|
Page 3 of 3
Caveats
There are some caveats that must be heeded for a trench-type wet well to be completely successful.
Caveat 1. Flow Splitter on Ramp
To retain the energy developed by the ramp and to obtain a swift flow of water along the floor during cleaning (as in Figure 5), the flow splitter must begin at the top of the ramp and continue without interruption to some point between the last two pumps. (See the "Model" section for an extensive discussion of flow splitters.) At super-critical velocities, any type of obstruction saps energy, may send a jet of water flying and certainly reduces cleaning capability. If the flow splitter begins at the base of the ramp, the water-now at very high velocity-strikes this obstruction, jumps off the floor, probably impacts the first pump bell, loses nearly all of its energy and never regains it. Quick cleaning is prevented. The trench can still be cleaned, but now it is by turbulence and attrition only. That process is slow, and the pump must run for many minutes under severe conditions.
As with flow splitters, the fillets also must begin at the top of the ramp. They should extend to the back wall to prevent the formation of side wall vortices at the last pump.
Caveat 2. Water Guide
If the water from the inlet pipe is allowed to spread wider than the trench itself, some of it runs up on the sloping wall above the trench, where it slows and then falls back into the main stream at low velocity and disrupts the main flow. To prevent this occurrence, raise the sides of the trench near the top of the ramp to form a "water guide" that keeps all the water confined to the width of the trench. Water may run up on the fillets, but they are small, and the deleterious effect is not pronounced.
Caveat 3. Elevation of last pump intake
Pumps in confined trenches often lose prime when the suction bell submergence is less than D/2. So place the bell rim at least D/2 below the sequent depth (given in Figure 5), and drop the floor under it to give a floor clearance of D/4. At greater floor clearances, larger grit and small stones may not be picked up by inlet velocities around 4-ft/s.
ANSI/I 9.8 allow inlet velocities between 3- and 8-ft/s with a recommendation of 5.5-ft/s for medium-size pumping stations. One manufacturer designs solids handling column pumps for 3.5- to about 5-ft/s, and the authors recommend these lower velocities to prevent large, heavy trash such as parts of bricks or concrete blocks from being sucked into the pump and damaging it. (Include a rock trap upstream or just remove large, heavy trash manually from the trench as necessary.) Submersible pumps are designed for very high inlet velocities, so if the last pump is a submersible, add a suction nozzle with a flare to reduce the inlet velocity to 5-ft/s or less.
Caveat 4. Cone Under Last Pump
Because the last pump has a floor clearance of only D/4 (so as to scour grit effectively) and any flow splitter must clear the rim by 3-in to pass solids, there is not enough room for an effective flow splitter in moderate-size pumping stations. Consequently, a cone under the last suction bell is a logical substitute for the flow splitter (see Figure 7).
During pump-down for cleaning, water has a strong tendency to circulate between the last pump and the end wall. That circulation results in an upstream current on one side of the trench that keeps the hydraulic jump far upstream. The circulation can be so strong that it can go under the suction bell (even though the pump is running) and travel upstream. To prevent this occurrence, a fore-and-aft vane on the cone is needed (as shown in the figure below).
Caveat 5. Anti-Rotation Baffle
The tendency of water to circulate behind the last pump was described in Caveat 4. To complete the suppression of flow between pump and end wall, an anti-rotation baffle (or barrier) is also required. Allow a minimum of construction clearance between the pump column and the baffle, but note that column pumps move slightly when pumping. Consult the manufacturer for details on how much to allow for that movement.
Caveat 6. End Wall
During normal pumping, the stagnant water behind the last pump tends to form a surface vortex. Such vortices take the form of the letter "J". If the wall is moved close to the pump (ANSI/HI 9.8 recommends 0.75 D from the pump centerline), it intersects the "J" and the vortex does not organize. Sloping back walls outward allows the creation of a vortex.
Forming Flow Splitters and Fillets
In the early stages of development, designing a flow splitter or fillets on a curving ramp was not an easy task with either models or prototypes.
Models
After a model test in a commercial laboratory showed that a flow splitter merging into the bottom of a ramp largely destroyed the energy of the flowing water and caused downstream flow to become subcritical, a machinist at Montana State University suggested making the curved portions of flow splitters and fillets of a two-component casting compound [6]. The one chosen is very strong but flexible.
Pour a slight excess into a wooden mold coated with floor wax. If the nose is gently tapered, one end of the mold can be suitably tapered to form a continuous nose. As soon as the compound cures enough to allow it, trim off the excess with a sharp knife. After final cure, warm the flow splitter with a bathroom heater and bend it over curves cut into wood with a band saw. The compound takes a permanent set upon cooling, but it does relax a little, so make the curves a little (say, 10 percent) sharper than the ramp curves. Fasten the pieces to the ramp with rubber cement, screws or both. Straight sections of fillets and flow splitters can be more easily made of wood well painted to resist moisture change and warping.
