2.4 Effect of the operating point location on the pump mechanical reliability
2.4.1 Reliability-based recommendable operating region of a VSD pump
Laboratory measurements were carried out for a Sulzer laboratory pump to determine its Rq curves at different rotational speeds. The pump was driven by an 11 kW ABB induction motor and an ABB ACS 800 frequency converter. The pump was connected to a system consisting of two water tanks, venturi tubes, piping and control valves as shown in Fig. 2.16. The pump was driven at seven different relative flow rates and at five rotational speeds ranging from 1080 to
1620 rpm by adjusting its operation with the frequency converter and control valves. In these operating point locations, the NPSHA/NPSHR ratio was at least 3, SE ratio was below 0.5, and there were no signs of cavitation, and thus, cavitation occurrence should have no effect on the measurement results. The relationship of the pump rotational speed and Rn was not tested during these measurements.
Venturi tube Control
valve
Water tank Water
tank
Fig. 2.16: Laboratory pumping system used in the evaluation. It comprises a Sulzer centrifugal pump, an ABB induction motor and an ABB frequency converter (not shown in the figure). The pump is attached to a system consisting of two water tanks and venturi tubes, piping and control valves.
Vibration of the pump was measured in each operating point location by a SKF CMSS786A acceleration sensor that was installed to the location of pump bearings. The measurement data was stored to a computer with a National Instruments (NI) PCI-6032E data acquisition card.
Measurements were carried out with a 20 kHz sampling frequency for a time period of five seconds. In addition, the pump has been equipped with pressure sensors for the determination of the pump flow rate and head, as described in Appendix B.
The measured vibration acceleration was numerically integrated into vibration velocity, andRq values were then determined for the pump based on the measured magnitude of vibration velocity at the pump blade pass frequency (BPF). This was selected as the indicator for the pump reliability, as it should be mainly affected by the pump operation, and it has been used as the Rq indicator in (Erickson, 2000). Calculation of Rq was carried out with the following equation:
v C R = − v +
max
q 1 , (2.21)
wherev is the magnitude of vibration velocity, the subscriptmax denotes its maximum value, and C is the constant to set the maximum value ofRq to 1.
Rq values of the Sulzer pump are given with their quadratic polynomial curves in Fig. 2.17.
Compared with the Rq curves published by Stavale, the resulting curves provide similar conclusions: a decrease in the rotational speed has an increasing effect on the Rq values, and generally the Rq curve shapes become flatter with a lower rotational speed. In addition, the maximum value of Rq is typically attained at the relative flow rate of 70–80 % instead of the QBEP. This might be partially caused by the location of pump energy loss minimum, which is located around the 90 % relative flow rate. A presumably erroneous exception to the Rq maximum locations is theRq curve at 1380 rpm, which attains its maximum at the 30 % relative
flow rate. The limited number of measurement points and the scatter of measured vibration velocities have also resulted in a possibly too large Rq maximum value at the 1620 rpm rotational speed. Because of these factors, the use of these Rq curves as an indicator of the recommendable operating region is questionable. However, exemplary limits of the recommendable operating region have been determined for the pump with the Rq threshold of 0.7. This has resulted in relative flow rate limits of 0–110 % at 1380 rpm, 48–98 % at 1560 rpm and 58–103 % at 1620 rpm. Compared with the HI guideline for POR and Barringer’s reliability curves, especially the minimum limits of the relative flow rate are smaller, and they are in the region with a higher specific energy consumption. The maximum limits of the relative flow rate, however, are more logical excluding theRq maximum at the rotational speed of 1620 rpm compared with theRq maximum at 1560 rpm.
0 20 40 60 80 100 120 140
0 0.2 0.4 0.6 0.8 1
Relative flow rate (%)
R q
1140 rp m 1260 rp m 1380 rp m 1560 rp m 1620 rp m
Fig. 2.17: Measurement-basedRq curves as a function of relative flow rate for the Sulzer pump at five different rotational speeds. The maximumRq is typically attained at the relative flow rate of 70–80 % instead of theQBEP.
Determination of the SulzerRtotal requires the use of an approximate model forRn, as it has not been determined during the laboratory measurements. Bloch and Geitner’s model has been selected for this case with the following equation:
8 . 1620 0
n =1− n ⋅
R . (2.22)
Based on the measurement-based Rq values, an assumption that RNPSH = 1 and (2.22),Rtotal of the Sulzer pump in each measured operating point location was determined. The resulting values with their quadratic polynomial curves are shown in Fig. 2.18. As previously, Rtotal curves show the benefits of driving the pump at a lower rotational speed. However, also these results demonstrate the arbitrary nature of reliability indices and difficulty of selecting appropriate thresholds for Rtotal: for instance, the Rtotal threshold of 0.15 would limit the recommendable operating region to 39–107 % ofQBEP at 1560 rpm. At 1620 rpm, the resulting
limit would be 65–95 % ofQBEP. On the other hand, the same threshold allows the pump usage at all relative flow rates, when the pump rotational speed is 1140–1380 rpm.
0 20 40 60 80 100 120 140
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Relative flow rate (%) R total
1140 rp m 1260 rp m 1380 rp m 1560 rp m 1620 rp m
Fig. 2.18: Rtotal curves of the Sulzer pump as a function of relative flow rate at five different rotational speeds. Because of the arbitrary nature of reliability indices,Rtotal-based determination may provide rather arbitrary limits for the recommendable operating region.
Based on these analysis results, Rtotal and Rq-based limits for the recommendable operating region have been added to the Sulzer pump’sQH characteristic curve. They are illustrated in Fig. 2.19 with theEs-based limit of the pump recommendable operating region. In this case, the Rtotal criterion has resulted in wider flow rate limits for the recommendable operating region than theRq criterion of 0.7.Rtotal-based limits are also more logically affected by the rotational speed change than the Rq-based limits. However, these reliability indices do not directly represent the reliability figures of the pump, and therefore the limit values can be rather arbitrary. Consequently, further research should be carried out concerning the applicability of reliability indices in this kind of an application.
In general though, the discussed energy efficiency and reliability-based criteria for this pump have resulted in limits inside which the pump should be driven, so it would not be wasting energy or it would not be prone to a mechanical failure as a result of the higher vibration of the pump. This consideration does not include the possibly limiting effects of the system, fluid characteristics and other factors, but it gives basic information on the recommendable operating region of a VSD centrifugal pump.
0 10 20 30 40 50
Fig. 2.19: Recommendable operating region of the Sulzer pump based on the energy efficiency and reliability criteria.