Computation Fluid Dynamic Modeling of a Proposed Influent … · 2019-07-18 · Computation Fluid Dynamic Modeling of a Proposed Influent Pump Station Edward Wicklein, Charles Sweeney,

Computation Fluid Dynamic Modeling of a Proposed Influent Pump Station

Edward Wicklein, Charles Sweeney, Constantino Senon1, Doug Hattersley1, Brian Schultz1, and Randy Naef 2

ENSR 9521 Willows Road NE Redmond, WA 98053

1) MWH Americas, 2) Clean Water Services

ABSTRACT

Traditionally pump intake hydraulics have been investigated using scale physical hydraulic models. Computational fluid dynamics (CFD) is a rapidly improving technology that in some cases can be used as a cost or time effective alternative to traditional physical model methods for evaluation of pump intake hydraulics. This paper presents a case study on the use of CFD modeling to re-evaluate and modify the design of a large wastewater treatment plant influent pump station. The pump station has two a self cleaning trench wet wells, which was physically modeled early in the design process. Changes were made to the pump station layout and capacity during subsequent facility design. A CFD model was used to refine the design of the wet wells and evaluate the effects of upstream sewer alignment on the pump hydraulics. This paper presents the results of the CFD model analysis and design refinement.

KEYWORDS

CFD modeling, Physical Modeling, Pump Intake Hydraulics, Wastewater

INTRODUCTION

A new wastewater treatment plant influent pump station is being designed and constructed as part of an overall wastewater treatment plant facility upgrade. The pump station has two parallel self cleaning trench wet wells with an initial pumping capacity of 180 mgd, and an ultimate capacity of 240 mgd. The parallel trench wet wells will be mirror images of each other, with the inside wall common to both wet wells and the pump and control rooms located on the outside of the wet wells.

Pump stations of this size and type are typically evaluated with a scale physical model to investigate and verify the hydraulic conditions approaching the pumps. Early in the design process a scale physical hydraulic model study was conducted to improve the hydraulic performance of the proposed pump station (ENSR 1999). Through extensive model testing, a series of geometric modifications were developed to improve the hydraulic conditions.

Subsequent analysis to reconcile station inflow rates led to several changes in the proposed pump station design. These changes included a reduction in flow rate leading to fewer required pumps, and a revised influent sewer design that simplifies construction. The design of the revised pump station was different enough from the previously modeled design that additional analysis was desired. An additional goal was to evaluate the effects of proposed upstream sewer changes on

7094

WEFTEC®.06

Copyright 2006 Water Environment Foundation. All Rights Reserved©

the pump hydraulics; therefore an innovative modeling approach was taken to aid in refining and verifying the hydraulic design of the proposed pump station utilizing a CFD model.

POTENTIAL PUMP STATION HYDRAULIC PROBLEMS

Hydraulic conditions have been identified such as air entrainment, vortex action, pre-swirl, and excessive turbulence in the approach flow to pumps can lead to fluctuating loads on pump impellers, vibration, cavitation, loss of pump capacity, and decreased efficiency (Sweeney and Rockwell 1982). It has been shown that these problems are strongly influenced by the approach flow hydraulics upstream from the pump, caused by the wet well geometry coupled with the influent conditions. Straight and uniform approach flow reduces the tendency for pump problems, whereas variable approach flow direction and non-uniform velocity distribution generate eddies and circulation patterns, which may adversely affect pump operation.

Providing uniform approach flow conditions may reduce the potential for pre-swirl and vortex formation. Tullis (1979) and Sweeney et al. (1982) have documented repeated cases in which preclusion of submerged vortices has required the installation of anti-vortex devices such as flow splitters, guide vanes, and/or cones.

The geometry of the wet well, operation of the pump(s), and depth of water in the sump influence the approach flow hydrodynamics and can result in the following adverse hydraulic phenomena (Sweeney and Rockwell 1982):

• pre-swirl of flow approaching the pump impeller;

• free surface vortex formation;

• submerged vortex formation;

• spatial asymmetry of the flow approaching the pump impeller; and

• temporal fluctuations (turbulence) in the flow approaching the pump impeller.

