Packo pump academy
The operation of a centrifugal pump is based on the centrifugal forces that arise when a liquid rotates at high speed in a housing. The construction of a centrifugal pump basically always means that an impeller rotates in a pump casing. The liquid in the impeller is carried along by the impeller blades and rotates with it. Because a radial force is created on the liquid outwards, the pressure on the outside of the impeller is greater than in the centre.
When the liquid can flow out at the outlet of the pump, the space that is released in the pump casing will be filled with liquid from the impeller. The liquid in the impeller moves from the centre to the outside and space is released in the centre of the impeller, which is occupied by newly drawn liquid from the inlet of the pump. Therefore, the inlet of a centrifugal pump is always in the centerline of the impeller.
In the impeller, the liquid rotates at high speed with the blades. So it enters the pump casing rotating around the shaft of the pump. As the fluid moves from the impeller to the outside of the pump casing, it is gently slowed down in this circular motion. This slowing down provides an extra thrust on the water, which increases the pressure even more. In science it is called that dynamic pressure is converted into static pressure. In order for this braking to take place as energy-efficiently as possible on the one hand, and to collect the liquid that flows out over the entire perimeter of the impeller on the other hand, the flow area in the pump casing increases towards the outlet. We say that the pump casing has a volute shape.
The outside diameter of the impeller determines which head (pressure) a pump will generate at a certain speed. The size of the inflow of the impeller (inlet of the pump), the height of the blades and the flow area of the pump casing all match to each other and determine at which flow the pump will operate energy-efficiently. For example, a pump with a narrow (flat) impeller with a large diameter is a pump that will mainly generate pressure at a small flow rate, and a pump with a high vane whose inlet diameter is large in relation to the outer diameter of the impeller is a pump that mainly will deliver a lot of flow at medium pressure.
The pumping action of the rotary lobe pump principle is generated by the counter rotation of two pumping elements (rotors) within a chamber (rotor case). The rotors are mounted on shafts that are kept synchronized by timing gears, in a way that the rotors rotate without contact with each other. The clearances between the rotors and between rotors and rotor case are only tenths of millimeters in order to keep internal leakage low.
As the rotors pass the suction port, the cavity generated increases creating a pressure decrease, which induces the pumped medium to flow into the rotor case. The cavity rotates over the outer diameter of the rotor case and moves the liquid from the suction to the discharge. When the rotors pas the discharge port, the cavity between the rotors is squeezed and the liquid is pushed into the discharge.
The purpose of a pump is to increase the pressure in the liquid. The liquid enters the inlet of the pump at a low pressure. During the flow through the impeller, the pressure is increased by the centrifugal forces. During the slowing down of the circumferential speed of the liquid in the pump casing, an extra thrust is created that further increases the pressure.
It is now the case that some losses always occur when entering the impeller, partly because the liquid that is not yet rotating at that moment collides with the impeller blades, which themselves are already rotating at high speed. These losses cause that the pressure when entering the impeller first drops a little and is then increased in the impeller and further into the pump casing. The magnitude of this pressure drop when entering the impeller is called the NPSH value (Net Positive Suction Head). The NPSH value is determined by the quality of the design of the pump and also the operating point of the pump. The less energy losses there are when entering the impeller, so the lower the NPSH value, the better the pump has been designed.
These losses when entering the impeller are first of all disadvantageous in terms of energy efficiency and cause a small loss of efficiency of the pump. But in addition, they sometimes have a much greater disadvantage: they can cause cavitation.
The interaction of pressure and temperature determine whether a medium is liquid or will boil and become vapor. We all know that water boils at atmospheric pressure when we heat it to 100 ° C. On the other hand, we can also boil water at, for example, 95 ° C in an enclosed space when we start to lower the air pressure above the liquid surface or, in other words, start sucking a vacuum. The minimum pressure under which a medium must be kept in order to keep it liquid at a certain temperature is known in science as the vapor pressure of that liquid at that temperature.
If we now start pumping that water of 95 ° C, we already know that a pressure drop occurs when entering the impeller. If this pressure drop is large (pump with high NPSH value), it could well be that the pressure at the inlet of the impeller falls below the vapor pressure of the water and the water at that point will spontaneously boil and thus become vapor. We say the pump is cavitating. With a better designed pump, one with a lower NPSH value, the pressure drop at the inlet is smaller, the pressure does not drop below the vapor pressure anywhere, the water does not boil and this better designed pump does not cavitate.
