This study establishes the relevant characteristics of sludge and how they affect the performance of pumps with a swept-back leading edge impeller and relief groove in the volute and head losses in straight pipes.

It is generally perceived that pump technology evolves gradually in the continuous improvements of products. The improvements are driven by Life Cycle Cost (LCC), in particular by the energy cost awareness.

Still, experience shows that innovations in product development could have a game-changing impact. In pumping complex fluids, the influence of the fluid characteristics on the pump performance must be considered. This involves both pump design knowledge and system know-how, and it is challenging.

For example, several years ago a pump with a swept-back leading edge impeller and relief groove in the volute was developed to improve the clogging resistance in sewage applications in combination with an efficiency level similar to state-of-the art clean water centrifugal pumps. An extensive development program resulted in a twin-blade semi-open impeller.

A centrifugal pump with this type of impeller reaches 80 percent pump efficiency in sewage applications. The wet end volute is traditionally designed, but with a relief groove in the insert ring. The stationary relief groove and the design of the impeller leading-edge is the reason for the excellent clogging resistance in sewage applications.

In this application, the clogging effect refers to particles in the wastewater, such as rags or flocs, causing unscheduled blockage in the pumps. The wear resistance is a core issue in pump technology. For this pump, accelerated laboratory wear tests as well as wear field tests showed promising results. The relative decrease in efficiency compared to conventional single-vane pumps was reduced by approximately 50 percent [1].

Several case studies conclude that a pump with a swept-back leading edge impeller and relief groove in the volute significantly reduces LCC in sewage applications, where displacement pumps (PC-pumps) traditionally have been used.

Rheology

Rheology is the science of fluid deformation and flow of materials. Rheological studies are when fluids are subjected to an applied force per unit area, stress, during a certain time. A fluid deformation is either due to an extensional force or a shear force that lead to a fluid stress. The resulting fluid response is of elastic or of viscous behavior. The behavior is given by a non-dimensional time ratio. In rheological studies, the behavior in the range between the ideal elastic and the viscous fluids behavior is of interest. A way to present rheological data is by a flow curve. The flow curve shows the stress ( ) versus the deformation rate ( ).

In Figure 1, this is shown for shear induced stress versus a shear deformation rate (shear rate). The flow curve slope represents the shear viscosity, thus shear stress to shear rate. In the case of linear behavior, the slope is constant and the viscosity is simply . These fluids are known as Newtonian fluids.

[[{"type":"media","view_mode":"media_large","fid":"295","attributes":{"alt":"In the case of non-linear behavior, or non-Newtonian fluids, the slope changes and an apparent viscosity are used.","class":"media-image","id":"1","style":"float: left;","typeof":"foaf:Image"}}]]

In the case of non-linear behavior, or non-Newtonian fluids, the slope changes and an apparent viscosity are used [2]. The apparent viscosity is the specific slope for every shear rate value with its origin in origo. In order to describe non-Newtonian fluids, the behavior has to be described with several parameters. One frequently used model is the power law model, Equation (1):

(1) where K is the coefficient of rigidity, n is the power exponent, is the shear stress, and  is the shear rate. In the case of n = 1, the Newtonian case, K represents the constant shear viscosity.

In Figure 1, a non-Newtonian fluid may further show an increasing apparent viscosity, so-called shear thickening behavior. The equivalent for a decreasing apparent viscosity is known as shear thinning. Finally, the yield stress is some very specific behavior for some non-Newtonian fluids. This yield stress point has to be exceeded to result in a flow motion.

Wastewater treatment plants (WWTP) involve a wide range of fluid transports of suspensions with different flow characteristics. The aqueous suspension changes due to the concentrations of particles and dissolved substances. However, the suspension of interest is the one with non-linear behavior, i.e. the sewage sludge. Additionally, the non-Newtonian characteristics of sewage sludge vary depending on its origin and process history.

Processes where polymer additives or mechanical steps like sludge drainage are used change the sludge characteristics. The way in which the origin is a municipal or an industrial sludge influences the sludge character. However, in a recent study [3], it is shown that municipal sewage sludge shows essential shear-thinning behavior.

The purpose of this project is to study the performance of a pump with a swept-back leading edge impeller and relief groove in the volute as influenced by handling the non-Newtonian fluids in WWTP applications. Also, the project involves rheological characterizations of sludge which are not further reported in this paper. The pump performance is studied in both laboratory tests and in field tests of on-site WWTP installations.    

