1. Introduction.

The large-scale use of regulation by changing the rotational speed was a kind of revolution in pump technology and made it possible to achieve significant energy savings. Compared to regulation by throttling, regulation by changing the rotational speed is always more energy-efficient, which results from the principle of operation itself, because in the case of throttling, the pump first transfers energy to the liquid, and then this energy is lost on the throttling valve, while in the case of regulation by rotational speed, reduction of the amount of energy transferred to the liquid. The question is not "will speed control bring savings" but "how long will it take to pay off the costs of using such control?" For a long time, the barrier to the use of speed control was the cost of frequency converters (so-called inverters). As the scale of production increased, this cost gradually changed and now inverters have come into widespread use, becoming an almost standard component of a pump set. Unfortunately, this beneficial phenomenon is accompanied by a dangerous trend in which, when the parameters can be adjusted, the principles of proper pump selection are neglected ("because they will be adjusted using the inverter"), which limits potential energy savings.

The author intended the following comments to facilitate the optimization of the use of frequency converters and to avoid common mistakes.

The issue of regulation by changing the rotational speed has been extensively discussed in [1], and the comments contained in this text are an attempt to supplement the knowledge provided in [1] for practical examples.

2. "Movement" restrictions.

The use of a variable speed drive may cause some operational problems. Older types of motors were generally not designed to be powered by an inverter. The problems were mainly related to cooling, because changing the power supply from typical sinusoidal alternating current to pulses generated by the inverter increases heat generation in the windings, and at the same time, the effectiveness of the engine cooling fan decreases with the rotational speed. Currently produced motors are generally adapted to work with inverters, but in each case it is advisable to consult this with the manufacturer.

In the case of pumps, the most serious problem is the possibility of resonance at certain rotational speeds. Pumps were traditionally designed for a specific nominal speed so that the critical speed at which vibrations increase does not coincide with the nominal speed and is most often above it. When changing the rotational speed over a wide range, there is a significant probability that the rotational speed will coincide with the critical speed divider (e.g. half of the critical speed), which leads to an increase in vibrations. When the rotational speed changes within a very wide range, this phenomenon is difficult to avoid. However, in consultation with the pump manufacturer, it should be determined in what range of rotational speeds this can be expected and the pump should be selected so that it does not operate within this range.

When the rotational speed is significantly reduced, the load capacity of the plain bearings may be reduced. Problems with lowering the rotational speed can also be expected when complex mechanical seal systems are used, in which the flow of coolant is forced by rotating elements. These phenomena are not often encountered in practice, but when selecting the scope of regulation, the possibility of their occurrence should be taken into account.

3. Characteristics of centrifugal pumps regulated by changing the rotational speed and possible energy benefits.

W [1] i [2] describes how the parameters of centrifugal pumps change with changes in rotational speed.

If the basic characteristic H(Q) at the nominal speed is known, the characteristics for other speeds can be determined by calculating point by point according to the formulas:

Q₂ = Q₁ n₂ / n₁,

H₂ = H₁ (n₂ / n₁)²,

where the index 1 denotes the parameters at speed n₁, and the index 2 parameters at speed n₂.

This way, knowing the Q parameters₁ family₁ at a point on the characteristic at speed n₁ Q parameters can be calculated₂ family₂ at a point on the characteristic at speed n₂. It can be assumed that the efficiency is the same at both points, and therefore the lines of constant efficiency are parabolas. The assumption of constant efficiency no longer applies at too low rotational speeds, at which centrifugal pumps stop operating properly due to the fact that the ratio of centrifugal force to viscous forces is too low.

By determining the characteristics at different speeds in the manner described, we obtain the so-called shell chart as per fig.1. Of course, the characteristics at different rotational speeds can also be determined by measurement, but the results obtained are generally close to those determined theoretically. The fundamental qualitative difference is that, according to the theory, the parabolas of constant efficiency run all the way to the origin of the coordinate system as in rys.1, and in practice, as mentioned, below a certain rotational speed the efficiency decreases.

Fig. 1. Lifting height and pump efficiency when changing rotational speed.  

Looking formally at the chart as: rys.1 one can come to the conclusion that by using regulation by changing the rotational speed, each required operating point with parameters Q and H can be obtained. From here we are close to another, dangerous conclusion: in the case of using an inverter, there is no need to follow the rules for selecting pumps. This is an incorrect conclusion because the permissible operating area of ​​the pump at variable speed is limited (fig.2).

The limitation from below results from the fact that, as mentioned, below a certain rotational speed the efficiency deteriorates and therefore it is not necessary to go below a certain nmin. In turn, the limitation from above is mainly due to strength reasons. Please remember that the pressure generated by the pump, as well as the torque of the shaft, increase with the square of the rotational speed. Modern inverters are generally capable of increasing the frequency from 50 to 60 Hz, which means an increase in rotational speed by 20% and therefore an increase in pressure and shaft torque by a ratio of approximately 1.22 = 1.44. This value is generally at the limit of the pump's design reserves, so further increases in rotational speed would be dangerous. Of course, in each case it is safest to agree with the manufacturer the maximum permissible speed of the pump.

Fig. 2. Recommended pump operating range when changing rotational speed. 

For a constant rotational speed, it is usually assumed that the recommended operating range is in the efficiency range from 80 to 110% of the optimal efficiency. Outside this range, not only does the efficiency decrease, but there are also unfavorable movement effects, such as an increase in vibration and noise. A similar range of recommended work applies to each characteristic for individual rotational speeds. As a result, the recommended working area is within the area marked at rys.2 dotted line.

