by Roland McKinney
Most process industries depend on rotodynamic pumps to achieve their business goals. Unfortunately in many cases, the approach is to “fit and forget” (at least until a critical repair is needed). There can be severe cost penalties associated with this approach, of which reduced productivity is probably the greatest as well as higher than necessary energy and maintenance cost.
Given that so many processes are dependent on pumps, how can it be that personnel in so many companies spend so little time optimizing their pump systems? The reasons are complex, but amongst the most important is cynicism. In a never ending quest for cost reductions, management may introduce successive new initiatives, such as LEAN, Management by Objectives, Six Sigma, Total Quality Management, Total Productive Maintenance, etc. While each of these can have a valuable impact on a business when well understood and implemented, all too often they are introduced as a “silver bullet” without providing the time and resources to get a proper understanding of the technique. When this happens, the initiative fails and employees view each as the latest management fad, and know that they can wait it out before resuming normal operations. Each poorly implemented initiative boosts cynicism in employees.
Cynicism is the greatest obstacle to the recognition of opportunities. A classic example is the energy consumption bounce following energy saving initiatives. Countless energy efficiency initiatives lose momentum, resulting in an energy consumption bounce, with consumption returning to where it was before the start of the energy efficiency program.
So what can be done to dispel cynicism and to recognize and capture opportunities? First, it is very important to thoroughly understand the system being implemented, whether it is an energy efficiency program or a management technique. Secondly, benefits must be measurable with an implementation schedule developed against quantifiable targets. Third, feedback from implementation must be recognized so that appropriate changes in implementation can be developed. Finally, the costs of implementation must be recognized, and it is this that is probably the most important element in dispelling cynicism. It is pointless for management to implement a new policy and say that no additional funds are available to assure proper implementation - if this happens, it will entrench cynicism not reduce it.
This article illustrates benefits that can be gained from a pump system optimization program. This will help individuals recognize whether or not optimizing pump systems within their process can bring worthwhile benefits.
Generally, optimization leads to reductions in energy use, providing lower operating costs and often it is the anticipated energy cost saving that is used to justify a project to improve a pump system. In fact, benefits extend well beyond energy savings, so savings from optimization programs tend to be underestimated. This can mean that some worthwhile projects are abandoned as they do not appear to meet return on investment hurdles when only cost reductions through energy savings are used.
A comprehensive energy management program should address all significant areas of energy consumption. Pump systems often account for a large part of electrical energy consumption, as shown in Figure 1. In this tissue mill (capacity 25,000 metric tons per year), while most of the installed power was due to tissue machine and rewinder drives, nearly 30% of the installed power was from pump motors.
Although the proportion of electrical power consumption varies according to the process, this example reveals
the significance of pump electrical power. In this mill, a 10% saving in electrical power consumption in pump systems would have a significant impact on energy costs alone. However, savings would not be limited to energy costs and when promoting pump optimization studies, it is important that all cost savings are recognized.
When a comprehensive energy management program is implemented, some capital must be made available, ideally based on energy reduction targets. For example, if implementation is predicted to reduce energy costs by X dollars in the first year, ideally 50% of X dollars should be made available for energy improvement projects. Assuming that targets are met, a percentage of the savings should then be made available in subsequent years for energy efficiency projects.
Lower energy consumption reduces associated emissions of the greenhouse gas, carbon dioxide. In areas where carbon emissions carry a cost penalty, this ultimately will provide a further cost reduction.
Relationship Between BEP and Reliability
The point at which the highest proportion of energy from the shaft of a pump is transferred to the fluid being pumped is the Best Efficiency Point (BEP) of that pump. At this point, internal forces in the pump are minimized. As the operating point of a pump moves away from the BEP, this energy transfer efficiency falls and axial and radial forces increase. Although some energy loss is inevitable (for instance, friction at wetted surfaces, at seal faces and within bearings) some of the energy that is not being transferred from the shaft to the fluid is wasted. Of this wasted energy, some is lost by internal recirculation within the pump and some as heat or vibration. These are destructive forces, and so the relationship between reliability and energy efficiency is clear - as more energy is wasted, destructive forces increase in intensity, so reliability is reduced.
Research has been carried out into the relationship between the operating point of a pump in terms of its BEP and measures of reliability, such as the Mean Time Between Failure (MTBF). Some results are depicted in Figure 21, showing that flow (as a percentage of BEP flow) has a marked influence on MTBF. For instance, if the pump was operating at its BEP, its MTBF is 10 times that expected when the flow is 70% of BEP flow.