Pump impellers are designed with the assumption that flow approaches the impeller axially. Pre-swirl of the flow in a pump inlet causes the flow to approach the impeller at an angle, which can result in a change in pump performance (head and flow). Pre-swirl may also reduce the minimum pressure on the impeller blade if the direction of pre-swirl is opposite the direction of rotation of the impeller. Excessive low pressure on the suction side of the pump impeller blades may ultimately cause cavitation damage. In addition, if the pre-swirl is not constant, it will result in load fluctuations.

Free surface vortices and submerged vortices can also influence pump operation. Strong free surface vortices may cause air to be entrained into the pump, potentially resulting in loss of prime and loss of pump capacity. Submerged and free surface vortices entering the pump, even without air entertainment, will impose a fluctuating load on the pump impeller blades as each blade passes through the lower pressure vortex core. Stable vortices produce load fluctuations at blade pass frequency (or multiples thereof) capable of causing vibration, accelerated bearing wear and, in extreme cases, impeller and diffuser component fatigue. If the natural frequency of

7095

WEFTEC®.06


the pump vibration approaches the blade pass frequency, destructive resonance results. The low-pressure vortex cores may reduce the local pressure at the impeller below the fluid pressure and induce cavitation of the impeller blades.

Spatial asymmetry in the distribution of velocities around the pump may cause an unbalanced loading on the impeller and vibration, while temporal fluctuations (turbulence) in the velocities at a particular point may result in broad-spectrum noise and vibration. Deviations in the spatial and temporal velocity distributions also can produce cavitation.

PUMP STATION HYDRAULIC ANALYSIS METHODS

Traditionally, scale model studies have been conducted to optimize the design of large pump stations; however, an emerging technology for pump station analysis through CFD modeling. Applications of CFD models to simulate flow fields associated with pump intakes have been underway for several years. Reports and papers have been published on the use of CFD modeling for analysis of pump station hydraulics, including: Constantinescu and Patel (1998); Nagahara et al. (2001); Li et al. (2001); and Ansar et al. (2002). Much of this research has focused on the simulation of vortex formation in pump sumps and circulation for pump stations with a single operating pump. Wicklein et al. (2002) have shown that a CFD model can accurately reproduce the flow field associated with cooling water pump intakes with multiple bays for a range of pump operations and water levels. Wicklein and Rashid (2006) have demonstrated that CFD models are very valuable tools for investigating pump station hydraulics and developing modifications to address performance deficiencies.

CFD Modeling

Commercially available CFD models numerically solve the fundamental equations of fluid flow, conservation of mass and momentum, known as the Reynolds-averaged Navier-Stokes (RANS) equations. These equations do not form a closed set (ASCE Task Committee 1988), owing to the non-linearity of the original Navier-Stokes equations and their temporal averaging. Current CFD models solve additional equations representing the turbulence characteristics of a flow field, which is a key parameter in determining the nature of flow, eddy formation, circulation, flow separation, and flow interaction with structures. The turbulence models commonly used in hydraulic engineering have been reviewed by the ASCE Task Committee (1988) and Rodi (1980). Commercial CFD models offer various turbulence closure models, the most common of which are based on second order closures using k-ε and k-ω formulations.

In their general form, RANS equations cannot be solved analytically. Commercial CFD models approximate the differential equations by the finite difference method which resolves the equations into a set of algebraic equations (Lomax et al. 2003). These algebraic equations are solved to provide hydraulic information (e.g., velocity, water surface elevation, and pressure) at a finite number of discrete points within the flow domain. Most finite difference-based CFD models use the finite volume method, as this approach allows the use of unstructured computational grids.

As the RANS equations are typically solved by the finite difference method, it is necessary to discretize the flow domain into a computational grid to define the actual locations where

7096

WEFTEC®.06


equations of flow will be solved. Traditionally the individual computational cells are hexahedral (6 faces), pyramidal (5 faces), prismatic (5 faces) or tetrahedral (4 faces), defined by the corner vertices. The task of grid generation is accomplished through the use of grid generating software that allows for definition of the model geometry, computational cell size, and grid density, and provides tools for grid quality analysis. Unstructured computational grids are the most common type, as they allow the greatest flexibility in defining the model domain and meshing properties.