Due to efficiency, the optimal operating range is even narrower. The pump should operate close to the highest efficiency parabola ƞmax. This is possible if the system characteristics are similar to those in circulation systems with zero static head. These types of characteristics are typical of district heating networks, so in this application frequency converters can provide the highest energy savings.

To initially assess what effects regulation by rotational speed will have in a specific pump system, it is enough to plot its characteristics on the shell diagram of a given pump. As stated, in the ideal case this characteristic may fall in the region of highest efficiency, which occurs for a circulation system with zero static head. We often deal with systems with flat characteristics, where static lifting height dominates and losses play a minor role. In this case, the system characteristics are approximately horizontal, and during adjustment the operating point moves against the background of the shell diagram as shown in rys.2 with an arrow from the nominal point towards point 1. As you can see, in such a case, as the efficiency decreases, the operating point first leaves the optimal and then the recommended operating area.

Typical examples of such systems include:

a) Water supply systems in which a constant supply pressure is maintained
b) Pumping stations for main drainage in deep mines
c) Drainage pumping stations

In this type of pump systems, the energy savings that can be achieved by using regulation by changing the rotational speed are smaller than in circulating systems. In such cases, it is advisable to use several pumps operating in parallel. Rough capacity adjustment should be made by changing the number of operating pumps, and changing the rotational speed should be used to precisely adjust the pumping station's capacity to the requirements. As stated in [2], it is more advantageous to regulate the rotational speed of all pumps, not just one.

In general, it can be said that the use of regulation by changing the rotational speed brings the following effects compared to throttling regulation:
a) The greater is the share of losses in the total lifting height
b) The wider the range of capacity regulation, the greater it is.

The permissible operating range of a speed-controlled pump, as in any other case, also depends on the suction properties. It is therefore necessary to check whether at each operating point the required NPSHr is lower than the available NPSH. The complication is that NPSHr characteristics are usually known at nominal speed. There are no generally recognized formulas that can be used to calculate NPSH requirements at changing revolutions. From the qualitative side, the impact of turnover on the required NPSH is shown in rys.3. Reducing the speed causes the values ​​of the required NPSH to generally decrease, so adjusting the speed down usually improves the anti-cavitation reserve. The danger may be that as rpm drops, the range of low NPSHr requirements shifts towards lower performance. If the regulation by changing the speed is carried out along a line in the direction of p.2 as in fig. 2, i.e. reducing the speed leads to a reduction in the lifting height at constant capacity, in which case the suction properties of the pump may deteriorate.

Fig. 3. The influence of rotational speed on NPSHr. 

Due to the above-mentioned lack of universal formulas allowing for the conversion of NPSHr characteristics, checking the suction conditions at variable rotational speed should be done in cooperation with the pump manufacturer.

4. The inverter should not replace proper selection.

The ability to change pump parameters using an inverter should not replace proper selection. For example, if the pump is selected for too high a lifting height and therefore tends to work with excessive efficiency, using an inverter, it is possible to adjust its lifting height to the required capacity. However, this will mean adjustment along the line shown in Fig. 2 towards point 2, which will cause the pump to leave the recommended range.

The author encountered incorrect recommendations from the designer of the heating system, which included four pumps regulated by inverters. The designer recommended that in the event of a decrease in efficiency, two pumps should be turned off and the remaining two pumps should be regulated by reducing the speed. This is an incorrect recommendation, because if we assume that the four pumps are correctly selected for the maximum efficiency and the corresponding lifting height in the heating pump system resulting from the flow resistance in the network, then when the efficiency is reduced, the required lifting height decreases along the parabola. Therefore, it would be advisable to regulate all four pumps. However, if we wanted to obtain reduced efficiency from two pumps, their optimal lifting height would be higher than indicated by the system characteristics and, as a result, the pumps would work as in point 2 on rys.2 outside the optimal range. Correct operating practice in such a case should consist in uniformly reducing the efficiency of all pumps.

W [3] a case was described when inverters were used to limit the efficiency of propeller pumps in a drainage pumping station, which were selected for excessive efficiency to such an extent that the water flow was so high that the edges of the channels leading to the pumping station were damaged by erosion. This is an example of incorrect selection, which was particularly costly because larger, and therefore more expensive, pumps were purchased first, and then inverters were used to reduce efficiency.

5. Summary and conclusions.

As can be seen from the above, when regulating by changing the rotational speed, the following rules should be followed:

1. At the selection stage, you should plot the characteristics of the system on a "shell chart" and check whether the range in which the pump is most likely to operate is within its recommended operating area. Outside this range, the pump should only run occasionally.

2. The best energy effects can be achieved in circulating systems, because then it is possible to operate only in the area of ​​high efficiency. If in a circulation system the required capacity is obtained not from one but from several pumps operating in parallel, when reducing the capacity, none of them should be turned off, but the capacity of all pumps should be reduced evenly.

3. In systems with flat characteristics (i.e. with a small share of losses in relation to the static height), regulation by changing the rotational speed is less effective. In such situations, it is advantageous to use several pumps connected in parallel and roughly adjust the capacity to the requirements by switching on the appropriate number of pumps. By changing the rotational speed, the efficiency should then be precisely adjusted; it is best if all operating pumps are regulated, not just one.

4. Changing the rotational speed should not only be used to correct the selection error (e.g. reduction of lifting height and efficiency). In such cases, minor corrections can be made by changing the impeller diameter, and in case of serious errors, it is advisable to replace the pump.