Experimental Set-up and Methods

This experimental study includes two parts: pump tests in a flow loop and tests in WWTP in Sweden, Norway and Finland. The pump pressure head, electrical input power, flow rate and the pipe pressure loss were measured.

In the closed flow loop, we tested several fluids. Water was used as a reference fluid. Tests with an aqueous polymer (Carbopol 676) mix and sludge of 3-6 percent total solids (TS) contents were carried out. Two different pumps with a swept-back leading edge impeller and relief groove in the volute were performance tested, a 5-kW and a 6.5-kW pump. In the WWTP installations, process water and gravity thickened sludge in the TS-range of 3-8 percent were tested with a 15-kW pump.

The characteristics of the different fluids were studied with a rheometer (Anton Paar) in a standard configuration, using the small gap approximation. This approximation results in a possibility related to the rheometer measurements to shear stress and shear rate, for laminar flow defined by Reynolds number [2]. However, rheometer measurements in a wide-gap configuration were used for sludge tests. This was due to the fact that sludge causes jamming problems in the standard configuration. This wide-gap configuration introduces difficulties in the rheological analysis, but one approach was used and further presented in [4].

The portable flow loop consisted of a closed flow loop of acrylic pipes (diameter 100-mm) and a container of 170 liters, a dry-mounted centrifugal pump and a gate valve to vary the flow rate. The pump pressure head was measured with static pressure taps upstream and downstream of the pump using a transducer by Rosemount 3051-CD3 (2.5-bar) (total performance ± 0.15 percent of full range).

The return leg (approximately 3-m) of the loop was used to measure the pipe pressure loss using a Rosemount 3051-CD2 (0.25-bar). The return length/diameter ratio indicated that a fully developed pipe flow was not expected for flow rates < 60-l/s. However, the pressure loss measurement was still interesting, as relative measurements compared to water measurements.

A magnetic inductive flow-meter (Danfoss, Magflow 3100) was mounted 10 pipe diameters downstream of the pump outlet. The electric input power was measured with a power meter (Hioki/Yokogawa) in order to obtain the pump performance curve.

It is well known that viscosity changes with temperature and that re-pumping results in an increased fluid temperature. In order to keep the test under constant flow conditions, the temperature was recorded with a thermocouple device and the fluid was kept at ambient room temperature during the test by cooling devices.

In the on-site WWTP tests, each pump with a swept-back leading edge impeller and relief groove in the volute (15-kW) installation was unique, but whenever possible the pressure taps were positioned to record a representative pressure. The pressure loss static taps were positioned in a low disturbance pipe region where fully developed pipe flow was expected. In order to record the power consumption, it was necessary to bypass the variable frequency drive (VFD). The different WWTP are shown in Table 1 with the type of sludge and total solids (TS) content and specific test carried out.   

[[{"type":"media","view_mode":"media_large","fid":"296","attributes":{"alt":"The non-Newtonian model fluid used in this project was Carbopol 676, polycrylat mixed in water.","class":"media-image","id":"1","style":"float: left;","typeof":"foaf:Image"}}]]

The non-Newtonian model fluid used in this project was Carbopol 676, polyacrylat mixed in water. Carbopol 676 is used in cosmetics, dishwashing and sanitary liquids and shows a shear-thinning behavior (http://www.carbopol.com/). A mass concentration of 0.6 percent was used and by adding sodium hydroxide (NaOH) the apparent viscosity changed due to the pH-dependency of the Carbopol 676 aqueous mix.

The wastewater treatment sludge of TS = 3 to 8 percent content was in focus during this study, as shown in Table 1. The sludge was received from the gravity thickener before the digester or the decanter in the plants. In the flow loop set up, the sludge was re-pumped, whereas in the on-site tests the sludge was virgin, i.e. only pumped ones through the test section.

In one of the WWTP, sludge polymer was added. We used FLOPAM 4240 of a concentration 2.0-g to 4.4-g polymer / kg TS-value in order to study the impact on the pump performance. In the WWTP this polymer was used to increase the decanter efficiency. The presence and possibility to use this polymer was of interest in this current study.   

Results

The rheometer test of the different Carbopol suspension shows a clear shear thinning behavior as expected.  The post-processing results in a determination of the non-linear model parameters, given by a power law model. The parameter result for different Carbopol is shown in Table 2. Carbo-3 shows the parameters, K = 20.0 Pa sn , n = 0.32 and roughly a TS-value of 10 percent, using the estimation shown by Frost [5].