Cavitation is a phenomenon with many adverse consequences. First, this spontaneous boiling of the liquid at the inlet of the pump causes noise and vibration, but undoubtedly the worst part is that the pump's capacity is lost. A vapor bubble is created in the inlet of the impeller that is very difficult to get out of there. The vapor is lighter than the liquid and is constantly pushed back to the centre of the pump by the centrifugal forces and eventually the vapor bubble obstructs the passage for the liquid. As more space in the impeller is filled with the lighter vapor, the pump also loses its head because the centrifugal forces on the light vapor are smaller than on the heavier liquid.
Shear is the difference in speed between 2 layers in the liquid. The degree of the shear is expressed by the shear rate, which represents the velocity gradient in the liquid. When 2 layers of liquid that are 1m apart have a speed difference of 1 m/s, then there is a shear rate of 1/s in the liquid between these 2 layers.
As soon as a liquid is in motion, shear is created. The velocity is never the same for each liquid droplet, so there is always a relative movement of one droplet versus the other, thus creating shear. Even with so-called plug flow, shear will form on the wall because this wall is stationary or moves faster than the liquid and thus slows down or entrains the liquid.
In pumps, there are moving parts as well as stationary housing. In view of the small distance between the moving parts and the housing, a large velocity gradient or high shear rate will develop in the fluid between them. The flow will also cause shear in the supply and discharge pipes, albeit a lower one.
Shear cannot be avoided and is not a problem in itself. Sometimes a high degree of shear is even desirable. For example, in a high-shear pump a high shear rate is generated to form emulsions or dispersions, or in general, to mix liquids that are difficult to dissolve into each other more or less stably.
However, the impact of shear on the microstructure of a fluid is not always positive or desirable. Shear, for example, can also damage sensitive liquids. Due to the shear stress between the liquid particles, an impact on the viscosity of the liquid is also seen in many complex liquids.
Liquids whose viscosity is not influenced by the shear rate are called Newtonian liquids. They are typically chemically simple liquids: water, oil, sugar solutions, ...
With much more complex liquids (preparations), the viscosity will decrease as the shear rate increases. These liquids are called thixotropic. As soon as the shear rate decreases, the viscosity increases again. The shear does not destroy the liquid structure and will not change it permanently. This means that a thixotropic liquid has the highest viscosity in a tank where it is at rest, so no shear. Due to the shear created by flow through a pipe, the viscosity decreases. In a pump, where the shear rate is much higher, the viscosity is at its lowest. When the liquid leaves the pump, the viscosity recovers as the shear rate falls back.
With a small group of liquids, the viscosity actually increases with a higher shear rate. These liquids are called dilutant. Examples of dilatant liquids are honey and concentrated starch solutions. Due to their dilatant behaviour these liquids are very difficult to pump.
The influence of the shear rate on the viscosity of these three types of liquids is shown graphically below:
To generate the centrifugal forces, the impeller of a centrifugal pump must always rotate at high speed in the pump casing and that requires energy. The energy that the motor supplies to the impeller can be divided into useful energy and losses. The useful energy is the energy that the blades use to set the liquid in motion from a standstill on the suction and make it rotate with the impeller. (A small part of this energy will be lost later, but we will leave that out of consideration here). In addition to moving the liquid, energy is also needed to make the impeller rotate on its own in the liquid-filled pump casing. When you spin a disc under water, friction is created between the disc and the liquid. The layer of water closest to the disc turns with it, the water on the wall of the casing stands still. So there is a difference in speed between the layers of water, or in other words, the layers of water slide over each other. It is precisely that sliding that requires energy, which increases more the more viscous the liquid is. The power that the motor of the pump has to deliver to overcome the shear forces in the water between the impeller and the pump casing can be regarded as a total loss. At a low flow rate, little new liquid enters the impeller that has to be brought up to speed (useful energy), but the shear forces due to the rotation of the impeller are already present (losses). Therefore, at a low flow rate, the useful energy / loss ratio or the efficiency of the pump is small. As the flow increases, there is more and more liquid that must be brought up to speed. The necessary power increases, but the shear forces remain almost the same. The useful part of the energy increases, while the losses remain almost the same, thus increasing the efficiency of the pump.