Dr. Eng. Grzegorz Pakula

Literature:
1. P. Świtalski, W. Jędral, Academy of pump technology, Variable speed control, pros and cons, Pumps Pumping stations, No. 3/2012
2. G. Pakuła, Performance regulation of two pumps operating in parallel, Pompy Pompownia, No. 3/2011
3. M. Świderski, Regulation of the propeller pump efficiency by changing the rotational speed. Pumps Pumping stations, No. 4/2012

1. Introduction.

The purpose of the pump overhaul is to restore it to a technical condition similar to that of the new pump. How close the condition after renovation will be to the initial condition depends on the renovation technologies used. The choice of technologies used during renovation is subject to optimization, because the expenditure incurred must be justified by the effects they bring. This article contains information helpful in making decisions about the selection of optimal technologies.

The second element to be optimized is determining the period between renovations, because for obvious reasons, carrying out renovations with excessive frequency increases the renovation costs, and on the other hand, extending the periods between renovations leads to increased operating costs due to increased energy costs. Therefore, there is a certain optimal length of the period between renovations. The methodology for determining it is discussed in point 3 of the article.

It should be emphasized that in the case of pumps, it is a serious mistake to try to recreate only the so-called mobility, which is defined based on parameters such as the level of vibrations, temperatures of structural nodes, noise generated during operation and the lack of leaks and leaks. Energy efficiency is very important for pumps, which determines energy consumption, which is the main source of operating costs. It is possible and probable that an incorrectly renovated pump has the mobility efficiency specified above, but operates with reduced energy efficiency. Optimal renovation management requires controlling the energy efficiency obtained after renovation, because, as shown below, differences in energy consumption costs due to differences in energy efficiency are much more important than differences in renovation expenditure.

2.  Mechanisms of pump degradation during operation.

 Pump parameters, including efficiency, change during operation. The efficiency of a new pump may increase slightly in the initial period after startup, which is caused by the running-in of the cooperating elements and, above all, smoothing the surface roughness resulting from the technological process by the flowing liquid. Then, during operation, the pump undergoes processes leading to the gradual destruction of its components, as a result of which the parameters deteriorate.

These processes can be divided into two basic groups:

• changes in geometry (increase in slot dimensions, wear of blades);
• changes in the condition of the surface (increase in roughness due to corrosion and/or formation of deposits).

The influence of individual destruction mechanisms on pump characteristics will be discussed below. In fact, these processes occur simultaneously and the impact of each of them on the pump parameters is difficult to separate.

2.1. The influence of roughness increase on wading losses.

Power is transferred from the shaft to the fluid through the impeller (impellers in a multistage pump). This transmission takes place essentially inside the rotor channels, mainly via the blades. The external surfaces of the rotor discs, rotating in the liquid, also absorb energy from the shaft, but the vast majority of this energy is dissipated in the form of the so-called wading losses, which have a significant impact on pump efficiency. Progressive corrosion and/or deposit formation increases the roughness of the external surfaces of the rotors and thus increases wading losses. This effect is particularly important in pumps with low speeds, i.e. in simple terms - for pumps with significant lifting heads and moderate efficiency, such as pumps feeding boilers.

2.2. The effect of increasing the roughness of flow channels.

Due to corrosion and erosion, the surface roughness of the flow channels increases during pump operation. An increase in the surface roughness of the impeller flow channels results in an increase in power consumption and may result in a slight increase in the pump lifting height, as the liquid is "taken up" more effectively by the impeller, which, however, adversely affects the efficiency. However, in the case of flow through guide channels, gaps in individual pump stages and channels in the suction and discharge casing, i.e. through all flow channels in the non-rotating elements of the pump, the increase in surface roughness does not affect the power consumption of the pump (the power consumption is determined by the condition of the impellers), it causes however, due to the increase in flow losses, the pump lifting height decreases, because increased flow losses in the channels cause a pressure drop. The impact of flow losses resulting from wall roughness on pump efficiency is most visible for pumps with high capacities and low lifting heads, such as diagonal, propeller or double-jet pumps.

2.3. The influence of changing the geometry of the rotor blades.

 Pump impeller blades gradually wear out. This process accelerates when solid particles are present in the pumped liquid. Mainly the initial fragments are destroyed, as a result of which the entire blade is shortened, and the ends of the blades, as a result of which the outlet angle of the stream is reduced, which reduces the pump lifting height. The wear of the blades therefore reduces the pump lifting height and power consumption. The influences of these factors on the efficiency compensate each other, but overall the efficiency deteriorates due to the wear of the blades.

2.4. The effect of enlarging sealing gaps.

During operation, the dimensions (widths) of the sealing gaps increase, as a result of which the volume losses associated with return flows of liquid from areas of higher pressure to areas of lower pressure gradually increase. An example of this is the return flow from the rotor outlet to its inlet through the gap sealing the rotor neck. This has an obvious adverse effect on efficiency, as the flow through the impeller is greater than the pump's capacity by the amount of the return flow, and the energy transferred by the impeller to the liquid returning to the suction side is lost. These types of volumetric losses have particularly severe consequences for pumps with high head and low efficiency. This is due to the fact that for reasons of movement (avoiding seizure), the achievable dimensions of the sealing gaps are similar, regardless of the pump efficiency, and with similar gap dimensions, volumetric losses have a higher percentage share for pumps with lower efficiency. The amount of volume loss with the same gap dimensions increases with the increase in the lifting height from the stage.

