The fine art of sludge pumping

Les Gornall explores the redesign and commissioning issues behind sludge pumping systems

Dr Sludge asks: "Why do sludge pumping systems need redesigning so often at first commissioning?"

Pumping water - it's intuitive

This may seem obvious but if a water pump's speed increases, the pressure in the outlet pipe (or resistance to motion) increases along with an increased flow rate. This is how Newton characterised fluids like water. Put another way, a Newtonian fluid is one in which the head loss in the delivery pipe above (pressure P1 minus pressure P2) is proportional to the velocity of the fluid in the pipe. This always assumes modest velocities and non-turbulent flow conditions in the pipe. Conversely, a decrease in the pump output causes a decrease in flow rate and a decrease in back pressure in the delivery pipe.

Pumping Sludge - it's an art

Water is a liquid - defined as a substance that will take on the shape of its container. Water resides in the bottom of a drinking glass and flows into vacant spaces in the inlet of a pump with little resistance. Unfortunately this is not true of sludges. A thick sludge, particularly if it is derived from faecal material and organic fibres, can be formed into 3D shapes independent of its container. This is perhaps best demonstrated in three 'paintings' by Chris Ofili in the Tate Modern that use elephant droppings for 3D effect. When applied, the dung was not a liquid (and fortunately did not run off the canvas, not least because Chris's work sells for more than $80,000 a piece).

Even thin sludges are non-Newtonian, pressure losses in pipes are not proportional to velocity and the viscosity of sludge is variable. Technically, sludge is best described as a Bingham plastic with a straight-line relationship between shear stress and flow - after flow has begun. The Bingham plastic equations are complex and may be useful for calculating the pressure losses in long pipes and interesting to mathematicians. However, for most practical work, simple experiment-based relationships plus some building tricks can be used to predict the behaviour of pumped systems and make them work for sludges. The first useful relationship is the '600 rule' that predicts trouble in pumping operations, noted in Metcalf and Eddie (P788):

The 600 Rule

When the product of the total solids percentage and the volatile solids percentage exceeds a value of 600 difficulties will be experienced in pumping.

Some theoretical examples are:

  • an old septic sludge, well degraded, with a VS of 50% and TS of 4%. The product is 50 x 4 = 200 and this is well below 600 and easily pumpable,
  • a primary sludge with a VS of 80% and TS of 4% would give a product of 320 - still below 600 and relatively easy to pump if the velocities in the pipe are kept up.
  • However, a thickened SAS (or primary sludge removed too infrequently) with a VS of 80% and TS of 8% would give a product of 640 and this would be difficult to pump reliably under the empirical 'rule'. The reason is that this sludge would not be a liquid.

    The pump moves sludge out of the inlet and transfers it to the outlet, a space is left in the inlet. However, in the example above, this will most likely be starved of fresh sludge since the semi-solid material may not 'flow' into the inlet as a liquid would do. In order for the inlet flow to occur, the velocity of the inlet sludge into the pump has to increase to a critical level, this velocity and the shearing action in the inlet pipe 'liquifys' the sludge much as 'gel' ceiling paints liquify when stirred.

    The difficulty is how to start the process. But there is another effect in the outlet. When pumping viscous sludges, special arrangements have to be made at pump suctions to accommodate the Bingham plasticity - chief amongst these is to make the pump suction line from a tank of sludge as short as practically possible for reliable operation. In this case above 1m would have been more appropriate. Sludge cakes of 15-30% TS can be pumped but these are treated as solids, sheared into crumbs that are dropped into a hopper and pushed into the pump inlet using a helical drive screw.

    Here, the vertical inlet pipe is 1m x 0.5m in section. A useful matrix (derived from a number of experiments on sludge head losses) is provided in the table above. The table is used for estimating the factor by which a 'head loss' based on water has to be multiplied for a range of sludges - excluding any thixotropy in the system. For example, a pump is designed to push water at 2.2m/s (7ft/s) though a given pipe is used to pump. On a 10% TS sludge it would have only 1/40th of the power required for start up on its performance curve and about 1/6th of the power required for 2.2m/s sludge flow once it got the mass of fluid moving.