When a pump is to be designed for a higher head, the diameter of the impeller must be larger: the head is proportional to the square of the diameter of the impeller. The area of the impeller also increases quadratically with the diameter. However, because the shear force acts further from the shaft of the pump (lever arm) on the one hand and the speed (and consequently the shear force) on the larger diameter on the other hand, the impact of the shear force on the larger diameter is greater. This causes the losses to increase much faster than the pressure and makes it impossible to design a high efficiency pump for a high head at low flow. In order to increase the efficiency, several smaller impellers with a smaller head (and better efficiency) are placed one after the other and a multi-stage pump is created.
The pump curve of a centrifugal pump gives the head, the absorbed power, the efficiency and the NPSH in function of the flow at a fixed speed. Usually the pump curve is published for water, or even: for a liquid with a specific weight of 1 kg / liter and a viscosity of 1 centipoise.
With centrifugal pumps there is a formula to convert the pump curve from one speed to another:
- Flow is directly proportional to speed
- Head and NPSH are quadratically proportional to speed
- Power input is proportional to speed to the power of three
- Efficiency is independent of speed
A lobe pump is a volumetric pump and therefore the flow rate delivered and the required power are initially directly proportional to the speed. This is also the case with highly viscous liquids and you can easily state that the flow and power vary linearly with the speed. The flow is the so-called stroke volume of the pump multiplied by the speed of the pump. The stroke volume of a rotary lobe pump is the volume of liquid that is moved per revolution of the rotors from the suction to the discharge of the pump.
However, with thin-viscosity liquids, leakage occurs between the rotors and between the rotors and the rotor casing, causing the pump to deliver less flow. Before explaining the influence of speed on the pump curve with thin-viscosity liquids, we will first discuss the construction of the pump curve with water as an example.
Per revolution, the pump initially moves the volume of water from the suction to the press again. However, under the influence of the pressure difference over the pump, a certain amount of water continuously flows back from the press (where there is a higher pressure) to the suction of the pump (where there is a lower pressure). This leakage flow rate is proportional to the square root of the pressure difference across the pump. So with a small pressure difference across the pump, this leakage flow rate will be small and most likely smaller than the flow rate of water the pump displaces through its rotation (the higher explained stroke volume * speed). Net, the pump still pumps water from the suction to the press.
If the pressure difference across the pump becomes large, the leakage rate can exceed the volume the pump moves through the rotation and eventually net water flows from the press to the suction of the pump. In the graph below this is the case with a pressure difference greater than 7.5 bar.
The leakage rate is independent of the speed. So if we increase the speed, then:
- the flow we transfer from the suction to the press through the rotation will increase linearly with the speed
- the leakage rate will remain the same
Since the net flow the pump pumps is the difference between the two, the pump curves will shift upwards in parallel with increasing speed.
The pump curve shows the development of the delivery head as a function of the flow rate: it indicates for each flow rate how much delivery head the pump delivers at a fixed speed. Usually we expect this course to be continuously decreasing: the greater the flow, the less head the pump delivers. Such a continuously falling curve is also called a stable pump curve. If the head decreases rapidly with increasing flow (as in the example below), this is referred to as a steep pump curve.
If the head is almost constant as a function of the flow (as in the following example), this is referred to as a flat curve.
With centrifugal pumps that deliver a fairly high head at a small flow rate, it sometimes happens that the pump curve rises slightly: the head increases slightly at higher flow rates. Such a pump curve (like the example below) is sometimes referred to as an unstable curve.
The term unstable curve is a bit unfortunate because it has a negative connotation. It is not necessarily the case that a pump with an unstable curve will work unstable, on the contrary. It is extremely exceptional that an ascending pump curve is responsible for unstable operation.
A pump always operates at the flow rate where the pump curve and the resistance curve of the pipework intersect. As long as this intersection is fairly sharp (as in the figure below), it is unambiguous and guarantees a stable flow rate.
When the pump is used to pump the liquid at a very high static height, or to pump the liquid into a vessel under high pressure, the resistance curve does not start at zero, but a lot higher on the vertical axis. You can see an example of such a situation in the figure below.
Even in this example, the pump still operates stably because there is a single point of intersection between the pump curve and the resistance curve.
