When and why should a melt pump be used?

When and why should a melt pump be used?
The screw in an extruder is basically a pump. There are other types of pumps: the piston pump, the peristaltic pump, and the gear pump are three common types.

At first glance, the screw type pump would seem to be an odd choice for use in an extruder. Because in an extruder constancy of output is very important, and the screw type pump is the only one of the four types that is not a positive displacement pump – that is, a pump which always pumps a specified volume of product for every revolution. The other three pump types – the gear pump, the piston pump, and the peristaltic pump – are all positive displacement types.

The screw type pump is chosen for the plastic extruder over the other pump types because it is an inefficient pump! That is not to say that it is a bad pump – just that it generates a lot of heat. But think about that for a minute. How else would you melt a Gaylord of plastic pellets? If you just put your pellets in a pot with a fire under it, there isn’t a polymer in the world that would not be hopelessly degraded on the outside long before the plastic in the middle was even warm. The only practical way to melt a lot of plastic is to take a little bit of it at a time and subject it to frictional heat generated by a screw type pump.

The screw type pump is also very good for mixing plastic. The barrier screw, the Maddock mixer, and a host of other types all mix by subjecting the plastic to a shearing action. This is a form of so-called dispersive mixing. An example of this type of mixing is the classic demonstration in high school chemistry class, where the teacher demonstrates how to mix carbon particles into oil. He spreads the particles on a hard with a spatula to break them up and disperse them into the oil: dispersive mixing!

The screw is also good at so-called distributive mixing. This is the type of mixing done in the common kitchen mixer. It does not change the composition of the melt; it just ensures that every part of the mixture has the same components.

More recently a third type of mixing screw has made its appearance. This is the so-called CRD mixer, which achieves its mixing through elongation and stretching of the plastic. It turns out that this type of mixing is superior to either of the other two types.

There are two things, however, that the screw type pump is not particularly good at – generating a constant pressure and delivering a constant throughput of plastic. The reason, of cause, is that the screw type pump is not a constant displacement type pump. There is always an open channel between the input and the output when pumping plastic forward and the channel is always open for the plastic to flow backwards as well.

An extruder can generate substantial pressure, of course. It does so because the plastic sticks to the barrel and slips on the screw. If it slips on both or sticks on both, we are in trouble.


But there are many factors that affect the degree to which plastic sticks on the barrel:
The melt index and other characteristics of the plastic
Whether virgin material or regrind is used
The past history of the plastic, and above all
The temperature of the barrel and of the plastic

The performance of an extruder is conditioned by the channel always open between the input and the output

All of these factors can be highly variable, and all affect the pressure generation. And obviously the throughput will also change when the pressure of the plastic varies as it goes through the die.

The throughput of the plastic which the screw can deliver while maintaining a given output pressure is also variable because the screw type pump is not a constant volume device. Pumping must be done with an open channel always in existence from output to input, see the fig. note that this statement is still true, even in the pressure in the plastic path of a mixer, such as a Maddock mixer, or a barrier, as in a barrier screw. For the mixing action in either case results when the plastic passes over the barrier and is subjected to the resulting shearing action. In neither case is the screw a positive displacement device, where a constant volume of plastic is pumped with each revolution, irregardless of back pressure and other phenomena.

The curves in figure 1 and The figure 2 illustrate these variations in two different ways. The data were created with the same general purpose screw with a 3:1 compression ratio, that is, the ratio of the pressure in the last section of the screw (the melting section) to the first screw section (the feed section).

The figure 2 is a conventional plot showing how throughput varies with back pressure at a constant RPM. The class method of analyzing the behavior of a screw type pump is to think of the delivery as consisting of two components – a drag flow, determined only by the dimensions of the metering section of the screw, minus a back flow which is an assumed reverse flow back down the open channel resulting from the back pressure at the output. The throughput drops 44% in the face of a 4,000 psi variation in b ack pressure!

The figure 3 shows the variation in RPM required by a screw to maintain constant throughput of plastic. Note that this screw was not selected to be bad. It is a perfectly typical general purpose screw designed for polystyrene. In addition to the usual feed section, transition section, and metering section, it had a Maddock mixer (equipment to a barrier screw) for dispersive mixing, and a Saxton mixer, for distributive mixing. These two are very important to mix the plastic and remove temperature variations, but they do not produce a constant displacement device. It can be seen that it takes a 68% change in the RPM to keep the throughput constant in the face of a 4000 psi increase in the screw output pressure.

Note that these data were collected within a short period of time with the same lot of plastic, with highly sophisticated controls for all parameters of the extruder, and with constant operating conditions in the extruder. Performance in an actual operating environment, with random lots of plastic, varying temperatures of the plastic feed, and the usual differences of opinion on operating conditions among operators, would be much worse.

These are those who may argue that constancy of output is not important in these days of closed-loop controls over the various facets of the process. This argument falls short for two reasons:
1. First, no corrective control can ever completely compensate for basic parameters that are not stable. Stabilizing those basic parameters – and throughput stability is certainly at the top of the list – in an important first step in taming any process.
2. Furthermore, some of the most potent closed-loop control schemes start with the assumption of constant throughput as a given. There simply is no substitute.