The reliability curve shown in Figure 2 is an example of a Weibull probability distribution, and it can be seen that it is skewed in that there is a steeper gradient at flows above BEP, and a longer “tail” at flows below BEP. Weibull analysis is used extensively in reliability engineering to make predictions about a product’s life characteristics, such as reliability or probability of failure at a specific time; the mean life, and failure rate. From Figure 2, it is clear that flow rate has a marked impact on pump reliability. When a pump is operating away from its BEP, its efficiency is lower than at BEP so the link between reliability and energy efficiency is simple - an inefficient pump system is an unreliable pump system.
On the basis of anticipated life, maintenance costs at different flow rates can be estimated, and these are shown in Figure 3, based on a typical pump life (η) of 35,000 hours. Maintenance costs at flows equivalent to 70% and 115% of BEP are almost 10 times those when the pump is operated close to its BEP. Maintenance costs are based on those estimated by Dupont2 from their experience with a chemical transfer pump, a pump rebuild estimated to cost $4,500.
In addition to higher maintenance costs, the costs of unreliability due to forced shut downs can be high. The cost of unplanned shutdowns vary, and in many cases are difficult to quantify. During a one year survey by International Paper at several of their mills, it was found that in these mills they had 101 pump reliability incidents, at a total of $5 million. Although the value of each of these varied, the average cost per breakdown was almost $50,0003. International Paper did not relate reliability with off BEP operation, but this was likely to have been a factor in some of the failures experienced. Assuming this to be the case, and using the MTBF data from Figure 2, allows a value to be placed on failure - where this occurs as a consequence of off BEP operation.
On this basis, the costs of poor reliability outweigh higher maintenance. And there can be no doubt that an inefficient pump system is a costly pump system.
Relationship Between Energy Efficiency and BEP
Using the same type of pump used to estimate maintenance costs, it is possible to estimate energy costs at different flows, and these are plotted in Figure 3 together with estimated maintenance costs. It can be seen that energy costs are much greater than maintenance costs at all flows, but energy costs fall as flow is reduced, suggesting a better outcome at lower flows. This is simple because lower volumes are being pumped and so even though pump efficiency is low, the energy cost goes down as the volume being pumped declines.
Of course, this lower cost is deceptive and there are several ways to show the energy cost penalty incurred by operating a pump away from its BEP, such as the additional cost to pump the same volume as is pumped over a fixed time period at BEP; or the additional cost incurred through pumping the nominal flow at the appropriate pump efficiency compared with the cost of pumping the same flow at BEP efficiency.
These are both illustrated in Figure 4, expressed as the difference from the energy costs at BEP operation for each option. Due to the lower specific energy (energy needed to pump a specific volume, such as kWh/100000 gals), pumping costs are lower when flows are higher than BEP, which is why the cost of pumping the same volume as that pumped at BEP is negative relative to costs at BEP. However, at flows higher than BEP maintenance costs increase, suggesting it is not possible to take advantage of the lower specific energy at high flows.
Real world situations for each of these conditions are when a discharge valve has been throttled in and is left in this position, so that a pump has to run for extended periods to pump a fixed volume, such as a transfer pump. In the second case, a discharge valve is throttled to reduce flow even though operation at lower speed could deliver the same volume but at higher pump efficiency. This is a very common real world situation.
There are clear differences between these two circumstances, and they reflect the complexity of real world situations and the need for an understanding of the real pump duty before deciding on “efficiency improvements” to a pump system.
Energy use is associated with carbon dioxide emissions, and in many parts of the world there are additional costs due to the emission of this greenhouse gas. This adds to direct energy costs, and so it follows that if energy is being wasted, carbon dioxide emissions and associated charges can be cut by improvements in energy efficiency of the pump system. As this varies so much and is not yet a cost in the USA, this has not been added to the benefits of pump system optimization - but in years to come this will likely become a significant cost associated with energy use.
Total Costs of Off BEP Operation
Having derived costs associated with inefficient energy use, cost of failure and maintenance, it is possible to present an estimate for the total benefits that can accrue from a pump optimization program, excluding the benefit associated from lower carbon dioxide emissions, and this is presented in Figure 5.
Energy costs used in deriving this total were those that would accrue from pumping the nominal flow, but at BEP efficiency. In this specific case, the pump has a relatively flat efficiency curve and so energy costs are low compared with other costs. With other pumps this is frequently not the case, and energy cost savings can be much higher than those shown here.
Although data has been taken from different sources, Figure 6 illustrates the benefits that may accrue from pump system optimization. Each individual case will be different, but it is very important to note that energy savings are only one of a possible range of benefits.
