Centrifugal Pumps- Part VII- Casing Hydraulic Design.
1- Intro
In this article we tried to cover the basic concepts of casing hydraulic design in order to enable reads to have better understanding of centrifugal pump design. Hydraulic design of the casing refers to the parts where the speed converts to pressure.
In this article we will not cover design for other parts in the casing such as mechanical seals, bearing and etc.
We are not trying to apply a set of rules for all pump manufacturers as they use their methods based on their design and researches.
In this article whenever pump mentioned, centrifugal pumps are meant.
2- Centrifugal pump principal
Pumps are used to increase the pressure of liquid flow. A basic cross section of centrifugal pump showed in the figure 1.
Figure 1- Cross section of basic centrifugal pump
As the flow enters the pump, it is directed to the impeller eye. Impeller rotates and increase the speed of the liquid. Depending on the impeller profile the pressure in this section may increase. The liquid leaves the tip of impeller with high speed, around 30-50 m/s, depending on circular velocity of the impeller and the dimensions of the impeller. This speed is too high for pipeline to handle and yet the pressure is not high enough. In the volute section (1) as the flow proceed, the volute will be gradually wider. Because of Mass-continuity law, the speed will drop. According to Bernoulli’s law (figure 2), as the speed drops, pressure will increase, considering "gz" is constant on both sides.
Figure 2- Bernoulli's law
In the pipe lines the speed shall be as low as 3 m/s. Higher speed will cause vortices to form, which increase the friction losses drastically. In addition high speed will speed up the erosion process. If speed is lower than a certain value, there is a risk of solid particle formation or the corrosion of the line.
3- Numbers for Designers
There are two numbers that we will use them a lot in this article, namely Specific Speed (Ns) and Suction Specific Speed (S or Nss). In the pump industries, in order to make the equations simpler, gravitational constants are neglected.
3.1- Specific speed
This number shows the PUMP PERFORMANCE at best efficiency flow rate (BEP) with the maximum impeller size.
Figure 3- Specific speed Ns- Courtesy of API 610, 11th edition.
In the right hand side of equation, n is speed of pump (RPM), q is the flow m3/hr or gallons per minute and H is head per stage expressed in m or ft.
3.2- Suction Specific speed
This number shows PUMP’s SUCTION PERFORMANCE at best efficiency flow rate (BEP) with the maximum impeller size. Problems such as cavitation can be seen using this number.
Figure 4- Suction Specific speed S or Nss- Courtesy of API 610, 11th edition.
In the right hand side of equation, n is speed of pump (RPM), q* is the flow m3/hr or gallons per minute and NPSHR is net positive suction head expressed in m or ft.
* Contrary to specific speed, in Suction specific speed, if the impeller has 2 suction eyes, q shall be halved. Double eyed impellers are used when the flow is high.
Based on our experience the NSS above 12000 US units will cause the vibration.
In both numbers one can use metric or US units. 1 Metric unit is equal to 51.64 US units.
In general lower values of these two numbers are desired. The values can be lowered by using lower speed pump or auxiliaries such as inducer.
4- Casing design
Back to main article, as mentioned in pervious chapters the pressure increment take place in the volute section, does all the pumps needs a volute? The answer is no. There are other methods that the pressure can be increase. However, all of these methods have same principal. The passing area of the fluid increases, speed will drop according to mass-continuity law and the pressure will increase according to Bernoulli’s law. Based on the manufacturers’ designs, one or more methods can be applied at the same time.
4.1- Vaneless guide ring
Vaneless guide ring, consists of two smooth discs. The distance between discs can be constant or increasing. The speed of flow reduce based on ratio between outer diameter (Do) to inner diameter (Di). As the (Do) increases, the speed of fluid will drop more thus pressure will rise more. This method can be used only in low viscosity fluid and low head. If the higher the head is, the more speed fluid needs. In order to convert this high speed to pressure, outer ring diameter shall be too large, which is not practical. This method can be used with low Ns numbers.
Figure 5- Vaneless guide ring
4.2- Concentric casing
This method is used with single stage pumps or the last stage of multistage pumps. As the name implies, since the casing and impeller centers are same, the flow pressure will not increase in the casing but only in the volutes. The casing diameter to impeller diameter ration shall is between 1.15 to 1.2. For low flows, this method has a higher efficiency than the volute casing (next paragraph). This method is only used when the Ns is below 600 US units. Above this figure, experiments proves that the efficiency of the pump will drop rapidly.