2.5. The total effect of pump wear on its characteristics.

The effects described above occur simultaneously and their influence on the course of the characteristics overlaps. The total effect on the characteristics depends on which of the destruction mechanisms described above dominates. Regardless of this, the efficiency and head decrease as the pump wears out. The associated power consumption also usually decreases while maintaining a constant rotational speed. In order to maintain unchanged hydraulic parameters (efficiency, lifting height) despite increasing wear of the pump, the rotational speed should be increased if such adjustment is possible. An increase in rotational speed causes an increase in power consumption compared to a new pump with the same hydraulic parameters as a result of reduced efficiency. If during operation there is an increase in power consumption at a constant rotational speed and unchanged hydraulic parameters, it usually indicates mechanical problems, such as internal rubbing of parts or bearing problems.

In addition to the above-mentioned deterioration of hydraulic parameters, the mechanical condition of the pump is also deteriorating during operation, which is expressed, among others, by: increasing the vibration level. The reasons for this are loss of balance of the rotating assembly due to its uneven wear, deterioration of the condition of the bearings and gradual degeneration of fits resulting in a change in the stiffness of the pump structure.

The rate and degree of deterioration of pump parameters during operation significantly depend on the type of pumped medium, in particular on its corrosiveness and the amount of solid impurities. In the event of significant contamination (e.g. water with a significant amount of sand), the rate of wear and the depth of decline in pump parameters may be significant. However, in the case of pumping clean water, especially treated water, such as boiler water, the drop in pump efficiency after several years of operation should not exceed a few percentage points. Measurement results of feed pumps are available indicating that those pumps with output efficiencies above 80% operate with efficiencies well below 70%. Such a deep drop in efficiency when working with clean water cannot be explained by pump wear. This indicates the use of inappropriate renovation technologies, which resulted in reduced efficiency after renovation.

3. Optimal period between renovations.

The simplest way to determine the moment of sending the pump for overhaul is to operate it until the pump no longer fulfills its function, i.e. until it is no longer able to provide the required parameters or it fails. Such behavior is rarely encountered in the energy industry, where preventive planning of renovations is used to perform them before a serious failure occurs, which helps reduce renovation costs and prevents emergency downtime of the installation.

Such planning is usually based on one of the two most commonly used methods:

1. Planning based on the number of working hours;
2. Planning based on monitoring technical parameters.

The first method involves using intervals between renovations determined based on the manufacturer's recommendations or the user's own experience. In some cases, the pump renovation period may result from the renovation period of the installation in which the pump operates. Such a situation may occur, among others: for pumps operating in power units, when the pump is overhauled during the period of overhaul of the entire unit, planned not for the technical condition of the pump, but for other machines and devices.

This method is simple in principle and does not require incurring costs for monitoring systems, but it is usually conservative, i.e. it usually leads to repairs being carried out more frequently than indicated by the actual technical condition of the pump.

In order to precisely determine the moment when renovation is required, planning methods are used based on monitoring the pump's operating parameters. In the energy industry, this usually includes monitoring the vibration level and/or temperatures of specific pump structural nodes.

However, it is possible and even probable that the pump is in good operating condition, i.e. it shows an acceptable level of vibrations and temperatures, but has reduced energy efficiency. In such a case, in order to reduce the costs of energy consumption, it is advisable to carry out renovation despite the proper operation of the pump.

In practice, to determine the optimal moment when the pump should be sent for renovation, it is enough to record the number of pumped meters³ and the amount of kWh used to drive the pump. It can be assumed that the cost of operating the pump consists of two main components: the cost of renovations and the cost of energy consumed (other costs, such as maintenance costs, are of a lower order). During operation, it is necessary to monitor how both of these components affect the cost of pumping, among others³. If you divide the cost of renovation, which can be approximately fixed, by the number of m³ pumped since the previous renovation, we will obtain the renovation cost per m³. This value decreases during operation as the amount of pumped liquid increases. If the amount of energy used to drive the pump is monitored, it is multiplied by the price of a unit of energy and divided by the number of meters pumped.³ we will obtain the average energy cost per pumped m³. This value increases during operation due to the deterioration of efficiency due to the progressive deterioration of the pump's technical condition. Adding the cost of renovation and the cost of energy gives the total cost of pumping m³. This value initially decreases during operation due to the decreasing renovation component, but at some point it begins to increase - due to the increasing energy cost. From an economic point of view, it is advantageous to send the pump for overhaul when such an increase is detected, even if the pump is still functional.

The above method of determining the optimal moment to send the pump for renovation is appropriate under two assumptions. Firstly, the pump operates (approximately) at constant parameters, otherwise the change in energy consumption may result not from a deterioration in efficiency, but from a change in the position of the operating point. Secondly, the renovation will restore the initial efficiency of the pump, because otherwise the expenses incurred for the renovation will not be recovered in the form of savings on energy costs.

4. Renovation technologies.

Broadly speaking, during renovation, individual pump components can be replaced with new ones or regenerated. The decision to replace or regenerate should be made based on a visual inspection of a given element and measurements of its geometry, taking into account the impact of the condition of a given element on the pump parameters.

Since the most expensive parts of the pump are the casings, if they are not suitable for regeneration, renovation becomes pointless, as the alternative is to replace the pump with a new one. Pump bodies should therefore be refurbished. Regeneration concerns the internal surfaces of the bodies in contact with the liquid, which undergo corrosion and "washing out" during operation. The resulting increased surface roughness of the flow channels reduces the efficiency of the pump (point 2) and should be removed during renovation and major defects repaired. In the case of flow channels in the bodies, the accuracy of reproducing the initial shape is not as important as in the case of flow channels of rotors and vanes, therefore their surfaces can be regenerated. Cavities and pits can be repaired by surfacing, and roughness can be removed by cleaning and grinding larger irregularities. Application of coating preparations is also used, which allows for obtaining a much better surface smoothness than in the case of the surface of raw metal. Experience and measurements show that the use of preparations increasing smoothness allows for a several percent (usually 2-5%) increase in efficiency. The problem in this case is the durability of their adhesion, which requires strict adherence to the technology of preparing the metal surface before coating with the preparation.