The flow field computed by the CFD model is a direct function of the flow conditions applied at the domain boundaries, known as boundary conditions. Typical boundaries include inflow boundaries, outlet boundaries, pressure boundaries, symmetry boundaries, and wall boundaries. Inlet boundaries provide a constant velocity in the three vector components into or out of the model domain, as well as constant turbulence characteristics. Pressure boundaries have constant pressure and turbulence characteristics, and flow can move in or out of the domain. Outlet boundaries only allow flow to travel out of the domain, and have no pressure or turbulence constraints. Symmetry boundaries allow no vector component normal to the boundary. Wall boundaries are considered solid with no flow through the boundary. The wall boundary type can either be no slip, with a roughness component, or a slip wall with no roughness component. Typically the law of the wall function is used to approximate the transition from the zero velocity at the boundary through the boundary layer into the free stream, which models the effective drag from the roughness at the wall, without requiring the large number of computational elements required to resolve the flow field within the boundary layer. Typically the boundary layer is not resolved when investigating large scale flow features due to significant computational overhead requirements in resolving this flow feature.

Model Performance Criteria

The Hydraulic Institute (HI) established criteria for evaluating performance of pump station designs through the use of physical hydraulic model studies. The details of physical modeling procedures and results interpretation are explained in ANSI/HI 9.8-1998. The summarized minimum performance criteria for physical models are:

• No organized free surface and/or subsurface vortices of greater magnitude than a Type 2 shall enter the pump for Froude-scaled model operation (referring to HI 1998 Figure 9.8.23). Dye cores must not be coherent for more than 10 percent of the time.

• The level of pre-swirl should be less than 5 degrees from axial and should be steady.

• Time-averaged velocities measured at eight locations in the pump throat should be within ± 10 percent of the spatial mean of time-averaged velocities.

• The temporal fluctuations of velocities measured at each of the eight locations should be less than 10 percent of the average measured at that location.

To date, a similar detailed set of performance criteria for evaluation of pump station performance using numerical model results has not been established by HI.

The key difference between current CFD model results and the results from physical model studies is that physical models are run in a quasi-steady state, whereas CFD models are run in an

7097

WEFTEC®.06


absolute steady state. A physical model has a fixed inflow, outflow, and average water level, but the velocity and water level at a given point fluctuate due to turbulence and local flow instabilities. Currently, CFD models provide the averaged solution of velocity at all points in the domain, and have a non-fluctuating water surface. The CFD model results therefore cannot be exactly compared with the current physical model criteria, as the fluctuating components of the flow field are averaged out.

It was previously established that the physical model study conducted by ENSR developed a set of geometric changes that brought the pump station performance within established HI criteria for physical model testing. Therefore, analysis of the same station with a CFD model should produce model results that would similarly meet HI criteria.

For comparison and presentation of pre-rotation and velocity results, point data were extracted from the CFD results in the pump suction piping to replicate the data taken from physical model. Eight points were taken on 45-degree increments on a radial traverse at the impeller elevation, and eight points were taken on similar 45-degree radial traverse at the point where a rotometer would be located in the physical model. The data taken at rotometer location were used to calculate a rotational velocity within the pump suction piping.

The angle of flow rotation approaching the impeller, θ, is reported in degrees from axial, and typically referred to as pre-swirl. The angle is calculated by equation 1:

)/U(UTanθ at1−= (1)

where:

Ut = tangential component of velocity

Ua = axial component of velocity

In this case the approach angle was calculated for each of the eight points, and averaged to find the average flow angle. For this study a positive angle indicates flow rotating clockwise looking downstream at the rotation location.

The maximum and minimum velocity is found by dividing the velocity found at each of the eight points by the cross sectional average velocity. The velocity is then expressed as a non-dimensional ratio, which facilitates comparison between different flow rates and scales.

BASELINE CFD MODEL

A CFD model of the final design of the physically modeled pump station was developed. The physical model included two parallel straight influent conduits 40-prototype feet long and 6.75 feet wide by 7 feet high. The conduits ended at an ogee crest at the start of the wet well. The wet well contained 4 pumps. The capacity of the two upstream pumps was 45 million gallons per day (mgd) each, and capacity of the two downstream pumps was 30 mgd each. The pump suction piping included a short radius flared turndown elbow, a horizontal spool, an eccentric reducer, and a long radius reducing elbow which ended at the pump flange. The wet well width

7098

WEFTEC®.06


was narrowed through several transitions from the upstream end to the downstream end. A fillet ran the length of the wet well between the walls and floor. A center triangular cross section flow splitter also ran the length of the wet well. A horizontal baffle blocked the top portion of flow on the centerline of the inflow jet to better distribute flow in the wet well.

