Resolving Operational Problems in Pumping Non-Settling Slurries
AFT Fathom™ Technical Paper
Authors: Daniel W. Wood, DuPont and Trey Walters, Applied Flow Technology
Presented at the Twenty-Eighth International Pump Users Symposium, September 24-27, 2012
ABSTRACT
A case history is presented pertaining to five pumping systems that operated satisfactorily until a new production requirement was imposed on the pumping systems. A new slurry product initially developed at lab scale was introduced into the production plant for an initial trial run. Problems began to surface immediately on three out of five batch process pumping systems when the slurry could not be pumped through the plant at the contract rate. Additionally, significant “heels” (unwanted fluid levels) were left in some of the suction vessels that were unable to be pumped out, resulting in considerable yield losses. This manufacturing problem had not been anticipated by the team, and without a quick resolution, a loss of customer confidence and a significant delay in the new product would have resulted.
The authors present an improved method for analyzing fitting losses in pumping systems when dealing with non-settling slurries operating in the laminar regime. In addition, design considerations are presented to minimize the impact that piping has on the pumping system when handling non-settling slurries operating in the laminar regime. A commercially available software package (AFT Fathom) was used to model the systems to better understand the hydraulics.
Conclusions & Recommendations
For a non-settling slurry operating in or near its laminar regime, the method proposed in this paper for estimating the pressure drop through fittings provided excellent results in the case history described.
- The NPSH available can be significantly impacted by pipe fittings in the suction line. Even though some of these fittings may be using for clean out (e.g., branch flow tees), their use should be balanced against providing adequate NPSH for the pump.
- Vessel nozzle outlet shapes are critical to the preservation of NPSH margins in non-settling slurry pumping systems operating in or near the laminar regime. Sharp edge outlets should be avoided if possible, and the use of rounded edge
outlets should be employed. - As a guideline, the NPSHA should exceed the NPSHR by a minimum of 5 ft (1.5 m), or be equal to 1.35 times the NPSHR, whichever is greater. For example, for an NPSHR of 10 ft (3.0 m), the NPSHA should be a minimum of 15 ft (4.5 m)
- Downstream fittings can also cause excessive pressure drop and flow reduction. It is important to minimize fittings where possible in the piping downstream of the pump.
- The shear added by a centrifugal pump to the liquid is significant. It is estimated per the Metzner-Otto rule that a centrifugal pump shears the fluid at 11 times the rpm of the pump. Some fluids do not quickly revert back to their presheared state after undergoing such a high level of shear,
and this may cause pressure drop calculations through
fittings to be overly conservative downstream of the pump. - Care should be taken when using hydraulic analysis
software to ensure that the fittings are being analyzed
properly from a pressure drop standpoint.
INTRODUCTION
Description of Pumping Systems
Five batch pumping systems were in place in an existing process plant. In each pumping system, the process liquid is fed into the pump from a large suction vessel which contains a mixer to keep the liquid in a sheared state. The pressure in the vapor space of the suction vessel is atmospheric. The piping from the suction vessel to the pumps is not a straight path in the systems, with some being more complicated than others. The liquid exits the pumps and goes through discharge piping.
There is a minimum flow recirculation line in the discharge of each pump that is regulated by a pinch valve. Pumping systems #1, #3, #4, and #5 are transfer systems moving fluid from one tank to another and a representation is shown in Figure 1 in the paper. Pumping system #2 moves the fluid to a machine which interacts with the fluid and this machine requires a minimum inlet pressure and a representation is shown in Figure 2 in the paper.
Newtonian fluids cover conventional fluids such as water, where the fluid shear stress is directly proportional to shear rate. The proportionality constant is the viscosity of the fluid. This relationship is observed in the solid line of Figure 3 at the left. Here it is apparent that the shear stress varies directly with shear rate. The solid line is for a Newtonian fluid.
The viscosity of a Newtonian fluid is not a function of the fluid dynamics (e.g., velocity, which is directly proportional to shear rate) or a function of time. This can be observed on the right of Figure 3 in the paper, where the viscosity has no dependence on shear rate (for the solid line, which is Newtonian).
Steady shear rheological behavior, shown with shear stress and viscosity as a function of shear rate. Dotted line is shear thinning fluid; solid line is Newtonian fluid. A fluid which exhibits a viscosity dependence on the fluid dynamics (e.g., velocity/shear rate) or time is referred to as non-Newtonian. A fluid in which the shear stress and shear rate follow a straight-line on a log-log plot is referred to as a power law fluid.
In practice, this means the viscosity varies for different velocities. In the case of a power law fluid, as the velocity increases, the viscosity decreases. This is also known as “shear thinning” behavior. In this case, the pumped media is a non-settling slurry. The slurry has the characteristic of being shear thinning and acting like a power law fluid. The fluid in this case is not drilling mud, but it looks and behaves somewhat like many drilling muds. The process followed to calculate the pressure drop for a power law fluid (and most other non-Newtonian fluids) is to first perform a rheological test on the fluid.
A viscometer is used to measure the shear stress at different shear rates. Often the test is done with increasing shear rate and then decreasing shear rate to check for hysteresis. A true power law fluid will not exhibit any significant hysteresis. This data is then used to determine the power law constants. The viscometer test was done on the fluid in this case study.
The process of determining power law constants was pursued and the raw rheological data followed a power law model quite well. This confirmed the fluid was non-Newtonian in its behavior and that the slurry was non-settling. If the slurry was of a settling nature the power law model would not have fit the data. Later in this paper the mathematical details of how these constants are used to calculate pressure drop will be discussed. Once the power law data was applied to the systems in question, it was apparent that the Reynolds number was in the laminar regime.