Many positive displacement pumps deliver the same flow rate regardless of the static head.

Chemical manufacturing includes complex processes. In fact, chemical manufacturing processes are so intricate that, typically, several unit operations exist within an overall process. These may include cracking, distillation and evaporation, gas absorption, and scrubbing and solvent extraction. Within these operations, transferring— is the process of transporting fluid from one point to another—stands out because it is important to the whole manufacturing process. Fluid transfer is a jack-of-all-trades with responsibilities along the whole chain. Some examples are moving raw materials into storage tanks, raw materials into blending or mixing tanks, final formulations into holding tanks and finished products into intermediate bulk containers for delivery or two-gallon jugs for store shelves.

Because of transferring’s importance, facility operators should identify the best pumping technology for the job—one that is versatile, reliable and efficient. For many years, centrifugal pumps were the go-to technology. However, positive displacement (PD) pumps—specifically sliding vane and eccentric disc pumps—can be the right pump technology for many chemical transfer operations.

The Challenge

In a basic explanation, the volume of fluid sent from Source Tank A will increase in Destination Tank B (see Figure 1). As this operation occurs, the only variable in the hydraulic system is the static head, which will change as the level in Tank A decreases and the level in Tank B increases.

In many cases, when the tanks are large enough, the static head variation is assumed to be insignificant, and a centrifugal pump is sized for a specific performance point. In reality, a centrifugal pump operates in a range in the curve of its hydraulic performance. The size of this range is specific to each application and should be evaluated.

Figure 1. Typical transferring process

In Figure 1, the performance of an equivalent PD pump is the yellow line (QM), which represents what a PD pump must do to deliver the same volume in the same time as the centrifugal pump that is operating in a specific range. Also, PD pumps, particularly those with self-adjusting volumetric efficiency capabilities, such as eccentric disc or sliding vane pumps, will consistently deliver the same flow rate across all pressure variations, regardless of the pumping system’s static head. As the discharge pressure changes, PD pumps provide a consistent flow rate.

A centrifugal pump’s operating range becomes more critical when the fluid must be transferred from one source tank to several points or tanks within the plant. In this case, the operating range will be wider, and the delivery parameters will be different from tank to tank. Chemical manufacturers have traditionally chosen centrifugal pumps for transfer applications for the following reasons:

  • They are commonly the first choice for moving water-like fluids. PD pumps are usually considered when the fluid is viscous.
  • They are a well-known technology and familiar to most operators.
  • They are believed to have a lower initial cost than PD pumps. However, this is not necessarily the case.

In reality, PD pumps can quantifiably counteract these advantages. PD pumps are appropriate for fluids with high viscosity, but they can easily move other fluids, from liquefied gases and water-like liquids (sliding vane pumps) to medium and very viscous fluids (eccentric disc and sliding vane pumps). PD technologies have operated successfully in the chemical manufacturing industry for more than a century, and their initial costs can be similar when all the equipment, accessories and controllers are evaluated. In many cases, the total cost of ownership is lower over a PD pump’s operational lifespan.

Centrifugal pumps work best when they operate at their best efficiency point (BEP). Unfortunately, the BEP is rarely realized for an extended period during fluid transfer operations, resulting in flow rates that can fluctuate constantly. Many facility operators are willing to accept fluctuations in flow rate. However, consistent off-BEP operation can lead to potential problems in the equipment’s operation, the production process and how the chemical is formulated. Note that the system, not the pump, dictates the operating conditions in which the pump must work.

During the chemical process, the amount of fluid sent must adhere to specific guidelines and quantities that are sometimes only known by the chemical manufacturer. In these instances, a centrifugal pump will not provide constant flow unless it is controlled with proportional-integral- derivative loops, flow meters, recirculation lines and variable speed drives. These components complicate the pumping system and introduce electric and electronic components that may be required to operate in hazardous areas and require special ratings.

By comparison, the fluid delivery rate of a PD pump that features self-adjusted efficiency will be more consistent than a centrifugal pump. In chemical manufacturing, these PD pumps also provide a more reliable production rates in terms of chemical quality. Contrary to self-adjusted volumetric efficiency technologies, centrifugal pumps lose efficiency as their components (wear rings, internal casing tongue or the impeller casing clearance) wear. Also, when a centrifugal pump operates to the left of its BEP, radial loads increase because the pump generates pressure along its volute by reducing the fluid velocity. This type operation will increase shaft deflection at the seal faces, increasing seal wear and adversely affecting the pump’s life expectancy. Working to the left of the curve will also increase axial loads that can potentially overload the thrust bearings, especially in open-impeller and diffuser-type, multistage centrifugal pumps. Finally, as a centrifugal pump operates close to the zero-flow point (zero efficiency), the temperature may increase to levels that can be harmful to heat-sensitive chemicals or products, negatively affecting safety.

When a centrifugal pump operates to the right of its performance curve, other problems can develop. Specifically, the net positive suction head (NPSH) required increases, which may cause cavitation. Because the fluid transfer process in the chemical industry is managed in batches, insufficient NPSH may be more complicated to detect. However, it will deteriorate the pump’s operational capabilities continuously, meaning that the pump’s ability to handle cavitation will be compromised.