As mentioned in point 2, the quality of the internal surfaces of the housings has the greatest impact on the efficiency of high-speed pumps, and for this reason, the regeneration of internal surfaces should be carried out with particular care for such pumps.

It should be borne in mind that the pump bodies are pressure elements and therefore their durability is not unlimited, because after a long period of operation, due to wear, the thickness of their walls may drop below the value ensuring the required resistance to internal pressure. In such cases, the body can be repaired by surfacing, but this requires the use of appropriate welding technologies to obtain the strength of the repaired body comparable to the strength of the new body. However, in accordance with the requirements of EU directives, the repaired pump body should always be subjected to a pressure strength test while maintaining safety conditions.

The efficiency of the pump depends largely on the condition of the impellers and guide vanes. As mentioned, due to the "washing out" of the end parts of the impeller blades, the liquid outlet angle decreases, which reduces the pump lifting height. This type of impeller wear is difficult to detect because upon visual inspection the impeller may appear undamaged and measurement of the blade outlet angle is difficult to perform. The resulting drop in parameters is only revealed during characteristic measurements. If this type of wear occurs, the impeller should be replaced because the regeneration of the blades, including the restoration of the outlet angle, is difficult to perform and, moreover, the durability of such a procedure would be questionable. In any case, verification of the condition of the rotors, including only measurements of the neck diameter, is insufficient to make a decision on replacement, because the further usefulness of the rotor is determined by the geometry of the blades, especially their outlet sections.

When replacing rotors and vanes, use original parts according to the manufacturer's documentation. Various types of "replacements", for example those made based on measurements of a worn component, can produce disastrous performance results. While it is possible to measure dimensions resulting from machining (such as outer diameter, hub and neck diameter) and, on this basis, make elements that fit the pump, precise reproduction of cast blades "from nature" is practically impossible, because cast elements are characterized by significant shape deviations, while the pump parameters are very sensitive to inaccuracies in the blade palisade geometry. The quality of the casting technology used is also of great importance, as it determines the surface roughness of the flow system, which has a significant impact on the efficiency. For example, rotors with identical geometry, but made using ceramic cores, allow for up to 3% higher efficiency compared to castings made in sand molds. Similar effects in terms of efficiency can be achieved by polishing the channel surfaces, which is usually only possible by manual processing. Sometimes, short-term pumping of water containing e.g. corundum is used, which allows for grinding the flow surfaces. The use of original rotors is also required for safety reasons. During operation, pump rotors are subject to significant loads from centrifugal forces, therefore the chemical composition of the material used, as well as the fine-grained internal structure of the rotor casting, are very important. Impellers with inadequate strength may burst when the pump is operated at high speeds, which in each case leads to complete destruction of the pump and also poses a health hazard to the personnel.

If the rotors are qualified for regeneration, it primarily involves restoring the dimensions of the sealing gaps. This is usually done by "legalizing" the rotor neck, i.e. equalizing it to a reduced size, and using an appropriate, undersized sealing ring. Selecting the appropriate dimensions of sealing gaps requires experience. The use of excessive clearances facilitates assembly, but prevents high efficiency. However, excessive tightening of the gaps may cause the pump to seize. Dynamic shaft deformations depending on the pump type should be taken into account here.

To achieve low vibration levels, it is necessary to balance the pump rotating assembly when assembled. Balancing the individual elements separately is not sufficient, because when the rotating assembly is compressed by axial force or due to the twisting of the nuts securing the individual elements, the individual elements may become twisted and the balancing will turn out to be ineffective. The design of many pumps makes it impossible to install a completely balanced rotating assembly. In order for the rotating assembly to remain balanced after disassembly and reassembly in the pump, it is necessary to use appropriate assembly technologies to maintain the relative position of the individual elements.

As a result of using non-original parts, even those seemingly less important, such as bearing housings, sliding bushes or stuffing boxes, despite the correct balancing of the rotating unit, the natural frequency of the pump changes after the renovation and operation at the previous rotational speed may be impossible, as it falls within resonance area.

For cost-saving reasons, there is a tendency to use self-made elements with simple geometry in renovations, such as spacer sleeves or even a shaft, which can be measured by nature and made on relatively simple machines. However, it should be remembered that for proper operation, such elements should be subject to appropriate surface treatment (hardening, surface hardening, etc.), otherwise they will easily degenerate, e.g. under lip seals. Cooperating pairs of elements (e.g. rotor neck - sealing ring) must be surface treated to obtain a difference in hardness, otherwise they may become fused. The shaft is a pump element exposed to significant loads, so using the wrong material to make it (e.g. not using forged rods as the starting material) may cause failure. Shafts in renovated multi-stage pumps must not only match their dimensions, but these dimensions must remain unchanged throughout the entire period of operation, which requires the use of pumps operating at rotational speeds over 3000 rpm and pumping hot liquids at 150°C. special methods of plastic forming and machining. Any change in the shape of the shaft, e.g. increased radial runout, may cause the pump to seize and lead to its serious failure.