This analysis is largely based on estimates of pump failure rates in Figure 2, and so it is reasonable to ask - what other information is there to support these conclusions?
Another way to relate pump reliability to flow rate is through the use of “Reliability Factors”. This concept extends beyond flow rate to include other hydraulic factors such as pump speed, suction energy and NPSH margin ratio and these have been discussed elsewhere4. The analysis below is limited to flow rate, but this does not imply that these other factors associated with pump reliability are not important.
Reliability factors are non-dimensional numbers used to provide a relative index ranging from 0 - 1 of one attribute as compared to the ideal for that given attribute. A rating of 1.0 indicates this is the best selection possible in terms of that attribute. Similarly, a rating of 0 suggests that this would not be a good selection, but does not indicate zero reliability.
The derivation of reliability factors was determined through both laboratory tests and field analysis of process pumps, and a summary of results is shown in Figure 6. The reliability factors are plotted against flow, as a percentage of BEP.
It can be seen that the shape of the curves are similar to that developed by Barringer and Associates, shown in Figure 2. Note that at low flow rates, the slope of the reliability factor curve is less, suggesting a longer pump life at lower flows. Another difference is that field data suggests an optimum reliability factor at 90% of BEP flow, not at BEP flow, whereas laboratory data suggests the optimum is at BEP. In this case, field data was based on 48 split case pumps in two process plants. It has been shown that reliability factors vary with pump type and capacity, but in general the shape remains similar to that illustrated in Figure 6.
The shape of reliability factors as determined under laboratory conditions is illustrated in Figure 7. The reliability factor was based on impeller vane pass vibration data, with the lowest level of vibration giving the highest reliability factor5.
During the laboratory studies, 100% pump capacity is at or very close to BEP and the general trend is again clear, with the reliability factor reaching a maximum at capacity (in other words, equal or very close to BEP).
However, there is one interesting inconsistency - for two of the three pumps tested, results showed that at flows above BEP, the reliability factor did not fall sharply. This was suggested by the authors to be due to high NPSHa (Net Positive Suction Head Actual) during the tests. If this finding was confirmed by analysis of field data it could have a profound effect on the selection of a “best operating point”, that is, lowest cost operating point. This is because at high flows (above BEP) specific energy is lower, so energy costs can be lower. At the moment, the most economical operating point is assumed to be the BEP. But this data suggests that in fact this may not always be the case, that it may be around 120% of BEP. In other words, it may be cost effective to allow maintenance costs to increase slightly in order to operate with lower energy costs. This finding could also suggest that in order to optimize energy consumption, efforts should be made at the pump design and installation stage to increase NPSHa well beyond NPSHr (Net Positive Suction Head Required).
This approach can also compare the relative costs of maintenance and energy costs, and the additional costs incurred due to off BEP operation. In the case of API pumps, Barringer developed different reliability values, as shown in Figure 86, though again this shows a strong relationship between flow and reliability. However, unlike the laboratory data at flows above BEP, the reliability decreases very rapidly with increased flow.
This apparent inconsistency between laboratory- derived reliability factor data and Weibull life derived from field experience may be explained by the fact that few installations provide a situation in which NPSHa>>NPSHr.
After selecting a pump to match the characteristics of one of the laboratory pumps, it is possible to complete the analysis of costs incurred as a function of operating point, as was done previously, and these results are shown in Figure 9. Again, the benefits of optimization are evident: as the operating point moves further away from BEP, the greater the benefits optimization will bring. In this case, the single greatest cost saving is from energy reduction and this increases sharply as the flow decreases. In this chart, the lowest flow is at 70% of BEP but many pump surveys have shown that pumps operate well below 70% of their BEP flow.
Although the data suggests that in any optimization program the best operating point should be BEP, this is based on cost savings that can be achieved by moving the operating point closer to BEP and tends to mask other potential benefits. Specific pumping costs (costs to pump 100,000 gal) show a slightly different story, and these are illustrated in Figure 10.
Energy costs are the highest, due to the higher reliability rating for this type of pump. As specific costs fall with higher volumes being pumped, even with the lower pump efficiencies due to off BEP operation, energy costs decline with increasing flow within the range illustrated. This again suggests that the best operating point may not always be at the BEP, as lower energy costs outweigh slightly higher maintenance costs and possibly even failure costs. If it were shown that pumps like this did not suffer from a reduction in reliability at flows higher then BEP when NPSHa>>NPSHr, the trend would be even more evident. In areas where energy costs are high, this trend is also likely to be accentuated. Properly analyzing and selecting a pump requires a confident understanding of the required duty and the cost of failure. The best energy driven business solutions are not always at the point of absolute lowest energy cost.