One may ask why to use this method. The answer is fabrication limitation. If one of these conditions are true, this method is better than volute method.
- Where the pump casing has to accommodate several impeller sizes.
- Where pump has to use a fabricated casing.
- Where volute passage has to be machined from a casting.
- Where foundry limitations result in higher impeller width.
Figure 6- Concentric Casing
The cutwater showed in the figure 6 is used to reduce the recirculations in the volute. Recirculation cause erosion in the volute and additionally they will cause extra pressure drop.
4.3- Volute casing.
This kind of casing is contrary to previous method, casing is eccentric, meaning that some part of kinematic energy to pressure head conversion is done inside casing. As showed in figure 8, the distance between tip of the impeller and casing wall increases constantly. As this gap become wider, the speed will drops and pressure increases more.
Volutes can have different cross sections as showed in the figure 7. Manufacturers may prefer to use the left one since it is easier to manufacture and costs less but the limit is Ns below 1100. Methods such as sand casting is used for manufacturing the casing, thus hard patterns are costly or even impossible to cast.
Figure 7- Different cross sections for volute area.
Volute area are gradually increase. The divergence angle is between 7° to 13°. The volute casing are generally used when Ns is below 1100 US units. Since in this design the casing is eccentric, a radial force is exerted on the impeller. The manufacturers design their pump in a manner that minimum force exerted on impellers when the pump working at BEP. As the flow increase this force increases and reaches its maximum in shutoff condition.
Figure 8- Single volute casing
As showed in the figure 8, while the pump works at BEP, the forces (almost) cancel out each other.
4.4- Double volute casing.
This method is same as previous one with a modification. The volute is divided into two section. This section is made in order to reduce the forces exerted on impeller for the pumps with high flow. Generally for the pumps with flow higher than 125 m3/hr, this method is used. However, manufacturing the pumps according to this method costs more comparing to single volute.
Figure 9- Double volute casing
4.5- Vaned diffuser ring
In this method a ring consists of number of vane installed in pump casing. The distance between these vane increases (BC in figure10). The space between each two vane acts as a volute. The center line of vanes (L in the figure10) can be curved or straight. The straight line has higher efficiency but results in larger casing. The number of vanes in diffuser ring is one more than the number of impeller vanes. If the ring has too much more vanes, the vanes themselves will cause pressure drop. If total number of diffuser vanes are less than impeller vanes, the pressure will not increase the flow leaves the pump with higher speed.
Figure 10- Vaned diffuser ring
5- Other parameters
There are other parameters that affect the hydraulic of the pump such as:
- Surface roughness
- Flow pattern inside the pump casing
- Manufacturing tolerances
- Gaps between impellers and casing
It is not cost efficient to produce a pump that all above mentioned parameters are controlled, thus API gives manufacturers a final tolerance for the head of manufactured pump, figure 11:
This table meant for example for a pump with differential head of 310m, rated head in test can be between 291m to 309m and shutoff head can have +- 15m tolerance, with only 1 condition; the curve from rated point to shutoff point shall be constantly rising.
6- Final words
Although these methods are mentioned, there are many more methods that can and will be used based on new researches. These methods can be used alone or together based on manufacturer design and experience and required pump conditions. For eager minds I strongly recommends to refer to the mentioned book in the first reference.
Related Articles
- Sand Casting- A brief introduction on method and defects
- Centrifugal Pumps- Part V- Inducers
- Centrifugal Pumps- Part IV- Water Hammer
- Centrifugal Pumps- Part III- Shutoff Pressure
- Centrifugal Pumps- Part II- NPSH and Cavitation
- Centrifugal Pumps- Part I- Mechanical seals
References
- Hydraulic Institute "Hydraulic Institute Standards for Centrifugal, Rotary & Reciprocating Pumps", 14th Ed. (1983)
- API 610, 11th Edition, September 2010.
#centrifugal #pump #casing #api #volute #hydraulic
KBESS2 Deputy Project Manager | Technical Leader | M.Sc. B.Sc.| MIEAust CPEng NER
5 年Thanks for sharing.