Since the energy efficiency of the pump after renovation is of fundamental importance for operating costs, measurements of the pump parameters should be carried out after the renovation. Replacing them with the so-called an operational test checking only the mechanical condition of the machine is insufficient.

5. Economic aspects.

The last statement is justified by the proportions between the costs of renovation and the costs of energy consumed resulting from the efficiency of the pump after renovation, which will be shown below on the example of the boiler feeding pump.

For a pump operating on parameters Q = 450 m³/ h, H = 1800 m with a specific weight of water of 9400 N/m³, which parameters correspond to the parameters of 50% of the pumps feeding the 200 MW unit operating at full power, the hydraulic power, i.e. the minimum power on the pump shaft required based on the laws of physics, is 2115 kW. If the pump, after a properly carried out renovation, has an efficiency of 81%, the actual power consumption will be 2611 kW. However, if the renovation is carried out incorrectly, an efficiency of around 70% can be achieved, and the power consumption will then amount to 3021 kW. The efficiency range of feed pumps after renovation at the level of 70-81% is a realistic range that can be found in practice, depending on the technical level of the company performing the renovation. As shown in the example presented, as a result of using inappropriate renovation technologies, the increase in power consumption due to failure to obtain the proper efficiency of the feed pump may be at the level of 400 kW, which, when operating for approximately 7000 hours a year, gives a difference in energy consumption of approximately 2800 MWh per pump (and two work at the same time). The cost of this additional energy, assuming a price of PLN 200/MWh, is PLN 560. PLN and exceeds the cost of major overhaul of the pump. As you can see, differences in the prices of major renovation offered by different contractors have a much smaller impact on the operating cost than differences in the efficiency obtained after renovation.

6.  Determining the optimal parameters of the pump after renovation.

When ordering a renovation, you should specify the required parameters that the pump is to achieve after the renovation. Since, as indicated in the previous point, failure to obtain them has serious cost consequences, it is necessary to carry out acceptance tests after the renovation, including measurements of energy characteristics.

There is therefore a need to define the required parameters. The most obvious practice is to specify the pump's original specifications as stated on the nameplate. However, there may be situations when these parameters are not optimal because there have been changes in the system since the pump was installed, resulting in changes in the parameters required from the pump. A major overhaul is an opportunity to adapt the pump parameters to current requirements. Such adjustment, to a certain extent, can be carried out using simple methods that do not involve an increase in costs, such as correcting the rotor diameter. For this reason, before ordering a major overhaul, it is advisable to conduct an analysis of the pump system to verify the required optimal operating point.

7. Summary and conclusions.

• After renovation, the pump should recover not only the so-called motor efficiency, but also, and even above all, energy efficiency.

• The increase in energy costs associated with failure to achieve adequate efficiency after renovation usually outweighs the differences in renovation prices offered by different contractors.

• It is necessary to check the energy parameters of pumps after renovation at a testing station by measuring the characteristics.

• To check the energy parameters of pumps, the so-called movement test. However, the latter should also be part of the acceptance tests after renovation, among others, to check, through vibration measurements, whether the required dynamic state of the pump has been achieved, using appropriate technologies for balancing the rotating assembly.

• The quality of the renovation and the resulting durability of the pump are also influenced by the quality of the materials used to make the pump components and the technology of their surface treatment. Materials and technologies should comply with the technical conditions established by the pump manufacturer. It is not always possible to check these requirements during acceptance tests. For this reason, requirements of this type should be included in the technical specifications and their fulfillment confirmed by material certificates, dimensional certificates and/or inspections carried out by the client during the renovation.

1. Dr. Eng. Grzegorz Pakula
2. MSc. Andrzej Wesołowski

The technical specifications accompanying requests for quotations for the purchase of pumps or included in tender documentation often include the requirement for stable characteristics. Sometimes one may get the impression that this requirement appears as a simple copy from previous documents, and does not result from an analysis that would indicate whether it is justified or not. In many cases, stability of characteristics is not necessary for anything, and requiring it makes the correct selection of a pump difficult, as it eliminates pumps with good parameters but with unstable characteristics. Introducing the stability of the characteristic "forcibly" by modifying the design of the pump rotor usually leads to a deterioration of efficiency. It is therefore worth considering whether and in what cases this requirement makes sense and when it is harmful.

The pump characteristic is said to be stable if the pump head increases as the capacity decreases and reaches a maximum at zero capacity. However, if the lifting height reaches its maximum at a certain efficiency different from zero, and from this point to the point of zero efficiency it decreases, the characteristics are called unstable. (fig. 1). This is the basic, most common type of characteristic instability. (In addition to this, there may be other types of instability, manifested by an inflection of the characteristics in a different performance range, but we will not deal with such cases, which occur relatively rarely).

Fig. 1. Stable (a) and unstable (b) characteristics. 

In textbooks on pump theory, it is often stated that the instability of the characteristics leads to unstable operation of the pump, i.e. to operation with abruptly changing efficiency, which is illustrated in the following graph:

Fig. 2. Cooperation of a pump with unstable characteristics with the pumping system. (pump lifting height at zero efficiency lower than the static height of the pumping system). 

Because the system characteristic (dashed line) intersects with the pump characteristic (solid line) at two points A and B, the pump can operate in a given system at both operating points, changing the efficiency step by step. However, this is a purely theoretical case that cannot occur in practice. The situation shown in Fig. 2 can only occur if the pump is selected in such a way that its lifting head at zero efficiency is lower than the static head of the system Hst. However, such a selection is practically impossible, because such a pump could not be started at all when the pipeline is full. Moreover, in practice, the pump should be selected so that its operating point (the point of intersection of the pump characteristics with the system characteristics) falls not in the low efficiency range but near the optimal efficiency point, which lies far to the right (i.e. at higher efficiencies) than the area instability located near the maximum lifting height.

