Positive displacement pump

Fix the speed and you have fixed the flow. Every other property of a positive displacement pump sits downstream of that. The Hydraulic Institute separates the two families on exactly this basis: rotodynamic pumps "impart energy to the pumped fluid by means of a rotating impeller, propeller or rotor", while positive displacement pumps "trap and move a volume of fluid with each shaft rotation".

The US Department of Energy pumping sourcebook, written with the Hydraulic Institute, states the consequence plainly. Positive displacement pumps "have a fixed displacement volume", so "the flow rates they generate are directly proportional to their speed" and "the pressures they generate are determined by the system's resistance to this flow".

That last clause is the one that bites. Shut a valve on the discharge of a centrifugal pump and it churns the same liquid round the casing, heats it, and eventually cooks itself. Do the same to a positive displacement pump and the trapped volume has nowhere to go. The sourcebook calls the condition deadheading and gives the sequence: "system pressure will build until a relief valve lifts, a pipe or fitting ruptures, or the pump motor stalls."

So a relief valve on the discharge is not an accessory. Size it for the full flow of the pump, pipe the relief back somewhere sensible, and test it on a schedule. A relief valve nobody has lifted in ten years is a fitting, not a protection device. It is also why a positive displacement duty is not selected off a performance curve the way a centrifugal duty is. You pick the displacement and the speed, then check the motor has the power for the pressure the system will actually impose.

Two things push a duty across the boundary, and only one of them usually gets measured.

Solids come first. Somewhere around 3–4% is where we stop offering a centrifugal and start offering positive displacement. That is a rough figure and it should be read as one. It earns its place because it is the number a site engineer actually has, and because it is right often enough to start from. Below it, a solids-handling impeller will normally cope. Above it, a centrifugal spends more and more of its energy fighting the medium instead of moving it.

Viscosity drives the same decision and gets measured far less often. A centrifugal pump loses rate of flow, head and efficiency as the medium thickens, which is why the Hydraulic Institute publishes a whole guideline, ANSI/HI 9.6.7, for correcting a pump's water performance onto a viscous liquid. It carries separate correction factors for flow, head and efficiency, and a flowchart to establish whether the procedure applies at all. A positive displacement pump has no equivalent problem, because it pressurises the medium directly rather than accelerating it. The Department of Energy sourcebook puts positive displacement pumps ahead of centrifugal pumps for viscous duty on exactly that reasoning.

The question we ask on the phone is blunter than either measurement: is it water, or as water? If the answer is that it goes like treacle when the plant is cold, the duty is already positive displacement, and the viscosity that decides the selection is the worst case, not the typical one.

Now the part a page trying to sell you a pump would leave out. Positive displacement costs more to own. The same sourcebook is direct about it: centrifugal pumps "typically suffer less wear and require fewer part replacements than positive displacement pumps", and positive displacement pumps carry "higher maintenance requirements than other types". You also inherit pulsation, which can excite pipework and loosen mechanical joints, and a longer spares list, because rotors, stators, discs, diaphragms and valve seats are all consumable. If a centrifugal will genuinely do your duty, buy the centrifugal.

Positive displacement divides into rotary and reciprocating, and the line is drawn at the mechanism. Rotary types carry the medium in cavities formed by meshing components: screws, gears, lobes, vanes. Reciprocating types draw it in and push it out on a cyclic stroke, using pistons, plungers or diaphragms. The rotary side has its own standard, ANSI/HI 3.1-3.5, covering sliding vane, axial piston, flexible member, lobe, gear, circumferential piston and screw pumps.

Rotary lobe is the workhorse of hygienic transfer and of sludge. Alfa Laval builds the SRU around a "low-shear, low-pulsation pumping action" for creams, gels, emulsions, aerated mixtures and organic solids in suspension, rated up to 106 m³/h, 20 bar differential pressure and 200°C. At the other end of the trade, Vogelsang's VX series primes from pits, evacuates air out of empty wastewater pipes and moves "primary sludge, through excess and secondary sludge, to concentrated digested sludge or sewage sludge". Same operating principle, two entirely different builds. If you want the mechanism in detail, the rotary lobe working principle is set out separately.

Progressive cavity is what you reach for when the medium is thick, laden, or both. Roto's range states the behaviour cleanly: "the capacity is proportionate to the speed" while "the developed head is independent of the rotational speed", giving "a uniform, metered and non-pulsating flow" with "minimum degradation of shear-sensitive media" from the low internal velocity. It is self-priming, reversible, and Roto quotes suction lift to 9.5 mWc. The price of all that is the stator, which is a wearing part with a maintenance regime of its own.