With model flow splitter and fillets installed all the way up the ramp, the improvement in cleaning was dramatic. Velocities were very high at the foot of the ramp and remained high to the end of the trench. But when the flow splitter was changed to end at the toe of the ramp, the energy and high velocity were destroyed. The downstream flow was turbulent but of subcritical velocity that would not scour quickly. There is no comparison between the performances of the two designs. Therefore, both flow splitter and fillets must extend to the top of the ramp for adequate cleaning potential.
Other Means of Fabricating Model Flow Splitters and Fillets. Although casting compound is by far the most satisfactory material when installed, its use is involved and time-consuming. One way is to make the flow splitter and fillets entirely of wood. Cut it into thin cross-sections over curves, and set the sections in bathtub caulking compound such as Dap. After the Dap cures, coat the entire unit with Dap to waterproof it and fill the cracks. The result is ugly and rough but reasonably quick and easy.
Another quick, easy method is to make flow splitter and fillets of thin (40-mil) plastic. It can be cut with tin snips and has about the right amount of stiffness. The strips for a flow splitter can be joined at the apex and to the floor with either Dap or long, narrow strips of duct tape pressed down firmly. Supports for fillets can be triangular pieces of plastic or thin (1/8-in) balsa wood, which can also be cut with tin snips. These are glued to the plastic strip with Dap or super glue at intervals of 3 to 6-in as appropriate. These methods are illustrated in Figure 8.
 
Plastic sheets (for models and steel plates for prototypes) over ramp curves must be cut to the proper curvature. If the ramp curve has a radius of r, the formula for the radius, R, of a flat sheet to fit it is  where is the angle between ramp and sheet.
(1)
Flow Splitter Noses. A flow splitter must have some kind of nose. One of the easiest to make and best in performance tapers linearly from full size where it joins the prismatic flow splitter to zero at the top of the ramp. If ß is the angle between the centerline of the ramp and the contact between nose and ramp (see Plan in Figure 9), the formula for the radius of curvature becomes
(3)
The nose of a flow splitter separates the incoming flow into two streams that must have enough depth and velocity everywhere to wash debris off the ramp. Tests of inflow from horizontal pipe (low fluid velocity) and from approach pipe (high fluid velocity) were made at flow rates of 50 and 75 percent of the last pump's capacity. The best of several noses tried is shown in Figures 8b and 9. The apex angle (2 ) is constant and the nose height tapers linearly over a length of approximately 2D from the bottom of the upper ramp curve to the top of the ramp. A very long (4.3 D) nose was very good, and a very short (D long) one was adequate.
Other Model Materials. Still another method is to cut fillets and flow splitters from plastic foam with a knife, table saw or band saw, bend them to follow the contour of the ramp and fasten them with rubber cement or even duct tape. One suitable material is Ethafoam [7], a polystyrene closed-cell, relatively rigid, strong foam. It holds a curvature if warmed and bent while cooling.
Choice. If the model floor can be removed for convenience in constructing fillets and flow splitters, the use of thin plastic sheet material is as easy and quick as any method. If not, the use of foam on ramp curves and either foam or wood for straight sections is easier. Either kind of flow splitters and fillets is easy to remove and replace--an advantage for demonstrating their value.
Prototype Fillets and Flow splitters
Prototype fillets are easily made with shotcrete anchored into the corners with dowels (and with a rebar or two running the full length), screeded and troweled smooth. If greater smoothness is needed as in long wet wells, coat the concrete with epoxy or line it with PVC. See Figure 10.
Practically the only load on flow splitters is during construction when the pumpcrete can exert a pressure of 1- or 2-lb/in.2, so fastenings are needed primarily to hold the shape. Flow splitters can be made of 1/4-in or even (for very large ones) of 3/8-in stainless steel (ss) plate Type 304L, 316L, or 347. Straight sections can be bent as shown in Figure 11a and held by 5/8-in anchors either cast in the concrete or set in two-component adhesive in drilled holes. Alternatively, the construction shown in Figure 11b can be used.
Along ramp curves, the plates must be cut to the radii given in Equation 1 (or 2 for noses), and then they must be flexed and stitch-welded at the apex. The type of construction in Figure 11b can be used, or the detail in Figure 11a can be followed if the plates are cut off a little below the floor and tabs are fastened to them for the bolts.
