Fluid Thinking
Brings Optimum Pump Solutions

In sizing the right rotary lobe pump (RLP) for any application it is essential to understand how the fluid will behave in the pump. At Alfa Laval Pumps Ltd’s Eastbourne manufacturing facility for its SSP Pumps brand of RLPs, the company has a rheology lab, which combined with a sophisticated pump selection and configuration program provides customers with the optimum solution. This in-house rheology lab service is the only one of its kind in Europe offered by a rotary lobe pump company. In this article Business Development Director, Marcel Verhoeven, discusses the importance of rheology and other fluid properties in pump sizing and the company’s procedures enabling customers to obtain the ideal solution and avoid costly errors.

For engineers specifying an RLP for a particular application it’s not just a matter of matching their required duty flow rate and discharge conditions to the pump’s specifications. A crucial consideration is the nature of the pumped medium, particularly an understanding of how it will behave in the pump as well as in the connecting pipe work and associated equipment.

The science of fluid flow is called rheology and it’s the rheological property of viscosity that is one of the most important factors to understand in sizing a pump. For some end users not appreciating this, has led to expensive errors in sizing a pump and/or its drive unit. Other equally important factors to consider include whether the fluid is corrosive and/or abrasive and its compatibility with the pump’s materials of construction.

Ensuring the best solution
Alfa Laval Pumps Ltd has developed a set of analytical procedures, which combined with an innovative pump selection program ensures the correct pump is sized first time and every time. If the characteristics of the pumped media are unknown by the customer, the company offers a service to analyse the pumped media for which a 0.5litre sample is required. From the sample supplied the rheology lab is able to carry out a number of tests to determine its viscosity and rheology. Other tests can also be undertaken when required such as specific gravity, pH, materials compatibility and abrasive solids content.

Abrasive solids carried in suspension are of particular importance as these may cause premature pumphead wear if the solids size, shape, hardness and density are not taken into consideration. For increased wear resistance to the pumphead and rotors, methods such as hard coating or other surface hardening treatments are employed. Also non-metallic rotors may be supplied which can give long-term wear resistance.

Understanding viscosity and avoiding vicious consequences
The viscosity of a fluid can be regarded as a measure of how resistive the fluid is to flow. The ease with which a fluid pours is an indication of its viscosity. For example, cold oil has a high viscosity and pours very slowly, whereas water has a relatively low viscosity and pours quite readily. High viscosity fluids require greater shearing forces than low viscosity fluids at a given shear rate. The cgs unit for measuring viscosity is the centipoise (cP), water at ambient temperature and atmospheric pressure has a value of 1cP and a very viscous substance such as printing ink typically has a value of 50,000 cP. Viscosity of the fluid in an RLP under pumping conditions will determine its efficiency, generally as viscosity increases so will the pump’s efficiency. This is due to a decrease in the amount of slip, which is the leakage of fluid through the pumphead clearances against the desired direction of flow. The amount of slip increases as the fluid viscosity decreases thereby reducing pumping efficiency.

In some fluids the viscosity remains constant regardless of applied shear rates. These are known as Newtonian fluids, where at constant temperature the viscosity is constant with change in shear rate or agitation (Fig. 1). Mineral oil and water are typical Newtonian fluids. Newtonian fluids will have a viscosity ‘at rest’ in the sample bottle similar to that in the pump.

However, other fluids can provide an unexpected surprise. Their viscosity ‘at rest’ may be very different to that which applies under pumping conditions. These fluids, known as Non-Newtonian fluids, show a change in viscosity with shear rate and are classified into various types depending on their characteristics. Some of these are briefly described below.

For pseudoplastic fluids, the viscosity decreases as shear rate increases (Fig. 2); examples include china clay slurries, adhesives and crude oil. Sometimes the initial viscosity may be so high that it prevents flow from starting in normal pumping conditions. Another very important implication for pumping a fluid of this type is that its viscosity in the pump may drop to just a fraction of its original ‘at rest’ viscosity, which if not recognised at the pump sizing stage may result in poor pump efficiency and failure to reach the pump duty requirements.

In the case of plastic fluids they need a certain applied force to be able to flow like a fluid. Non-drip paint is an everyday example that requires agitation to overcome its solid-like structure thereby allowing it to flow. The considerations for plastic fluids in pumping are much like pseudoplastic fluids.

Fluids showing a decrease in viscosity with time under shear are termed thixotropic fluids. A fluid may show both thixotropic and pseudoplastic behaviour. Fluids displaying the opposite rheological property of increasing viscosity with time under shear stress, such as vanadium pentoxide sol, are known as anti-thixotropic fluids.

Dilatant fluids such as paper coatings exhibit an increase in viscosity with shear rate. Again, their behaviour in the pump needs to be understood before any pump sizing can begin.