The CFD model geometry was developed from the CAD drawings of the final modified wet well design reported by ENSR (1999). The drawings were in two dimensions and at the model scale of 1:4.8. They were scaled to project dimensions, and the three-dimensional geometry was developed from the files.

Figure 1 shows and isometric view of the model grid of the physically modeled wet well configuration. The model grid is made up of over 567,000 hexahedral (6 faces) and prismatic (5 faces) cells. The grid is sufficiently refined to include a 1.2-inch thick wall on all pump piping within the wet well.

Figure 1 - Baseline Model Grid

Baseline Test Program

The CFD model was run for 5 different flow rates/pump operating conditions that were run in the physical model. Table 1 summarizes the tests run, and the corresponding physical model documentation test numbers.Table 1 – Baseline Pump Operating Conditions

Pump Flow (mgd) Test No. Water Surface

Elevation (ft) Pump 1 Pump 2 Pump 3 Pump 4 Station Flow

(mgd)

1 92.2 45 45 2 93.3 45 30 75 3 93.6 45 45 90 4 94.5 45 45 30 120 5 95.6 45 45 30 30 150

The tests chosen represented the range of operating flows, pump operations and water levels tested in the physical model. In addition, tests were chosen that included operation of pump 2,

7099

WEFTEC®.06


which had the most detailed instrumentation in the physical model. In all cases the CFD model was run to a converged steady state solution, defined for this project as a global residual of key computed values of less than or equal to 0.001.

Baseline Run Model Results

The results of the CFD model runs were compared with the physical model results using the typical parameters that are used to present physical model results: pre-swirl, and the maximum and minimum velocity at the impeller location.

The tabular results of point data locations from the CFD model are compared with the physical model results for all 5 tests in Table 2. Results are reported from all pumps operating in the CFD model. Velocity information is only reported from pump 2 in the physical model as this was the only pump with an instrument to record velocity data. The velocity information was included for all pumps in the CFD model, as it was readily available.

Table 2 – Baseline Model Result Comparisons Between CFD Model and Physical Model

Degrees from Axial Vmin/Vavg Vmax/Vavg Test No. Pump CFD Physical CFD Physical CFD Physical 1 2 -5.3 0.7 0.97 0.98 1.03 1.05 2 2 -2.1 0.6 0.97 0.99 1.02 1.04 3 0.4 0.7 0.95 1.04 3 1 0.4 1.1 0.97 1.03 2 -1.1 -1.2 0.96 0.97 1.03 1.03 4 1 0.6 2.8 0.97 1.03 2 -4.6 1.7 0.97 0.98 1.03 1.03 3 4.6 -0.2 0.95 1.04 5 1 0.9 1.6 0.95 1.03 2 -6.9 0.5 0.97 0.97 1.03 1.04 3 4.5 -3.4 0.95 1.04 4 4.5 0.8 0.97 1.03

The pre-swirl of flow approaching the impeller is typically higher in the CFD results for most of the cases. This is not surprising, as the means of calculating the angle is slightly different for the two models. In the physical model the tangential velocity is measured by a rotometer for a fixed amount of time, which is assumed to be representative of the flow field. Because of flow instability, the tangential velocity is not steady and can reverse direction during a course of measurement. In this case the reversal reduces the average, leading to a lower reported value. This type of flow instability is very common in trench wet wells with a strong inflow jet. Review of the complete data from the physical model testing reveals that for very similar operations with minor or no geometric changes, the angle reported could be either positive or negative. The CFD results represent a true average condition. The established criterion for the

7100

WEFTEC®.06


approach angle is 5 degrees from axial. The CFD model results show that this criterion is met in almost all of the cases, as it had been in the physical model study.

Both the maximum and minimum velocity compare very well between the CFD model and the physical model. In both cases the velocity is nearly uniform at the impeller location. This uniformity is expected, as the piping has both an eccentric reducer and a long radius reducing elbow with a 61 percent area reduction. Flow acceleration attenuates velocity profile variations.