Only in very exceptional cases does it happen that the static head of the system is almost the same as the head of the pump. All pressure required from the pump is required to overcome a static pressure difference. The component that represents the flow losses in the pipes is almost zero. The consequence of this is that the resistance curve is very flat and is almost parallel to the pump curve. If the pump curve itself is also flat or slightly rising, then there is no longer a sharp intersection of pump curve and resistance curve and the pumped flow will become unstable. Such an exceptional situation is shown in the figure below.
In such systems it is also not possible to control the flow rate by frequency control. Because the pump curve and resistance curve are almost parallel, a minimal change in the pump speed will result in a complete loss of flow. In the example below, you can see that at a speed of around 2900 rpm, the intersection between the pump curve and the resistance curve is not clearly defined. A very small change in the pump speed makes the flow jump in all directions and makes the system uncontrollable.
This (control) problem does not occur in the example below with a steep (stable) curve.
Why can 2 centrifugal pumps not do the double work of 1 single pump?
Many believe they can double the flow when putting 2 identical centrifugal pumps in parallel. Nothing could be further from the truth.
If you want to push more flow through a piping system or installation, you will need more pressure to overcome more resistance against that flow. And not a little bit because the resistance (= required pressure) is typically quadratically proportional to the flow. (double flow needs 4 times more pressure).
Since putting centrifugal pumps in parallel does not increase the pressure, the gain in flow will be very limited. The operation point will move from point 1 to 3 which is far less than a doubling in flow.
This can have several causes.
First, it must be checked that the pump is running in the right direction as indicated on the nameplate and manual. A centrifugal pump that turns in the wrong direction takes a lot of motor power, but delivers almost no flow. In addition, the capacity of the pump also depends on the speed of the pump. For the remainder of this explanation, we assume that the direction of rotation and speed are correct.
The flow that a centrifugal pump delivers, is always a balance between what the pump can do and what the pipework connected to the pump allows. The pressure that a centrifugal pump (at a fixed speed) creates, will decrease slightly with the flow rate, but it is quite constant all in all. However, to push a larger flow through a pipe work, more and more pressure is needed. This necessary pressure consists partly of a static height difference (you want to use the pump to pump the liquid from low to high) and partly of friction losses. The static height difference is independent of the pumped flow, the friction losses with centrifugal pumps (which pump low viscous liquids) are usually quadratically proportional to the flow. So to pump twice the flow through a pipework, the pump has to press 4 times as hard, so 4 times more head is needed.
When pumps deliver less flow than expected and than indicated on the pump nameplate, in practice the pump usually delivers the pressure indicated on the nameplate, but not the desired flow rate. The cause for this is not caused by the pump, but by an incorrect estimate of the friction losses. In other words: the pump delivers the expected pressure, but the pipework allows less flow at this pressure. In order to be able to pump the desired flow through the pipework, the pump would have to press harder, in other words more pressure is needed; requires more pressure than indicated on the pump nameplate. In other words, the pump delivers the pressure expected, but more pressure is needed to push the desired flow through the pipework.
There are 2 possible reasons why a pump does not deliver the indicated pressure:
- Either the pumped liquid is gaseous. Due to the gas phase in the liquid, the average specific gravity ρ is lower than that of water and the formula ρ * g * h produces a lower pressure. At higher gas concentrations, even a gas bubble collects in the impeller eye and the pump delivers even less pressure.
- Either the pump is cavitating. Cavitation also collects a gas bubble in the impeller of the pump.
The purchase price of a pump is often the main reason to select one or the other pump; the pumps efficiency is often forgotten or undervalued. However, in many cases you pay every year more than the purchase price of the pump to your energy supplier. Below chart shows the comparison of the initial investment and the yearly energy cost for centrifugal pumps with a differential head of 4 bar. The chart shows the comparison for pumps with different capacities, both for 40 hours and 100 hours of operation per week.
The importance of the energy consumption in the total cost over 5 years is shown in this chart:
A difference of 10% in purchase price can be a decision maker. However, a hydraulic efficiency of 60% compared to 63% means 5% more energy consumption. The energy consumption is what you’ll have to pay every year again for the lifetime of the pump. This chart shows that you will pay the 10% gain in purchase price to your energy supplier already during the first 2 years. From the 3rd year on, the penalty begins and will be repeated every year again.
