When designinga pumping station there are many factors that can affect the product such asclient needs, cost, civil constraints and design best practice with the latterbeing the starting point for an initial concept. Seldom does the finalconstructed pumping station reflect the initial concept but it is essentialthat commonly accepted best practice is only deviated from where there is agenuine need. One of the most commonly used design guides in the UK is the,’Hydraulic Design of Pump Sumps and Intakes’ by M. J. Prosser. It is commonly accepted that “the flowapproaching the intake section should be uniform, steady and of a single phase”(Prosser, 1977). What this means is the velocity and direction of flow shouldbe consistent and there should be no entrained air due to the flow conditions.
Although M. J. Prosser accepts that these ideal conditions cannot always beachieved, the recommendations for achieving this are often taken to extremes atthe expense of feasibility and therefore requires further validation or thedevelopment of alternative solutions.M.J. Prosser, discusses several differentflow conditions that could produce undesirable effects on the pump and the intake.Air Gulping – During low water levelconditions due to un-expected draw down of the pump sump, air would be allowedto enter the pipe, gathering in large pockets inside the pipe. Figure 4.
6ashown below, gives an example where the water level of the sump has beenallowed to drop significantly below bottom water level allowing substantialpockets of air to enter the pipe while still maintaining a full pipe flow. Thiswould lead to a significant drop in pump performance and would result higherthan normal levels of vibration caused by a non-uniform pressure acting on thepump impeller. This could also result in cavitation within the pump casing dueto the sudden drop in pressure caused by large pockets of air in the pipework,a phenomenon that will be explained later within this report.Figure 2.
4 – Image showing the airgulpingAir Entrainment – When a fluid is from aheight above the top water level, tiny air bubbles can be introduced into thewell. If this is a particularly small well or the suction flow rate issufficiently high, the entrained air can be dragged into the pump suction. Thiscould result in loss of pump prime, increased chance of cavitation and damagingvibrations in the pump assembly.
Figure 2.5 shows a typical cause of airentrainment where the fluid is dropped from above top water level (TWL) intothe sump. Entrained air has been known to cause a drop in efficiency of up to15% when only 3% free air has been pulled into the suction intake (HydrotecConsultants Ltd, 2015).Figure 2.5 – Image showing airentrainmentSurface vortices – Surface vortices cancause air to be pulled into the suction of a pump resulting in reduced pumpperformance can result in cavitation at the eye of the impeller. This type ofvortices is encouraged by regions of swirling and stagnant flow in the wet well(forced vortices) and a suction intake below a minimum recommended submergencedepth (free vortices).
M.J. Prosser gives a direct recommendation for theminimum submergence depth of a pump suction intake, which is a minimum of 5times the diameter below bottom water level (BWL) within the sump (GRUNDFOS Management A/S, 2004).Vortices may be surface or submerged formations.They can be classified in six different types; type one being a minor surfacedimple and type six being a fully developed air core as shown in figure 2.6 (HydrotecConsultants Ltd, 2015), (Hydraulic Institute, 1998). The acceptance criteriafor surface vortices can be summarized by the following statement, “Surfacevortices of type 3 or more and subsurface vortices of type 2 or more areunacceptable” (Jones, 2014).Figure 2.
6 – Image showing surfacevortices classifications Submergedvortices are said to be more detrimental to pump operation as the pressurewithin the vortex can sometimes drop below atmospheric pressure. This canresult in cavitation damage to the suction intake geometry and result gassesentering the pump. Submerged vortices can be classified in three main typesranging from a type one, which is simply swirl in the fluid flow leading up tothe intake to a type three fully developed air core. This is illustrated infigure 2.7 (Hydrotec Consultants Ltd, 2015), (Hydraulic Institute, 1998).Figure 2.7 – Submerged vorticesPre-swirl – Pre-swirlat pump intakes is not something that is pump manufacturers design their pumpsto be able to cope only to a certain degree.
When designing a pump sump, thepump manufactures assume that the flow conditions within the sump areacceptable therefore the performance characteristics they quote depend on this.If the velocity of the flow approaching the pump intake is Asymmetrical;meaning the velocity profile across the sump is not even, pre-swirl can occur this is illustrated in figure2.8 (Prosser, 1977). It is commonly stated that swirl angles should be lessthan 5 Degrees or 6 Degrees if only occurring during infrequent pump operation (Joneset al.