Figure 2 shows three views of the velocity magnitude in the wet well for test condition 1. The top image shows the velocity magnitude at the water surface, which is generally symmetric along the centerline. The middle image shows a section view through pump 2, the only operating pump in this condition. The velocity is generally low in the top portion of the wet well, and higher near the pump inlet. There is considerable flow separation within the flared suction elbow in, as evident by the low velocity region on the downstream side of the elbow. The bottom image shows a vertical section view along the wet well centerline. The flow dives below the baffle location, to the operating pump, and the velocity is generally low in the rest of the wet well. Figure 3 shows streamtraces through the wet well. Flow enters the wet well, and dives below the baffle, setting up two vertical circulations. Flow moves up the ogee, where it is re-entrained in the inflow jet. A second circulation cell develops in the downstream end of the wet well, with flow moving toward the baffle near the surface, and then being pulled down at the baffle location.

In general, we can conclude that the CFD model results compare reasonably well with the physical model results, and the CFD model is therefore representative of the system. As the physical model and the CFD model of the same configuration met established performance criteria, we assumed subsequent CFD model runs should similarly be held to the same performance standards.

7101

WEFTEC®.06


Figure 2 – Three Views of the Initial Flow Field, Test 1

7102

WEFTEC®.06


Figure 3 – Streamtraces of Flow Path For Initial Flow Field, Test 1

REVISED WET WELL CFD MODEL

The pump station design was modified to account for a new flow capacity, geometric changes to the influent sewer, and a new pump phase-in program. The phase 1 maximum pump station capacity will be 180 mgd, or 90 mgd per wet well. The wet wells were moved adjacent to each other with so that separate pump mechanical rooms could be located on the outside of the station. The first two pumps in each side of the station will have a capacity of 25 mgd, and the last pump will have a capacity of 40 mgd. The piping through the wall and in the wet well will be 42 inches in diameter for all pumps so that a future capacity expansion can be carried out without the need to replace piping within the wet well. All pipe fittings used were standard ductile iron. The wet well was shortened by 9 feet and the three pumps located on 15-foot centers. The invert of the ogee crest was raised to an elevation of 92.15 feet, and the entire wet well elevation was adjusted to account for the change in ogee elevation. The box culvert at the influent was replaced with a 72-inch diameter pipe that enters a 36 foot long transition section, which ends at the ogee.

The model was revised to reflect the new pump station geometry, which included raising and shortening the wet well, replacing the pump piping, and modifying the upstream sewer system. The model grid depicted in Figure 4 shows an isometric internal view of the wet well and suction piping.

7103

WEFTEC®.06


Figure 4 - Revised Model Grid

Revised Model Test and Results

The new wet well will operate at different flow rates, water levels, and pump operations than the previous physical model. One test was carried out with the new design. For this test pumps 1 and 2 operated at 25 mgd each, and the water level was set to maintain critical flow depth at the ogee crest. At wet well water levels below the critical depth at the ogee for the given flow, the pumps will pump down the wet well and the condition therefore couldn’t be maintained. The test results are summarized in Table 3

Table 3 – Revised CFD Model Initial Results

Pump Discharge

(mgd) Water Level (ft) Degrees from

Axial Vmin/Vavg Vmax/Vavg

1 25 94.51 5.87 0.97 1.03 2 25 -6.54 0.95 1.04

With the revised pumps, the velocity distribution at the pump impeller location was not changed significantly from baseline results. The pre-swirl increased significantly in the two operating pumps, as compared with the initial tests.

Figure 5 shows contours of velocity magnitude for the run, similar to the previous Figure 2. The flow field in the wet well was driven by the strong inflow jet from the 72-inch diameter pipe.

7104

WEFTEC®.06


Replacing the inflow box culvert with a pipe significantly reduced the flow area, and lead to a corresponding increase in flow velocity entering the wet well. The expansion transition from the circular pipe to the wet well did not significantly dissipate the inflow jet energy. The jet attached to the wall adjacent to the pump mechanical room. The top of the horizontal baffle extended above the water surface, which forced the jet to dive toward the first two operating pumps. Significant flow separation was evident at the inside bend of the flared suction elbow in pump 2. In addition, it appeared that much of the flow entering the elbow was coming from the right side of the center floor flow splitter which contributes to pre-swirl.