If the lifting height of the pump at zero efficiency is higher than the static height of the system, the pump characteristic (also unstable) intersects with the system characteristic only once and the pump operation with step-changing efficiency cannot take place regardless of the efficiency with which the pump operates in the system. cooperation with the system. (fig.3).

Fig. 3. Cooperation of a pump with unstable characteristics with the pumping system. (the lifting height of the pump at zero efficiency is greater than the static height of the pumping system).

It can therefore be concluded that the instability of the characteristics does not in practice lead to any problems in cooperation with the pump system, except for cases of completely incorrect selection, which are unacceptable for other reasons, regardless of the stability of the characteristics.

The answer to the question why some pumps have stable characteristics and others have unstable characteristics is not easy, as it would require a detailed understanding of the mechanism of energy transfer from the impeller to the liquid depending on the impeller geometry. The author of this article did not find a simple answer to this question in the literature. However, it is known that the stability/instability of the characteristics is related to the pump speed, and therefore to the rotor geometry. Characteristic instability is typical for pumps with low speed characteristics (slow-speed centrifugal pumps), while it usually does not occur for pumps with high speed characteristics (helicoidal, diagonal and propeller pumps), which have inherently steep H(Q) characteristics. For medium-speed centrifugal pumps with spatial curvature of the blade, both stable and unstable characteristics occur, although the latter occur less frequently the higher the speed factor. It can therefore be concluded that the instability of the characteristics of slow-speed centrifugal pumps is a natural feature. Attempts to construct such pumps with stable characteristics would mean the need to depart from proven, optimal design methods, which could result in pumps with reduced efficiency.

It can be concluded that high efficiency to some extent favors the instability of the characteristics. To justify this statement, let's imagine measuring the pump characteristics at the station shown in rys.4.

Fig. 4. Measurement of pump characteristics.

In order to determine the characteristics, we measure the power consumed by the pump (N), its efficiency (Q) and the suction pressure (ps) and discharge pressure (pt). Let's imagine that an orifice is installed between the pump discharge port and the discharge pipeline, throttling the flow. In such a system, a pressure gauge connected to the discharge pipeline will register the pump discharge pressure minus the pressure loss on the orifice. Assume that the pump characteristic measured without the orifice is unstable (solid line in Fig. 5). The pressure loss at the orifice is proportional to the square of the capacity, so the orifice characteristic is a parabola (dotted line in Fig. 5). In the measurement system shown, the pump characteristic is measured (dashed line in Fig. 5) will arise as a result of reducing the lifting head resulting from the characteristics of the pump without an orifice by the value of head losses at the orifice. As these losses increase with capacity, the characteristics of an orifice-throttled pump will tend to go from unstable to stable. This example shows the effect of throttling with the orifice at the pump nozzle only to clearly demonstrate the throttling effect. In practice, such an orifice does not occur, but a similar effect will be achieved by internal throttling resulting from the roughness of the pump surfaces in contact with the liquid. This means that treatments involving smoothing (polishing) the inside of the pump aimed at improving its efficiency result in a deterioration of the stability of the characteristics and even instability as a side effect. It should be added that the above considerations are not only theoretical, but are confirmed in practice. The author is aware of cases where reducing the surface roughness of the pump flow system, resulting in several percent (2-3%) improvement in efficiency, resulted in the described effect of deterioration of the stability of the characteristics.

Fig. 5. Pump characteristics including orifice throttling.

You should be aware that the instability of the characteristics within a certain range may be a measurement effect. First of all, the discharge pressure of the pump at zero capacity usually shows significant pulsations. These pulsations could be observed on traditional manometers in the form of pointer vibrations. In such conditions, the reading result, affecting the lifting height at zero capacity, and therefore the stability of the characteristics, depended to some extent on the subjective assessment of the person conducting the measurement. When using automatic (computerized) pressure measurements, appropriate filtering and averaging of the readings should be used, because recording a single, accidental result may lead to taking into account extreme values ​​that distort the course of the characteristics.

Another measurement effect that influences the stability of the characteristics is the pre-curve effect. To determine the pump lifting height resulting from the difference in discharge and suction pressure, the pressure at the nozzle axis should be taken into account. In practice, the pressure meter is not connected in the axis but at the pipeline wall. For low efficiencies, a reverse effect of the impeller is observed, resulting in the liquid swirling in the section of the suction pipeline adjacent to the pump (initial twist). Due to this swirl and the associated centrifugal force, the pressure at the wall of the suction pipeline is higher than at the axis. Such an apparent increase in suction pressure reduces the calculated lifting height at low capacities and may lead to instability of the characteristics. For this reason, to eliminate the influence of the pre-turn, the suction pressure measurement should not be made directly near the suction port, but at some distance in front of it. (of course, flow losses in this section of the suction pipeline must be taken into account).

As mentioned, for some types of pumps, the instability of the characteristics is not caused by the above measurement effects but is a characteristic feature of the given pump. Such instability can be eliminated by applying certain structural measures. One of them is the oblique rolling of the rotor. (fig.6).

Fig. 6. Inclined rotor rolling. 