, 2014).Figure 2.8 – Diagram showingpre-swirl formationSwirl can form in boththe clockwise and anti-clockwise directions and will be most detrimental topump performance when the swirl acts in the opposite direction to the rotationof the impeller. If the rotation of swirl does oppose the impeller rotation,then this will act as a torque that artificially increases the head that thepump operates against. When pumps operate at a higher head, more energy must betransferred to the fluid by the impeller. The additional energy used will meanthe pump will have a reduced efficiency as it will be consuming more electricalpower. The effect of swirling within the pipework can increase eddy formationand turbulence which can triple the head loss due to pipe fittings. Pre-swirlcan also reduce the Net Positive Suction Head Available (NPSHA) andcause cavitation within pump volute casing (Jones et al.
, 2014). NPSHA,is the amount of suction head, measured in metres, above the centreline of thepump impeller. Pre-swirl and excessive velocities can reduce the NPSHA toa value below that of the vapour pressure leading to the formation of bubblesof gas that will ultimately explode and produce a shockwave capable of damagingthe impeller, pump casing and cause excessive vibration.Another key piece ofliterature that discusses desirable flow conditions is the American National Standard for PumpIntake Design 1998. This standard is widely accepted and the informationcontained within it is referenced quite commonly within the water and wastewater industry. The standard briefly discusses flow conditions but tends tofocus more on the geometry of suction intakes; something that M.J.
Prossersummarizes quite also. The standard briefly outlines several key ‘designobjectives’ for mump intake design. It states that “specific hydraulicphenomena have been identified that can adversely affect the performance ofpumps” (Hydraulic Institute, 1998).
The phenomena it isreferring to is listed below for further discussion.· Submergedvortices· Free-surfacevortices· Excessivepre-swirl of flow entering the pump· Non-uniformspatial distribution of velocity at the impeller eye· Excessivevariations in velocity and swirl time· Entrainedair.As the above quotestates that these phenomena must not be present to an excessive degree only itdoes afford the designer some flexibility as it is unlikely that these can beeliminated for every design.Of the above flowconditions, the first three will be of interest within this research. Thevortices can be predicted by the CFD software and their presence can bevisually confirmed. Pre-swirl will also be measured and the criteria for thiswill be discussed later in the lit review. The standard goes on to explain thatthe impact of these phenomena on pump performance varies on a case by casebasis. It also highlights the fact that such a detailed impact analysis has notbeen carried out.
Quantifying the amount of air entrainment within the pumpsuction was an initial objective of this research, however due to complicatednature of Two Phase CFD models this was ruled out of the scope. This will bediscussed later in the report.2.2.
2Pumping Station Geometry Thedesign guide by Prosser and the American Standard also outline dimensional requirementsfor pumping stations and this will be discussed within this section of thereport. There are several main styles of pumping station commonly used,rectangular, square and circular. For each type, the designs can vary incomplexity and size to quite a large degree and this would be impossible to fitwithin the scope of this project. For this reason, only rectangular, single suctionpump sumps will be discussed. Further justification for this scope limitationwill be given within the methodology section of this report.The design guide byM.J Prosser is the favoured guide for information on this kind of pump sump.
The guide goes in to quite a bit of detail on the dimensions that designersshould attempt to achieve to ensure the pump intake operates most efficiently.There are several key dimensions that Prosser outlines within the design guide.It is explained that most pump manufacturers give these dimensions in terms ofbell mouth diameter, i.e. dimension x should be 5 x D, where D is the bellmouth diameter and d is the pipe internal diameter. These dimensions will beused to construct the 3D models for my research.
The type of suction fittingthat should be used is also discussed. Prosser states that a Bell Mouth fittingis commonly specified as a requirement by most pump manufactures and defines aratio of D/d of 1.5 to 1.
8 with a dimensional tolerance of +/- 20%. The guidealso briefly discusses the impact of using a ‘sharp edged intake’ whichconsists of a simple pipe with no graduated inlet. Using such a suction isexpected to cause flow separation resulting in swirl within the suctionpipework, whereas the bell mouth allows the flow to remain attached to thesurface of the pipe. Figure 2.
9 below illustrates this principle (Prosser,1977). Another reason to use a bell mouth instead of a sharp edge intake isthat the bell mouth has a head loss coefficient of around 0.04 compared to 0.5for a sharp edge (Vano Engineering, 2012)