[[{"type":"media","view_mode":"media_large","fid":"297","attributes":{"alt":"The rheological Carbopol results were used with a semi-empirical expression to estimate the pressure loss.","class":"media-image","id":"1","style":"float: left;","typeof":"foaf:Image"}}]]

The rheological Carbopol results were used with a semi-empirical expression to estimate the pressure loss [6, 7]. This semi-empirical expression is based on a friction factor and a modified Reynolds number [6, 7].

The agreement between the pipe loss measurement for Carbopol and the estimated results was good and somewhat expected. However, it validated the approach and the pipe measurements. This approach was applied for the sludge tests, but unfortunately a difficulty in the post-processing was not completely solved. However, the sludge pipe loss tests indicated that the re-pumping of sludge showed a minor impact on the rheology compared to pumping virgin sludge in the WWTP test.

In Figure 2, the pipe pressure loss in the flow loop shows a significant increase with flow rate, as expected. A shift due to an increased viscosity is obvious for Carbopol compared to the water pressure loss curve and to the theoretical curve, obtained from [6].

[[{"type":"media","view_mode":"media_large","fid":"298","attributes":{"alt":"For lower flow rates, all tested fluids showed a non-linear behavior and a significantly larger pressure loss compared to water curve.","class":"media-image","id":"1","style":"float: left;","typeof":"foaf:Image"}}]]

For lower flow rates, all tested fluids showed a non-linear behavior and a significantly larger pressure loss compared to water curve. For higher flow rates, the results show, for Carbopol as well as for the WWTP-Ore sludge, a parabolic curve shape. The WWTP-Ore had a belt drainage step and used polymers within the sludge handling process.

For some of the tested sludge, a recovery in the pressure loss was indicated. A recovery means a pressure loss less than water for the actual flow rate. Such pressure recovery is reported in results of pulp fiber suspension, see [8]. For all tests in the flow loop, a local loss increase is observed at maximum flow rate, even for water measurements. The reason for this local increase was the valve introduction in the flow resulting in an undesired impact on the results. This was verified by replacing the gate valve with a dummy piece of pipe.

In Figure 3, the pressure loss measurements in the WWTP installations, the non-Newtonian behavior of the sludge, is shown compared to the water reference curve. The tests of added polymer in the sludge showed a change in the characteristics towards Newtonian behavior, not shown in the figure.

In the case of process water, the difficulties in performing on-site measurement are exemplified. In process water all undesired disturbances are developed, whereas in sludge the disturbances are damped due to the viscosity. However, the agreement of the measurements and to the theoretical water curve was on an acceptable level in this study.

[[{"type":"media","view_mode":"media_large","fid":"299","attributes":{"alt":"In the case of process water, the difficulties in performing on-site measurement are exemplifiied.","class":"media-image","id":"1","style":"float: left;","typeof":"foaf:Image"}}]]

In Figure 4, the pump performance curves show a marginal change for different fluids in the flow loop. The solid line represents the water reference curve, second order least-squared curve fit to water data. The result shows a derating in pump performance over the flow range, apart from close to zero flow rate (shut-off).

In this region, the sludge results show an increase in head compared to the water reference curve. This is not completely understood, but a vague explanation is the sludge rheology. The pump performance results in sludge fluids (TS of 5 to 6 percent) showing a decrease up to 6 percent in pressure head at maximum flow rate. For Carbo-3, this is up to 14 percent head decrease. The derating may simply be estimated to be equivalent to the TS-value and simply to be assumed as constant for the complete flow rate range. Thus, for example, pumping a sludge of 5 percent TS-content reduces the head to 0.95 x Head in water.

The input power shows a progressive increase with increasing viscosity. In Figure 5, in the test of the sludge (TS of 5 to 6 percent) showed a power increase at maximum flow rate of 7 percent and of 15 percent for Carbo-3, respectively.

This result and the assumption of a constant rating, similar to the one for the pressure head, give the derating. The increase in the power consumption is equivalent to the TS-value, i.e. pumping a sludge of 5 percent TS-content, requires a power of 1.05 x Power in water.

[[{"type":"media","view_mode":"media_large","fid":"300","attributes":{"alt":"In this region, the sludge results show an increase in head compared to the water reference curve.","class":"media-image","id":"1","style":"float: left;","typeof":"foaf:Image"}}]]

In Figure 6, the power consumption of a WWTP field test shows similar results for sludge of different TS-contents as for the flow loop result. The same result for pump head in field tests was obtained as the one reported above, although the tests were more challenging. The challenge is keeping the flow condition constant during the measurements.