Figure 6 shows streamtraces colored by velocity magnitude within the wet well. A recirculation zone developed in a portion of the expansion, adjacent to the jet after it attached to the wall. Flow dove beneath the baffle and reached the floor near pump 2. From there flow traveled along the floor and rose at the downstream end of the wet well. The flow then traveled upstream at the wet well surface to the baffle wall, where it was pulled down to pump 1. There was swirling of flow between pumps 2 and 3.

DESIGN DEVELOPMENT TESTING

Extensive changes were made to the wet well design through a series of tests. These changes addressed the excessive pre-swirl of flow, reduced the flow separation in the suction piping and reduced the tendency for surface vortices to develop. Figure 7 shows typical coherent flow swirl from the surface that reached the pumps.

Modifications included replacing the suction elbows with long radius elbows and flares to reduce the observed flow separation. This change necessitated lowering the wet well floor elevation, which reduced the tendency for the surface swirl develop and reach the pump inlet locations. The wet well was shortened and the water levels were raised above critical depth at the ogee. Finally, non-clog anti-rotation vanes were added in the suction piping to reduce pre-swirl of flow approaching the pumps. The vanes significantly reduced pre-rotation of flow approaching the pumps. The wet wells still included the fillets and center flow splitter developed in the previous physical hydraulic model study that were necessary to minimize development of submerged vortices, and aid in the self cleaning cycle. The final model grid is shown in Figure 8.

7105

WEFTEC®.06


Figure 5 – Three Views of the Revised Model Flow Field

7106

WEFTEC®.06


Figure – 6: Baseline Results

Figure 7 – Typical Surface Swirl Reaching the Pump Suction Inlets

7107

WEFTEC®.06


Figure 8 - Final Model Grid

FINAL MODEL RUNS

A series of documentation tests were run with the final model to verify the pump station performance for a range of pump operations. The test program used for documentation testing is presented in Table 4. The wet well water level was set based on the normal depth of flow where the 72-inch diameter influent entered the wet well transition section. Only one wet well was modeled in detail for these operations.

Table 4 – Pump Station Documentation Tests Operations

Pump Flow (mgd) Test No. Pump 1 Pump 2 Pump 3

Total Station Flow (mgd)

Minimum Operating Water El

1 0 25 0 25 94.30 2 25 25 0 100 95.30 3 25 25 40 155 96.90

Three tests were run documenting the flow conditions in the wet well for phase 1. The results of the three tests are summarized in Table 5. For these tests the pre-swirl was very low, with a peak computed value of 2.6 degrees which was less than the peak value calculated for the physical model configuration of 6.9 degrees. The velocity distribution was almost uniform, with only slight variations.

7108

WEFTEC®.06


Table 5 – Documentation Tests Results

Test No. Pump Discharge

(mgd) Water

Level (ft) Degrees from

Axial Vmin/Vavg Vmax/Vavg 1 2 25 94.30 1.3 0.99 1.01 2 1 25 95.30 2.6 0.98 1.01 2 25 2.3 0.99 1.01 3 1 25 96.9 1.0 0.98 1.01 2 25 0.4 0.98 1.01 3 40 0.7 0.96 1.04

Figures 9 through 11 show the results of the runs. The figures all show the wet well surface velocity magnitude, the velocity magnitude at a centerline section through the wet well, and a section through pump 2. The results of all three tests were well within the performance criteria established previously during analysis of the physical model.

SUMMARY AND CONCULSIONS

The wet well hydraulics of an influent pump station were investigated with a CFD model for the original recommended sump design developed through a physical model study. Performance criteria for evaluating the CFD model were established through comparison of the CFD results with the physical model study results. The CFD model was then used to evaluate the proposed system changes.

The flow details at the pump impeller location were very dependent on the suction piping. The tendency for surface vortex formation was dependent on the wet well geometry. Modification tests were conducted that led to a wet well design that decreased the potential for vortices to develop, improved the velocity distribution, reduced flow separation in suction elbows, and reduced pre-swirl of flow to acceptable levels for a wide range of operating conditions.