The optimum slope for the sides of flow splitters is 45-deg, but if the trench is narrow and the width of the floor too confining, steeper slopes (up to 2:1 = 63.4-deg) are acceptable. In the 18-Mgal/d pumping station example, a flow splitter with sides sloping 45-deg leaves floors only 6-in wide, whereas a 2:1 slope provides floors 11-in wide. It is designer's choice.
For the detail in Figure 11b, specify that on ramp curves the contractor shall (before cutting expensive stainless steel) make wood patterns of the flow splitter (and its nose) of thin (1/8- to 1/4-in) plywood, transport them to the field and scribe accurate offset lines on them so that the steel sheets follow all the imperfections of the concrete surface.
The cost of flow splitters varies greatly with contractor and circumstance, but a budget figure today is about $550/linear ft for the flow splitter in Figure 11b-mostly for labor. It may cost more on the ramp. The flow splitter in Figure 11a is probably less expensive.
Conclusion
A trench-type wet well built to the recommendations of the American National Standard for Pump Intake Design [1] augmented by the advice herein is a superb performer and needs no improvement. However, model study is required if a pump exceeds a capacity of 40,000-gal/min, if the station capacity exceeds 100,000-gal/min, if pump operation and reliability is exceedingly critical, if approach flow is non-uniform or asymmetric or if the geometry of Figure 1 is altered.
Acknowledgements
This paper was sent to all commercial laboratories that test hydraulic models of pumping station wet wells. The correspondents were Thomas Demlow of Northwest Hydraulic Consultants, Inc., Andrew E. Johansson of Alden Research Laboratory, Richard E. Long of ENSR, and Tatsuaki Nakato, recently retired from the University of Iowa. There were suggestions for additions and word changes but no disagreements. Other reviewers were Garr M. Jones and Lawrence Oeth of Brown and Caldwell, Arnold Sdano of Fairbanks Morse Pump Corp., Sateesh Nabar of Nabor Stanley Brown, Inc., Sarwan Wason and Mike Zappone of Carollo Engineers, Constantino M. Senon of MWH Americas, Inc., and William Wheeler of Wheeler Designs. Arnold Sdano converted the penciled figures into electronic form. The authors express their gratitude to all those who contributed.
List of References
1. ANSI/HI 9.8-2008. American National Standard for Pump Intake Design, Hydraulic Institute, Parsippany, NJ (scheduled for publication in 2008).
2. Sanks, R.L., G. Tchobanoglous, B.E. Bosserman, and G.M. Jones. Pumping Station Design, 2cd Ed., Butterworth-Heineman, Boston (1998).
3. Jones, G.M., R.L. Sanks, G. Tchobanoglous, and B.E. Bosserman. Pumping Station Design, 3rd Ed., Elsevier, Boston (2006).
4. Trench2.0. A user-friendly computer program developed by Prof. Joel Cahoon, to calculate water depths, velocity, Froude numbers and sequent depths along the trench in trench-type wet wells. Available free at www.coe.montana.edu/ce/joelc/wetwell.
5. UnifCrit2.2. A user-friendly computer program developed by Prof. Joel Cahoon to calculate flow rate, velocities, water depths and critical water depth in circular and trapezoidal open conduits. Available free at www.coe.montana.edu/ce/joelc/wetwell.
6. Obtainable from McMaster-Carr Supply Co., Santa Fe Springs, CA. (www.mcmaster.com) as two-part casting compound No. 8644K11 for 1-lb (25-in3) or as No 8644K12 for 10-lb (250-in3).
7. Obtainable from many sources including McMaster-Carr Supply Co. as Item No. 86155K33.
Professor Emeritus Robert L. Sanks, a consulting engineer since 1944, has taught graduate or undergraduate courses in civil engineering for 15 years and in environmental engineering for 16 years, has authored or co-authored six books (he was co-editor or editor-in-chief of four), and has published numerous papers and participated in many seminars. He has studied models of pumping station wet wells since 1992. He was instrumental in improving the performance of trench-type wet wells, publicizing that design, serving on the draft committee for ANSI/HI 9.8 and promoting the popularity of the trench-type wet well. He has been a consultant on many pumping stations-several larger than 80-Mgal/d in size.
Theodore T. Williams, MS, F.ASCE, is professor emeritus at Montana State University, where he taught 47 years in civil engineering. He served as department head and associate dean of engineering. Prior to his academic career, he spent 5 years as engineer to an irrigation canal system in Colorado. He has completed a number of consulting engineering assignments in the areas of open channel and pipeline hydraulics. He has studied hydraulic models of pumping station wet wells since 2001.
<< Start < Prev 1 2 3 Next > End >>
|
Subscribe to this comment's feed
Joe Evans, PhD
Dr. Lev Nelik
Robert Perez 