    When designing pumps for sludges, use design flows of 2-3m/s and use the above factors to size a mighty motor/gearbox to provide the required power through start-up without stalling - expect the worst case total solids because when the pump receives an unexpected slug of high solids sludge, even for a few seconds, it will most likely stop the pump unless there is provision for the 'out of spec sludge' design factors. It also helps on variable speed pumps to ramp up to full power very quickly.

    redesign issues

    Why do sludge pumping systems need redesigning so often at first commissioning? In these more competitive times when construction bids are won on the last 1% of contract price, pumps are mistakenly sized for the average sludge conditions as expressed in the contract, for example, 'the sludge will be thickened to 7%'. This makes the pump undersized relative to the maximum total solids in the sludge found in field conditions. In practice, the TS will vary above and below the design average value by a few percentage points.

    The pump may operate well at the design values until the TS increases for external reasons (a high polymer setting on the drum thickeners or at the start of a fresh and more active batch of polymer for instance). Once the '600 rule' limit has been broken a slug of high TS sludge moves through the pump inlet pipe, the pump inlet struggles to provide a sludge flow to the pump body and the discharge pipe velocity decays to below 1m/s.

    The reduction of sludge flow into the pump then causes the sludge in the discharge pipe to thicken - the viscosity increases, the back pressure on the pump increases, which further slows the flow. The slower pump rate causes a yet further reduction in velocity, which further increases the resistance to flow, which puts yet more load on the pump - a strong feedback that progresses quickly until the pump stalls and flow stops. In design tenders, the upper and lower limits for both total tolids and volatile solids need to be specified for all sludges in order to design the pumps.

    Pumping sludge is counter-intuitive

    Water, and any other Newtonian fluid, responds with a lower head loss for a reduced flow rate and less work is required from the pump as the flow slows down (this is intuitive). However, when sludge is pumped at velocities in the range 0.1-3m/s, a reducing flow can produce an increasing head loss, more work is required from the pump as the flow is slowed down. This is contrary to common sense - its non-Newtonian, it's not water!

    Dr Sludge's casebook:

    At site A duty/standby 3kW sludge transfer pumps failed to push blended 8% sludge 30m through a 150mm delivery pipe even when operated together in hand. The pump inlet pipe comprised 10m of 150mm bore pipe fitted into the base of a 4m high filled tank of 8% sludge with a VS of 80%.

    There was some polymer in the sludge from a cake/sludge reblending operation. In an experiment, a flange was removed to leave the inlet pipe open to the air close to the pump. The flow was monitored. Actually, to everyone's surprise the visibly free running sludge in the propeller mixed tank simply 'gelled' in the pipe.

    For some of the older operators on site it was hard to imagine that (in old money) a 1ft head of sludge would not push the liquefied sludge out of a six-inch pipe. When the pump was connected, pump rates were no more than 0.2 l/s. The solution was to dilute the sludge with very little water to yield a maximum of 7.5% TS sludge. This produced a pump rate of 6-10 l/s without compromising the performance of the downstream anaerobic digester. At this duty point The 600 Rule equation yields 7.5 x 80 = 600 - the '600 rule' held true again. Beware; we ignore this rule at our peril.

    A good idea from this experience: Fit the pump inlet line with three water flushing points to enable the operators to rectify a blocked sludge line with the aid of a small wash-water hose - much cleaner than the other option of dismantling the line at the flanges, flushing out each section onto the ground and re-assembling (we tried that - it ended up looking like the next exhibit for the Tate Modern - unfortunately we could not get a buyer).



    Click a keyword to see more stories on that topic, view related news, or find more related items.


    You need to be logged in to make a comment. Don't have an account? Set one up right now in seconds!

    © Faversham House Group Ltd 2005. edie news articles may be copied or forwarded for individual use only. No other reproduction or distribution is permitted without prior written consent.