Measuring success
Some fluids can change permanently when subjected to shear. Such delicate pumping media needs to be handled carefully. For example, the fluid structure of rubber latex can be irreversibly destroyed when under excessively high shear conditions, rendering it useless for further processing.

To measure a fluid’s viscosity under pumping conditions samples are tested at Alfa Laval’s rheology lab with a computerised rheometer. Fig. 3 shows a graph of viscometer results for glycerine which is a Newtonian fluid showing no change of its viscosity with shear rate. The results of an analysis of detergent in Fig. 4 show that it is a Non-Newtonian fluid with viscosity dropping sharply from about 2,150 cP ‘at rest’ to about 1,000 cP in a shear environment typically found in pipes, and 90 cP under shear conditions typical of a mid size RLP. If the viscosity of the detergent was taken ‘at rest’ to predict the pump’s efficiency this would give a significantly erroneous optimistic result compared with the actual pump’s efficiency, which would be about half of the expected efficiency under certain operating conditions.

Sizing a pump and drive combination on the assumption that a fluid’s ‘at rest’ viscosity is the same as that under pumping conditions can result in wrong and costly decisions being taken. For example a water company sized 30 pumps with electric motor drives to deliver chemicals for water treatment. The company was perplexed to discover that the pumps were only performing at fifty percent of their expected capacity. It was subsequently discovered by rheological examination that the fluid’s ‘at rest’ viscosity of 2,000 cP drastically plummeted to 6 cP in the pump, corresponding to a reduction in pump efficiency due to slip from 97% down to 50%. This meant that to achieve the required pump capacity the pumps’ speed needed to be increased, necessitating in the replacement of existing drive units.

A further example can be quoted from a mining company having problems with pump seizures when pumping talc slurry. Upon consultation with the rheology lab at Alfa Laval Pumps Ltd it was discovered by testing that the fluid was strongly dilatant and that under shear conditions the fluid could solidify in the pump, thereby becoming unpumpable.

Temperature can also greatly affect viscosity. In the case of an insecticide manufacturer, one morning it was discovered that six motors fitted to pumps handling insecticide had mysteriously burnt out. On subsequent investigation it was discovered that the fluid’s high daytime temperature of 130°C dropped to 20°C at night, corresponding with a dramatic increase in viscosity from 30cP to about one million cP. On start-up in the morning, the power required to move the still cold, highly viscous fluid, caused the starting load rating of the motors to be exceeded. Examination of the fluid’s rheology would have avoided the costly down time involved in rectifying this problem.

Once a set of sample viscometer results is obtained a best-fit match to a mathematical rheological model from a range of models is made. The selected mathematical model, for example the Cross model, generally used for pseudoplastic fluids, is used to calculate viscosities in pipes and pumps for any particular customer application. This is employed in the SSP Pumps LobeSelect selection and configuration program.

Other tests conducted in the rheology lab on the customer’s sample fluids include compatibility tests with the pump’s materials of construction such as elastomeric seals to check for absorption, and possible disintegration. Additional information from these tests is then used to specify the pump.

LobeSelect – making pump selection easier
To select the optimum SSP rotary lobe pump, LobeSelect, a sophisticated Windows based selection and configuration program has been designed. This prompts the user to enter pump duty information, and using data from either the existing comprehensive fluids database, or new information on the customer’s fluid from the rheology lab, selects the pump from the product range most suited to the specific application. The database contains the crucial rheological and other properties of about 10,000 fluids built up by the rheology lab over the past three decades.

The first step in pump selection is the input of the required flow rate and the resulting pressure differential. LobeSelect then extracts from the fluid database the fluid’s ‘at rest’ and ‘in pump’ viscosity, specific gravity, and optimum running speed required to deliver the desired flow rate at a specified temperature or temperature range. From this, pumps matching the duty criteria of the customer’s application (see Fig. 5 – image of LobeSelect screen output) are then selected. For each pump selected, its efficiency, Net Positive Suction Head requirement (NPSHr), power required and running speed are obtained.

For abrasive fluids such as clay slurry, LobeSelect will also automatically specify pumps with hard faced mechanical seals. In the case of delicate media, for example latex, the maximum pump speed is set by the program’s fluid’s database.

In ensuring the optimum pump selection and configuration it is essential that the fluid’s characteristics, particularly its viscosity, be considered. For a definite Newtonian fluid or a fluid matching or closely matching the database predicting the viscosity in the pump is straightforward. However, for many applications it is better to play safe and analyse the fluid to determine its rheological properties. Even slight changes in composition of a fluid can have dramatic consequences. For example a paper manufacturer, which changed just its coating’s thickening agent, suffered a significant drop in pumping performance compared to the expected performance. Downtime caused by the removal of the existing pump unit and installation of a new more efficient pumping solution could have been avoided if a sample of the fluid was analysed and used prior to the pumps being sized, ultimately saving time and money.