The model was also used to aid in design of an upstream sewer junction. The model was used again after the studies reported on in this paper closer to the time of field construction to evaluate the effects of changes in the sewer design geometry on the hydraulic conditions at the pump suction locations. This proved to be a great advantage over traditional physical model methods as the upstream sewer is seldom included in the physical model, and the physical model is seldom kept until field construction is completed. The CFD model is digital, and is available for use at any time after it is developed. Further, the modeling is conducted at prototype scale with fewer practical restrictions on the overall domain of the model.

7109

WEFTEC®.06


Figure 9 – Three Views of the Final Model Flow Field, Test 1

7110

WEFTEC®.06



7111

WEFTEC®.06



7112

WEFTEC®.06


REFERENCES

Ansar, M., T. Nakato, and G. Constantinescu. 2002. Numerical Simulations of Inviscid Three-Dimensional Flow at Single-and Duel-Pump Intakes. Journal of Hydraulic Research, 40(4), pp. 461-470.

ASCE Task Committee. 1998. Turbulence Modeling of Surface Water Flow and Transport: Part I, II, III, IV, V, Task Committee on Turbulence Models in Hydraulic Computations. Journal of Hydraulic Engineering, 114(9), pp. 970-1073.

Computational Dynamics Ltd., 2002a, Methodology, STAR-CD Version 3.15, Computational Dynamics Ltd. London, U.K.

Computational Dynamics Ltd., 2002b, User Guide, STAR-CD Version 3.15, Computational Dynamics Ltd. London, U.K.

Constantinescu, G. S., and V. C. Patel. 1998. Numerical Model for Simulation of Pump-Intake Flow and Vortices. Journal of Hydraulic Engineering, 124(2), pp. 123-134.

ENSR. 1999. Final Report, Durham Wastewater Treatment Plant 3 Expansion Influent Pump Station Hydraulic Model Study. Prepared for HDR Engineering, Portland, Oregon.

E. Wicklein, and M. Rashid. Use of Computational Fluid Dynamic Modeling to Evaluate Pump Intake Performance and Develop Design Modifications. Published in the Proceedings of the 2006 ASCE Environmental Water Resources Conference, Omaha, Nebraska, May 2006

Hydraulic Institute. 1998. American National Standard for Pump Intake Design. ANSI/HI 9.8-1998.

Li, S., L. Yong, J. Silva, and V. C. Patel. 2001. CFD Modeling of Three-Dimensional Flow in Practical Water-Pump Intakes. IIHR Technical Report No. 419.

Lomax, H., T. Pulliam, and D. Zingg. 2003. Fundamentals of Computational Fluid Dynamics, Springer, Berlin, Germany.

Nagahara, T., T. Sato, and T. Okamura. 2001. Effect of the Submerged Vortex Cavitation Occurred in Pump Suction Intake on Hydraulic Forces of Mixed Flow Pump Impeller. Presented at CAV 2001: Fourth International Symposium on Cavitation, California Institute of Technology, Pasadena, CA USA. June 20-23.

Rodi, W., 1980. Turbulence Models and their Application in Hydraulic Models. International Association for Hydraulic Research, Delft, The Netherlands.

Sweeney, C.E., D. Hay, and R.A. Elder. 1982. Pump Sump Design Experience: Summary. Journal of Hydraulic Division, A.S.C.E. Vol. 108, No. HY3: pp. 361-78.

7113

WEFTEC®.06


Sweeney, C.E., and G.E. Rockwell. 1982. Pump Sump Design Acceptance Through Hydraulic Model Testing. 11 pp. In: Proc. of the International Association for Hydraulic Research: Symposium on Operating Problems of Pump Stations and Power Plants. Amsterdam, The Netherlands, 13-17 September 1982.

Tullis, J.P. 1979. Modeling in Design of Pumping Pits. Journal of Hydraulics Division, A.S.C.E. Vol. 105, No. HY9: pp. 1053-63.

Wicklein, E., C. Allaben, and M. Rashid. 2002. Optimizing Cooling Tower Pump Intakes Using Computational Fluid Dynamics Models. Published in the Proceedings of the 2002 Industrial Water Conference, Orlando, Florida.

7114

WEFTEC®.06


Documents

Computation Fluid Dynamic Modeling of a Proposed Influent … · 2019-07-18 · Computation Fluid Dynamic Modeling of a Proposed Influent Pump Station Edward Wicklein, Charles Sweeney,