In centrifugal pumps, the pressure is generated by centrifugal forces. These are themselves generated by the rapid rotation of the liquid in the pump casing. Centrifugal pumps therefore require speed to generate pressure. The greater the pressure required, the greater the internal fluid velocity in the pump.
Viscosity creates friction in the fluid located between the impeller and the pump casing. This fluid friction creates resistance to spinning. So power is needed to overcome this friction. An increasing viscosity therefore increases the power consumption of the pump. This increase is all the faster at high speeds or therefore with pumps that have to deliver a high pressure. The impact of the viscosity on the power of centrifugal pumps delivering low pressure is smaller. Therefore, a centrifugal pump is not a good choice for generating high pressures with highly viscous liquids.
With lobe pumps and screw spindle pumps, there is always a small play between the rotors and between the rotors and the pump casing. In order to make the pump as efficient as possible, these clearances are kept as small as possible, but to prevent the rotors from running into each other or in the pump casing, they are always present. These clearances form an open connection between the outlet and the inlet of the pump. With thin viscous liquids, these clearances between the rotors and between the rotors and the pump casing create an internal leak: some liquid is forced from the outlet of the pump (where there is a higher pressure) through this narrow passage back to the inlet. (where there is a lower pressure). This internal leakage rate increases as the viscosity of the pumped liquid decreases. Therefore, volumetric pumps are less efficient with low viscous liquids.
The question now arises as to which viscosity a centrifugal pump is the better choice and from which viscosity it is better to choose a volumetric pump. The answer to this question is that it depends not only on the viscosity, but also on the required head and flow. So you cannot give a fixed viscosity value where it is better to switch from centrifugal to volumetric. With a high pressure head and a small flow rate, switching is better at a lower viscosity, while at a high flow rate with a low pressure head you can stay longer with a centrifugal pump. The example below should make this clear.
When selecting a pump for 5 m³/h - 4 bar pressure, even with water (viscosity 1cP) less power is required with a rotary lobe pump than with a centrifugal pump.
When selecting a pump for 40 m³/h - 1 bar pressure, a centrifugal pump still requires less power than a volumetric pump, even at a viscosity of 1,000 cP.
In addition to the necessary power, the initial cost and maintenance costs also play a role in the choice of pump technology.
The importance of selecting and using pumps for the brewing and beverage sector as efficiently as possible is still greatly underestimated. In new projects, pumps are often selected too large (with too much reserve) or the wrong pump technology is chosen. Once in use, when changes are made to the production line, the required capacity or the process, it is forgotten to check whether the pumps are still suitable for the new conditions. This sub-optimal pump selection or operation results in too much energy being consumed and sensitive liquids (e.g. liquid yeast, mash, etc.) being damaged, or at least changing its natural structure and properties.
Everything starts with the choice of the right pump technology. The 2 most common mistakes are (1) that it is assumed that sensitive liquids should always be pumped at the lowest possible speed and (2) that the choice between a centrifugal or volumetric pump is made only on the basis of viscosity while also flow and head influence the choice of the right pump technology.
Lower rpm is not always lower speed
From the gut feeling it is generally assumed that vulnerable liquids must be pumped at a low speed. In a centrifugal pump, the pressure (head) is created by centrifugal forces. The magnitude of these centrifugal forces is determined by the circumferential speed of the outer diameter of the impeller, or in other words by the product of rotational speed and the diameter of the impeller. If we assume that the head that the pump has to deliver is determined by the process, then the circumferential speed of the tops of the blades is fixed. If a pump with a lower speed is chosen, it will inevitably have to have an impeller with a larger diameter. The speed-related shear forces and collisions in the liquid are therefore not reduced by choosing a pump with a lower speed. On the contrary, due to the larger surface area of the larger impeller, more shear and product damage occur. That is why, when selecting a centrifugal pump, the required power of a smaller pump that runs faster is often lower, especially with slightly more viscous liquids. Damage to the liquid occurs precisely when the pump transfers too much energy to the liquid. Due to the law of conservation of energy, this excess energy is converted into heat via all kinds of shear and collisions (which damage the liquid), which often has to be returned later in the process, which again costs money.
