of flow, a gear ratio of 35:1 would be needed. Since a gearbox with a ratio
of 50:1 was readily available, the adjusted propeller performance was
calculated.
7. Using Equation 4 to determine the flow rate for the same propeller
at 35 rpm yields

1.70 m 3/sec _ 50 rpm

Q235rpm

therefore

Q2 = 1.19 m 3/sec

8. The blade power requirement is computed (assuming the density

ratio of air to water is 1.22 x 10-3), using the manufacturer's propeller
performance curve (Figure Al) and Equation 5:

0.0015 hp= (50 rpm) 3 1.22 X 10 -3

P2 (35 rpm) 3

A3
 Ib3M~d3SIH0H

G 0s

ZI
70
W 0
>D
0~0

0)

CIO

Z C)

W L 0
rro~ 0q -

cc 0

0(L

> 0

EID wn-I

QI x

A-4
Therefore,

P2 = 0.42 hp (or 0.31 kw)

Using the curves for specific blade pitch settings as provided by

9.
propeller at
the manufacturer (Figure Al) and the information above, the same
rpm:
different blade pitch settings has the following performance at 35

Flow Rate Velocity Blade Power

Blade Pitch
m3 /sec m/sec kw
deg
1.19 0.45 0.31
22
1.95 0.74 0.67
30
2.12 0.80 0.82
32

A5
 10. To determine the approximate flow rate necessary for whole lake
destratification, the normal volume of the anoxic hypolimnion should be used.

From the author's previous experience at Ham's Lake, the required flow rate
for destratification of the lake is about equal to pumping the volume of the

hypolimnion every 8 days. With a volume of the hypolimnion of approximately

3
6 x 106 m and a recommended destratification time of 8 days, a flow rate of
3
8.7 m /sec is required.

6 3
6 x 10 m x 1 l
1 day =8.7 m 3/sec
Y -d
ays 77 s ~~ec

11. Using the above performances of a 1.83-m propeller at 35 rpm,

4.4 pumps will be required at a blade angle of 30 deg; 4.1 pumps will be
required at 32 deg. The design specifications for four pumps having the

1.83-m propeller at 35 rpm were written. Appendix B provides copies of the

actual specifications, parts list, and drawing supplied to contractors for
bids and ultimately for construction of the pumps.

Conclusions

12. A detailed discussion of the destratification system is given by

Punnett (1988). Conclusions from that report were as follows:

a. The epilimnion was increased (the major objective).

b. The pumps were sufficient for destratifying Beech Fork Lake

even though a strong thermal-density difference existed prior
to the start of pumping.
C. Mixing occurred throughout the lake, even though the shape of
the lake did not appear to be suited to mixing and pumping
was conducted at only one location.

d. The water in the vicinity of the dam did not become anoxic.
Although at times the overall DO was low, only less than
1 percent of the lake volume became anoxic for a short
period.

A6
 APPENDIX B: SPECIFICATIONS FOR AXIAL FLOW PUMPS

General Description

1. These specifications are for the construction of four axial flow

(fan-type) pumps. Each pump consists of a flotation raft and support frame,
electric motor, right-angle drive gearbox, shaft and bearings, and a 6-ft-diam
fan (propeller). The motor and gearbox are mounted on the raft. The support
frame, suspended below the raft, is used to stabilize the shaft and propeller
as well as provide a base when the pump is on dry land. A sketch is provided
as Figure Bl; a generalized parts list is given in Table Bl.
MOO ANO STAND

i RAFT

,,[i' fFRAME

,,SHA r

PROPELLER N YPE)

Figure Bl. Axial flow pump

(shown without optional shroud)

Technical Description

2. The raft shall be constructed of pressure-treated wood (approved

by US Environmental Protection Agency for water use) or a steel hermetic con-
tainer, with at least a 6.5-ft square deck. The raft shall be supported with
sufficient flotation to give the pump a freeboard of about 18 in. above the
waterline. The flotation material, if used, shall be gas and oil proof, and
enclosed for protection from waterfowl. The frame shall be constructed of

BI
metal and painted with high-performance epoxy paint. The electric motor shall

be a 3-hp, 3-phase, 240-volt, TEFC (total enclosed fan cooled) motor. The
motor shall be mounted directly on the gearbox. The right-angle gearbox shall
have a 50:1 reduction and be rated for continuous operation. The shaft shall
be made of 2-in.-diam, cold-rolled steel (about 8 ft long), and painted with a
high-performance epoxy paint. The coupler that connects the gearbox output

shaft to the propeller shaft can be either a steel sleeve type or a flexible,
chain-type coupler. The propeller shall be six-bladed, 6 ft in diameter, with
a variable pitch (e.g. Model 72R6xx from Aerovent, Inc., Piqua, OH). The
propeller shall be suspended about 6 ft below the water surface and suffi-

ciently supported with guide and thrust bearings to be stable under full speed
(about 35 rpm). The pump will be used in a lake and subject to weather;
therefore, the construction and all components shall be compatible with an
adverse environment.

B2
 Table BI
Generalized Parts List

Part Description
Frame Metal, designed to stand on dry land as
well as provide a stable support for the
shaft and propeller while operating. The
metal should be painted with a high-
performance epoxy paint unless made of
corrosion-resistant metal.
Flotation Hermetic container or styrofoam (or equiv-
alent) that is gas and oil resistant and
enclosed for protection from waterfowl.
Motor 3-hp, 3-phase, 240-volt, weatherproof,
continuous operation.
Gearbox Right-angle drive (horizontal input, down-
ward output), 5-hp input, weatherproof,
continuous operation, 50:1 reduction,
motor-mount flange and coupling.
Coupler Steel sleeve or flexible, chain-type.
Shaft 2-in.-diam, about 8 ft long, cold-rolled
steel, high-performance epoxy coating
(painted).
Bearings For 2-in.-diam shaft, submersible, with
thrust bearings if needed.
Propeller Fan-type, 6-ft-diam, six-bladed, adjust-
able pitch (e.g. Model 72R6xx from
Aerovent, Inc., Piqua, OH). Shroud
optional.

B3