The influence of skew rolling on the instability can be explained in such a way that the lifting height (discharge pressure) with the gate valve closed depends on the speed at which the liquid accelerated by the impeller rotates in the pump. This speed depends on the diameter of the rotor, so the bevel does not significantly reduce the head at zero capacity, because the rotor blade at the front disc still has its initial diameter. However, when the efficiency is greater than zero, the liquid flowing along individual streamlines in the rotor obtains different lifting heights, depending on the diameter of the rotor in the place where a given streamline leaves it, and therefore the lifting height achieved by the liquid flowing near the rolled disc is reduced. As a result of mixing individual streams, the pump produces a lower lifting height compared to the lifting height before the rotor rolls. The scale of this reduction increases with efficiency, resulting in a steeper and more stable characteristic. However, mixed streams with different energies cause losses and therefore this method of removing instability comes at the expense of pump efficiency.

The stability of the characteristics is also improved by bringing the spiral tongue closer to the outer diameter of the rotor. This can be explained by the fact that the lifting height (pressure) of the pump with the gate valve closed depends on the speed at which the rotating liquid "falls" into the inlet of the discharge port, where the kinetic energy associated with the speed is converted into the stagnation pressure of the retained liquid. When the spiral tongue is close to the impeller, this speed is close to the maximum spinning speed of the liquid produced by the impeller. However, if (as is the case in typical pumps) the difference between the outer diameter of the impeller and the diameter of the spiral tongue is from several to several dozen millimeters (especially with a turned impeller), then in this area the liquid spinning speed, in accordance with the free spin principle decreases with increasing radius and for this reason the speed at the inlet to the discharge port is lower than after the impeller, which causes a decrease in the pump head at zero efficiency.

As mentioned above, stability is promoted by internal throttling caused by the surface roughness of the flow system. Therefore, you can try to achieve stability of the characteristics by artificially introducing roughness, e.g. by cutting radial grooves around the outer diameter of the front rotor disc. However, this procedure clearly impairs performance.

It should be noted that pumps react quite "capriciously" to the above-mentioned methods of eliminating instability. Experience shows that in some cases stability of the characteristics can be achieved using one or a combination of several methods described, but there are cases of "persistent" instability that cannot be removed in this way.

As can be seen from the above, artificial elimination of instability in the characteristics of pumps for which this instability is a natural feature always takes place at the expense of efficiency. Therefore, you should ask yourself whether the stability of the characteristics is really necessary for something? The author was prompted to write this article by the case of a certain contract for the supply of several dozen pumps, in which the recipient established and consistently enforced the requirement of stable characteristics during acceptance tests. In the case of several pumps, this forced the manufacturer to eliminate the instability of the characteristics using the methods described (mainly by introducing an inclined slope), as a result of which pumps with very good efficiency lost their efficiency by 1-3%. As a result, the power consumed in the entire installation, as a result of the administrative enforcement of the requirement for stable characteristics, increased by several dozen kilowatts. Therefore, the question must be asked: what was it intended for?

As mentioned above, the instability of the characteristics does not in practice interfere with the cooperation of a single pump with the pumping system. The situation is slightly different in the case of parallel operation of pumps.(fig. 7).

Fig. 7. Parallel operation of pumps.

Let us consider two pumps 1 and 2 with unstable characteristics (assume that in the ideal case these characteristics are identical). Point A on the curve marks the point at which the pump achieves the same head as at zero capacity. The total characteristic of two pumps operating in parallel (1 + 2) is created by adding the capacity at the same lifting height.

If the pumps were selected in the system in such a way that the operating point of a single pump in operation would be at a capacity lower than the capacity at point A (as is the case in Fig. 7 in the case of steeper characteristics of the system a), switching on the second pump to parallel operation would be impossible, because the lifting height of the pump starting from zero efficiency would be lower than the height of the second, currently operating pump.
Correct selection is based on the fact that the operating point of a single working pump is for a capacity greater than the capacity at point A (as is the case in Fig. 7 in the case of the flatter characteristics of system b). Then starting the next pump does not pose any problems.

It should be emphasized that the latter case is a natural case, because the range of optimal pump operation is always to the right of point A, and therefore pumps should be selected in this way. However, in case of incorrect selection, the mentioned startup problems may occur when pumps with unstable characteristics are operated in parallel.

Another reason why it is advantageous to use pumps with stable characteristics is the clear relationship between head and capacity. In some cases, this simplifies automatic pump control algorithms, because the current pump operating point can be clearly determined based on a simple pressure measurement.

SUMMARY

Characteristic instability is a natural feature of pumps with specific combinations of parameters corresponding to lower speed characteristics. Pumps designed for these parameters in an optimal way, i.e. to obtain the highest possible efficiency, usually have unstable characteristics.

There are ways to eliminate the instability of the characteristics, but they always lead to a deterioration of the pump efficiency.

In general, the instability of the characteristics for a single pump does not lead to any serious problems. The stability of the characteristics only facilitates the introduction of automatic control of the pump operation, but such automatic control is also possible for pumps with unstable characteristics (which only causes some complexity of the control algorithm).

Problems may arise when pumps with unstable characteristics operate in parallel, but only if pumps are incorrectly selected for the installation (pumps operate outside the optimal range). If selected correctly, the operation of parallel pumps with unstable characteristics is not a problem.
In this situation, the requirement of stable characteristics when selecting and purchasing pumps is not always justified, because in many cases this requirement brings nothing and always makes it difficult to select a pump with optimal efficiency.

According to the author, the requirement for stable characteristics is often set not on the basis of an analysis of actual needs, but on the basis of copying previous specifications. This practice leads to energy losses and for this reason the requirement for stable characteristics should only be made when it is truly justified.

Dr. Eng. Grzegorz Pakula