For the same reason, pumps with a vortex impeller will damage the fluid more than a conventional centrifugal pump. As the fluid continuously rotates out of the impeller into the impeller clearance and back, the fluid collides with the front of the vanes, loses speed and is abruptly accelerated back up to speed upon entering the impeller. This requires extra energy and damages the liquid. In practice, we see that pumps with vortex impellers are used to much too quickly to pump liquids with small solid particles (such as mash). University studies have shown that using classic centrifugal pumps with an optimized 3-dimensional blade design leads to a lower ß-glucan content in the mash and faster filtration. Also when pumping curd, more damage occurs with a vortex impeller compared to a classic open or closed impeller, if necessary with a slightly larger passage.
Typical high turbulence in a vortex pump causing product damage
The importance of the Best Efficiency Point
Besides paying attention to the mentioned design elements (impeller geometry and choice of speed - impeller diameter), it is perhaps even more important to ensure that the pump is operating (approximately) in its Best Efficiency Point (BEP). There is no point in buying a pump with an efficiency of over 80% and then actually using it in process conditions where it is only 40% efficient. The reason why pumps do not run in their BEP in practice is often that they are selected with too much reserve. This not only makes the pump too big and too expensive (CapEx) for what it is supposed to do, but also causes it to consume more energy than necessary (OpEx) during operation, energy that is going to heat up the liquid causing more cooling to be required further down the line (OpEx), and increasing the risk of product damage.
Does it still need to be said that, from an energy point of view, you cannot make a bigger mistake than regulating the flow by means of valves? First we let the pump generate too much pressure, for which it consumes too much energy (cost and risk of product damage), then it works at a flow rate lower than the BEP and a bit further on we throttle the pressure away in a control valve. This very concentrated energy conversion of pressure into heat in the valve causes additional product damage. By setting the desired flow rate via speed control, it is possible to ensure that the pump runs in its BEP, only the strictly necessary energy is used and less product damage occurs. The power consumption of a centrifugal pump is proportional to the third power of the speed. This means that the speed only needs to be reduced by 20 % to save half the power!
Effects of speed control on power consumed
The right pump technology
Just as power consumption (the efficiency of the pump) should be the most important parameter when choosing between different impeller types and sizes of centrifugal pumps, power consumption must also be evaluated when choosing between a centrifugal pump and a volumetric pump. In a centrifugal pump, speed is needed to generate centrifugal forces. So the higher the pressure demanded, the higher the velocity and consequently the shear forces. In a volumetric pump, the flow rate is proportional to the speed. Shear forces are therefore more likely to be flow-related. It is important to be aware of the completely different cause of the energy loss and consequent product damage between centrifugal and volumetric pumps when choosing a pump technology. That is why, for pumping a small flow rate at a high pressure, a volumetric pump can be the right choice and for pumping a very large flow rate of the same liquid at a low head, a centrifugal pump should definitely be chosen.
Even in the case of volumetric pumps, a pump with a lower speed will not necessarily damage the liquid less. In order to pump the same flow rate at a lower speed, a larger pump is required. Here too, the difference in peripheral speed of the outer diameter of the rotors is rather limited (if any). This will still create the same shear force in the fluid entering the gap between the rotors and the casing. As the pump is larger, more fluid enters these gaps and undergoes this adverse shear. With low-viscosity liquids, a larger pump will also have more internal leakage, which is again a loss of efficiency and leads to higher energy consumption and more risk of product damage.
More energy-efficient pumps pay for themselves
The risk of product damage when pumps are not energy-efficient has been described and proven above. In addition, the energy cost is also becoming increasingly important. High efficiency motors (IE3 – IE4) are rightly required by law. However, by switching from IE2 to IE3 and IE4, only a few to 5 % of energy efficiency is gained. By a correct pump selection (correct pump technology, correct impeller type and use of pumps in their BEP) tens of percents of energy can be saved. Did you know that a centrifugal pump that runs 40 hours a week, consumes its own value of energy after 1 to 3 years? Centrifugal pumps that run 100 hours a week annually consume a value of energy that is higher than their own value. In practice, we see examples of investments in more energy-efficient pumps that pay for themselves within a year.
When choosing pumps in the brewing and beverage sector, energy efficiency should be the determining factor. Always choosing the pump with the highest efficiency automatically selects the right pump technology, minimizes product damage and saves energy throughout the pump's life. The initial purchase price of the pump is in many cases responsible for less than 10 % of the total life cycle cost of the pump.