Internal Wear as a Factor in Reduced Pump Efficiency
Operation away from BEP is not the only reason for reduced efficiency in pump systems, and reasons for this include wear of internal components (through erosion, cavitation, corrosion, etc), increased friction losses in piping (due to partial blockages, corrosion, deposition of calcium carbonate), etc. Many of these changes occur slowly, and so are not easily recognized at a single point in time as having occurred, but they can have a significant impact on pump system efficiency. Opportunities to save energy by replacing or repairing system components are frequently overlooked. There are many examples of this type of effect in the literature, and one of these is summarized below as an indication of the loss of efficiency that can occur.
In an audit of their pumps, Monroe County Water Authority discovered that many of their pumps were operating below the OEM specifications, often by as much as 20% below the OEM rated efficiencies7. One of the pumps had an OEM rated efficiency of 88% at its BEP, but after 6 years in service this had dropped to 77.8%. Mechanical refurbishment plus sand blasting and coating increased the pump efficiency at BEP to 88.5%, equivalent to an annual saving of more than $17,000 at a power cost of $0.085/kWh. The causes of reduced efficiency were mechanical wear and tuberculation on wetted internal surfaces.
Another example is shown by the worn impeller illustrated in Figure 11 - the efficiency of this pump was deteriorated over time and was only detected when the bearing failed. The energy wasted in getting to this stage must have been considerable.
SKF has recently launched a service - “Energy Monitoring Service - Pump Systems” - that is designed to help pump operators track efficiency, as well as actual operating efficiency relative to BEP. Periodic measurements of pressure, flow and power for each pump being tracked provide the data needed for the analysis, which is completed by SKF @ptitude software. This is based on SKF’s Operator Driven Reliability concept and provides plant operators with the information they need to establish the optimization opportunities available within their pump systems. It can be incorporated into any condition monitoring program - and in fact SKF recommend that this should be done, as condition monitoring does not provide all the information needed to assess the health of a pump: energy efficiency is a key component of the information needed.
Pump system optimization provides real opportunities to improve reliability and to reduce costs. There are many tools and training courses available to help provide the knowledge needed to implement a program to realize these opportunities. Where investment is needed to capture the benefit, the full array of cost savings should be estimated to provide a realistic ROI for that investment.
In most cases the best operating point will be at the BEP of the point, but in some specific cases the lowest total operating cost may be at an operating point other than the BEP. Circumstances that may support this include regions with high energy costs or where NPSHa>>NPSHr.
Given that energy costs are by far the largest component of the life cycle costs of owning a pump, more research should be targeted at establishing why some pumps can operate beyond their BEP without an increase in maintenance costs or reduction in reliability and this should include field studies.
Roland McKinney is a Senior Consultant, Environment and Sustainability, with SKF. As a pump system specialist, Roland McKinney specializes in improving the energy efficiency of pumping systems and in training other SKF employees in this and other areas of pump operations. Roland previously spent 30+ years in the Pulp and Paper Industry, working mainly in areas such as the design of paper recycling facilities, tissue production, water and waste water treatment. In many of these roles, Roland was responsible either directly or indirectly for the specification of pumps used in these systems. Roland has wide experience in system start up, having participated in the start up of mills in the UK, USA, Algeria, India and Hungry. He can be reached at roland.mckinney@ skf.com
1. Barringer, P., A Life Cycle Cost Summary, ICOMS, Perth, 2003
2. Hodgson, J., & Walters, T., Optimizing Pumping Systems to Optimize First or Life Cycle Costs, Proceedings of the 19th annual pumping Symposium, 2002 Houston, TX
3. Segrest, M., The Growing Push for Energy Efficiency in Pumping Systems, Pumps and Systems, May 2008
4. Budris, A.R; Sabini, E.P; Erickson, R.B, Pump Reliability - Correct Hydraulic Selection Minimizes Unscheduled Maintenance, PumpLines, Fall 2001 Publication by Goulds Pumps
5. Erickson, R.B; Sabini, E.P; Stavale, A.E, Hydraulic Selection to Minimize the Unscheduled Maintenance Portion of Life Cycle Cost, Whitepaper download from Goulds Pumps
6. Barringer, P, Pump Practices and Life 2004
7. Verosky, K; et al, Pump Refurbishment and Coatings Pump & Systems, November 2008