Discussion and Conclusion

As a result of the test described above, the derating of the pump performance, as pressure head and power consumption, was found to be as simple as:

(2)  for  TS< 8 percent

where TS is the percentage total solids, H is pump head and P is input power. The recommendations in Equation (2) show agreement with results in the recent study [9] for non-Newtonian slurries.

The derating recommendation for pumps with a swept-back leading edge impeller and relief groove in the volute in sewage sludge application is given a restriction for sludge of TS-value < 8 percent. Sludge with larger TS-values has to be further studied in order to improve the accuracy of the suggested recommendation. Further studies are needed to completely understand and close the rheological issues of sewage sludge.

[[{"type":"media","view_mode":"media_large","fid":"302","attributes":{"alt":"As a result of the test described above, the derating of the pump performance, as pressure head and power consumption, was found to be as simple as","class":"media-image","id":"1","style":"float: left;","typeof":"foaf:Image"}}]]

In the case of Newtonian fluids, the Hydraulic Institute (HI) defines a method of derating based on the known viscosity [10]. According to this method, pump performance with viscous fluids shows that the pump head versus flow rate curve falls below the water reference curve, while the shut-off head point remains, regardless of the fluid viscosity.

This empirical method is an approximation, and HI stresses that pump geometry is simplified and flow conditions are idealized. In the case of pumps with a swept-back leading edge impeller and relief groove in the volute, the HI method results in a somewhat conservative derating in the Newtonian fluid case.

Acknowledgements

It is generally perceived that pump technology evolves gradually in the continuous improvements of products. The improvements are driven by Life Cycle Cost (LCC), in particular by the energy cost awareness.

Still, experience shows that innovations in product development could have a game-changing impact. In pumping complex fluids, the influence of the fluid characteristics on the pump performance must be considered. This involves both pump design knowledge and system know-how, and it is challenging.

For example, several years ago a pump with a swept-back leading edge impeller and relief groove in the volute was developed to improve the clogging resistance in sewage applications in combination with an efficiency level similar to state-of-the art clean water centrifugal pumps. An extensive development program resulted in a twin-blade semi-open impeller.

A centrifugal pump with this type of impeller reaches 80 percent pump efficiency in sewage applications. The wet end volute is traditionally designed, but with a relief groove in the insert ring. The stationary relief groove and the design of the impeller leading-edge is the reason for the excellent clogging resistance in sewage applications.

In this application, the clogging effect refers to particles in the wastewater, such as rags or flocs, causing unscheduled blockage in the pumps. The wear resistance is a core issue in pump technology. For this pump, accelerated laboratory wear tests as well as wear field tests showed promising results. The relative decrease in efficiency compared to conventional single-vane pumps was reduced by approximately 50 percent [1].

Several case studies conclude that a pump with a swept-back leading edge impeller and relief groove in the volute significantly reduces LCC in sewage applications, where displacement pumps (PC-pumps) traditionally have been used.

References

  1. Arbeus, U. 1998. The N-Pump - a new concept, Scientific Impeller 1998,pp.23-27 ITT Flygt.
  2. Chhabra, R. & Richardson, J.F. 1999. Non-Newtonian flow in the process industries. Oxford: Butterworth-Heinemann.
  3. Seyssieecq, I., Ferrasse, J-H., & Roche, N., 2003. State-of-art: rheological characterization of wastewater treatment sludge. Biochemical Engineering J., 16:41-56.
  4. Uby L., Hallgren G., Fahlgren M. Holm R. and Svensson T. 2006 General accuracy improvement for Searl and Couette rheometery: corrections for arbitrary gap size and end effects of non-Newtonian fluids,  AERC- 2006 conf. proceedings, Greece.
  5. Frost, R.C. How to design sewage sludge pumping systems, tr185, Tech. report, Water Research Centre, Medmenham, UK, 1983.
  6. Miller, D.S. 1990. Internal flow system flow. BHRA publ.
  7. Brown N.P. Heywood N.I 1991, Slurry Handling - design of solids - liquid systems, Elsevier applied Science, London.
  8. Fellers, C. & Norman, B. Papperteknik , KTH Stockholm, 1998. 
  9. Sery, G.A & Slatter P.T., 2002, Centrifugal pump derating for non-Newtonian slurries. 15th International Conference on Slurry Handling and Pipeline Transport Hydrotransport 15, Banff, June 2002: 679-692.
  10. SIS-ISO/TR 17766, 2006, Technical report SIS, Centrifugal handling viscous liquids-performance corrections.

Pumps & Systems, March 2007

Issue