Other potential performance-robbing concerns for centrifugal pumps in chemical-transfer applications include:

  • Mechanical issues—Caused by vibration, which may reduce mechanical seal life, during off-BEP operation
  • Overheating—Caused by low-flow operation
  • Product leakage along the pump shaft—Because of shaft deflection (overhung impellers)
  • Dry running—Can cause pump problems but may result in a catastrophic failure for a magnetically driven pump because the pumped media lubricates it
  • Inability to strip lines
  • Inability to self-prime
  • Susceptibility to cavitation—Can occur because of entrained gases
  • Fluid-handling capabilities affected—Can be compromised with changes in fluid viscosity, which can occur because of modifications and adjustments in the process or temperature changes

Centrifugal pumps are acceptable options in many fluid transfer applications and have been proven to perform reliably for many years. However, a more efficient and reliable option may be available for many fluid transfer operations in chemical manufacturing.

Table 1. Real-world example of the total cost of operation for a PD pump versus two centrifugal pumps
Notes:
1. Pump types compared are PD sliding vane and ANSI centrifugal.
2. Yearly operating cost is based on 3,000 hours of operation at a rate of $0.1 per kilowatt hour, or $223 per pump horsepower.
3. These costs are based on pumping an aqueous solution with a specific gravity of 1 at 300 Saybolt seconds universal at 126 gallons per minute (gpm) at 80 psig for the PD pump and a range of 114 gpm at 190 feet to 130 gpm at 165 feet for the centrifugal pumps (see Figure 2).

The Solution

Unlike centrifugal pumps, the design of PD pumps allows them to produce a constant flow at a given speed regardless of discharge pressure, which is critical in chemical manufacturing, which require precise dosing rates. Specifically, two pump types are ideal for chemical fluid transfer applications: sliding vane and eccentric disc.

Sliding vane pumps feature a series of vanes in the pump rotor that slide out as they wear, which means that the pump delivers volumetric consistency throughout its life or until the vanes require replacement. Sliding vane pumps also offer zero shaft leakage (magnetic coupling); non-galling operation, stainless steel or ductile iron construction for corrosive liquids; chemical-duty mechanical seals; low-to-medium shear and agitation; and self-priming and dry-run capabilities, even in an explosive or hazardous environment.

This pump type features a disc inside a pump cylinder. The disc is driven by an eccentric bearing on the pump shaft. This creates four distinct pumping chambers that increase and decrease in volume as the disc is rotated by the eccentric bearing, producing suction and discharge pressures as the chambers move in pairs that are 180 degrees apart. This operation ensures that the fluid passes through the pump at a constant and regular flow rate and eliminates pulsation within the pumped fluid. Because the pump does not depend on clearances to facilitate product flow, any slip or loss in volumetric efficiency is negligible. Additionally, with the mechanical sealless option, products that are difficult to seal and prone to crystallization cannot adhere to any surfaces and cause damage, which eliminates a maintenance concern.

Figure 2. Combined operation

Because of the sliding vane or eccentric disc pump’s method of operation, they offer many benefits to fluid transfer:

  • Constant flow across the required range of pressures
  • Low-shear operation, important when handling many raw materials in chemical production
  • Dry-run capability and ability to strip discharge lines
  • Ability to work to some levels with compressible fluids

One drawback of rotary PD pumps is that they should not operate against a closed valve on the discharge side because they do not have shutoff head. This potential problem is overcome, however, with the placement of a relief/safety valve on the discharge side of the pump. Table 1 illustrates a real-world example of the total cost of operation for a PD pump versus two competitive centrifugal pumps. While the initial cost of a PD pump could be a few hundred dollars more than a centrifugal pump, the monetary savings through the first five years can be significant.

Because a PD pump relies on less horsepower to operate, its annual operating costs can be nearly 60 percent lower than that of the centrifugal pump. Because of this, the total cost savings that are realized when a PD pump is used increase throughout its life cycle. Evaluation of the range of operation of a centrifugal pump and the equivalent PD pump is shown in Figure 2.

Equation 1
Where:
SA and SB = the areas of source Tank A and destination Tank B, respectively.
Volume VA = Volume VB is the amount of fluid transferred.
H1, H2, R = the points where the curves intercept the “H” Axis (Q = 0, or zero flow), which are known for a given system
H1 and H2 = static heads geometrically defined
R = a characteristic of the pump curve
HA and HB = the final condition at a given variation in time of h
tf = the time that the PD pump operates to deliver the same amount of fluid as the centrifugal pump while operating in a range

To compare the pumping technologies, the range of operation of the centrifugal pump and the equivalent PD pump must be fully defined (see Figures 1 and 2 and Equation 1). H and Q are the total dynamic head and the flow. When the curves of the system and the pump are defined, c, M, Q represent known constants of the quadratic equations. To simplify the mathematical analysis, the pump curve is expressed as a second-degree curve, although depending on the nature of the impeller, this can be another polynomial or logarithmic equation. Because the volumes (VA, VB) and their changes over time (H1, H2, HA, HB) are known, QM can be simply evaluated by dividing the volume pumped by tf.

Issue