Peristaltic and diaphragm pumps buy you something the others cannot. The Department of Energy sourcebook notes that these types "do not require seals and thus do not leak", and that on corrosive or hazardous media, taking seal maintenance out of the system is where the money is saved. Nothing but the hose or the diaphragm touches the medium.

Double disc deserves naming because of what it does on a sewage works. Penn Valley rates its double disc pump to run dry without damage, with a "seal-less design, no packing gland or mechanical seals, no seal water required", self-priming on high suction lifts, passing line-size semi-solids, and routinely put on primary sludge, septage, scum, surplus activated sludge, digested sludge and lime slurry. Rag, grit and dry running are what kill pumps on a sewage works, and those are precisely the three it shrugs off. The Alfa Laval and SSP double disc came out of production in 2017. The Penn Valley pump replaces it directly, and in our experience it outlasts the pump it succeeds.

Reciprocating covers metering and high-pressure low-flow work. Because each stroke displaces a known volume, counting strokes is the same as counting litres, which is why chemical dosing is a reciprocating job rather than a centrifugal one.

Abrasive media and shear-sensitive media both want pumping slowly, and nothing on a performance curve tells you so.

The mechanism differs at each end. With an abrasive medium (grit, ash, lime slurry) wear tracks how fast the wetted surfaces slide past one another, so taking the speed down buys a long extension on the wearing parts. With a shear-sensitive medium the damage is to the product rather than the pump. Shear breaks emulsions, thins yoghurt, shortens fibre and knocks the structure out of anything held in suspension. Both Roto and Alfa Laval design against it: Roto attributes its low degradation of shear-sensitive media to low internal velocity, and the Alfa Laval SRU is built around low shear and low pulsation for exactly that reason.

The commercial consequence is the one customers resist. You buy a bigger pump than the flow alone suggests, and you turn it down. That costs more on the order and less over twenty years of running. It is also the opposite of how most duties get specified, because the margin usually gets added to flow and pressure rather than taken out of speed. Running a positive displacement pump hard in an abrasive medium is one of the quickest ways we see a pump's working life cut short.

Three failure modes, and the order barely changes from site to site.

Dry running comes first, and the families do not tolerate it equally. On a progressive cavity pump it is catastrophic: the rotor burns through the elastomer stator, and you replace both. On a rotary lobe pump the usual casualty is the mechanical seal, which is lubricated by the medium it is sealing. Take the medium away and the faces run dry and fail in minutes. In most pumps it is the seal that gives out first, whatever else is wrong, and where a seal choice is genuinely difficult the seal manufacturer's advice is the one to take. Saying so is practice, not an admission.

Deadheading is second, and usually not a design fault at all. Somebody shuts a valve, forgets it is shut, and starts the pump. The relief valve on the discharge exists for those five seconds, which is the whole argument for testing it.

Grit is third, and it is a slow failure that presents as a sudden one. Flow falls off as clearances open, somebody winds the speed up to compensate, and the wear rate goes up with it. On a wastewater works the root cause is almost always grit removal upstream not doing its job, which is a plant problem the pump then gets blamed for.

The exception is worth knowing, because it is the whole reason double disc pumps still sell. A pump with no mechanical seals cannot fail one, it will sit through a dry start, and a pump that passes solids does not care much about rag. That is a different answer to the problem, not a tougher version of the same pump.

Send the medium first. On a positive displacement duty the medium decides more than the duty point does, and it is the part that usually arrives last.

• The medium, named, with its solids content and what those solids are.
• Viscosity at its worst, not its typical: cold start, end of batch, whichever is thicker. Say whether it is Newtonian or thins as you work it.
• Flow, in l/s or m³/h, and whether it is continuous or batched.
• Suction and discharge conditions, so the total dynamic head can be worked out rather than guessed. Suction is where positive displacement duties go wrong.
• Temperature, which with the medium sets the sealing arrangement and the elastomers.
• Hygienic or not, which sets the wetted materials and whether the pump has to clean in place.
• Product sensitivity: shear, abrasion, and whether the solids have to come through intact.

Most enquiries do not arrive with all of that, and they are not expected to. A site engineer has to be expert in a great many things and pumps are rarely one of them, so the job is asking the right questions rather than expecting the right brief. The pump selection sequence sets out the order we actually work in.

The duties this comes up on most are high-viscosity transfer, low-shear handling, food and hygienic processing and water and wastewater. Tell us the duty and we will size it, including when the answer is that a centrifugal will do the job for less money.