Stabilizing the throughput
By far the way to get a stable throughput from an extruder is to use a melt pump on the output. In concept the melt pump is very simple. It uses a pair of closely meshed gears, placed in the main plastic stream. The teeth are filled at the input, leveled off by the housing, and then emptied at the meshing point. The device can be thought of as “measuring cups on a wheel”. The teeth are loaded at the input and emptied at the output, see fig.4.

The important thing to note is that the gear pump is a positive displacement device. There is no channel open between input and output, as in a screw type pump. Every time the gears advance one tooth, precisely two teethful of plastic are delivered from input to output, and insofar as the primary pumping mechanism is concerned, the amount of plastic delivered will be independent of back pressure.

Overall, the throughput is not really completely independent of the pressure, of course. Extrusion grade gear pumps are made to extremely close tolerances. Both the gears and the housing are precision ground to size rather than being simply machined. Nevertheless, there will still be clearances required for the gears to turn, and where there are clearances, there will be leakage. The gear pump in fig.4, for example, has 1200 psi tending to push the plastic back from output to input – along the sides of the gears, around the periphery of the gears, and through the bearings.

The “efficiency” curve, or throughput plot, of the gear pump used with simple input pressure control will typically look like fig.5. It has the same shape as the curve for the extruder shown earlier, but the throughput variations with pressure are reduced by a factor of about 15:1. As we shall see in the following, this reduction in throughput variation can be reduced even further through proper control.

4 Microprocessor melt pump control
As mentioned earlier, the performance of any extrusion line can be significantly enhanced though the use of automated closed-loop controls. It is, however possible to control the extrusion line manually, and many people still do.

Manual control, however, is not an option for melt pumps. In the following we will explain why it is simply not practical to control a melt pump manually.

The problem lies in the nature of the interaction between the melt pump and the extruder. Consider the back pressure in the extruder when it is feeding plastic into a melt pump turning at a constant speed while the speed of the extruder is gradually increased, see fig.6.

As long as the RPM of the extruder is such that the plastic pumped into the melt pump is not sufficient to fill the teeth of the pump, the back pressure is minimal. The melt pump simply takes away whatever the extruder is pumping.

As soon as the output of the extruder is sufficient to fill the teeth, however, the situation changes dramatically. Now it is just as if the extruder is pumping into a brick wall. The melt pump can accept only so much plastic and no more. Attempting to pump more plastic skyrockets the pressure upwards until it blows the safety blow-out plug, at which time the pressure drops to zero, and the ball game is over (see fig.6)

Our problem, however, is that if we are going to use the melt pump to meter out the plastic, operation has to be in the steep vertical part of this pressure characteristic, because the melt pump can’t do any metering unless the teeth are filled, and that only occurs on the steep vertical portion of the curve. But no human operator can maintain that exact balance for very long. Manual control is simply not feasible!

An examination of the plastic path in the extruder/melt pump combination turns up some other interesting facts:
·Most of the change in output shown in fig.6.5 is not proportional to the level of the pressure in the input to the melt pump. It is proportional to the difference in pressure across the pump! In the melt pump of fig.6.4, for example, there is a difference of 1200 psi trying to drive the plastic through the leakage paths around the sides of gears, around their periphery, and through the bearings.

·We can never eliminate these leakage paths. But as long as the pressure drop across them remains constant, the amount of plastic passing through them will remain constant too, and so will the net throughput of the melt pump / extruder combination, see fig.6.7.

·Since a microprocessor controller is generally used to control the temperature anyhow, it poses no problem to control not just the inlet pressure to the pump but also rather the differential pressure across that pump (fig. 8). Now the curve looks like the one in fig.7, where operation is always at the same point in the pressure / throughput curve. Controlling the differential pressure across the pump in this manner improves the throughput stability of the extruder/melt pump combination to about 0.1% or better.

Throughput stabilized by holding differential pressure across pump constant

Differential pressure controller

The prime reason for adding a melt pump to an extrusion system is to improve the throughput stability – from between 5% to 8% typical for an extruder used alone to the 0.1% of the typical melt pump with differential pressure control. There are, however, other major advantages in using a melt pump at the output of the extruder:
· The output can almost always be increased – typically by 15 to 25%.
· The throughput curve of an extruder used alone as shown in fig.2 illustrates that the screw type pump in an extruder is a very inefficient pump. If there is any significant pressure at the die, the output of an extruder used alone is reduced by a significant amount, up to 45% if the back pressure is 4000 psi.
· But when a melt pump is used, the extruder only has to generate enough pressure to fill the teeth of the gear pump. The gear pump will generate the pressure required to push the plastic through the die.
· The pressure required to push the plastic through the die is now being generated by an efficient pump, the melt pump, rather than by the very inefficient screw type pump. Even though there are now two motors instead of just one, the total power consumption is almost always less.
· The melt temperature of the plastic is almost always significantly lowered.
· Reducing the total power consumed also reduces the excess thermal energy imparted to the plastic. This can be a significant advantage for heat sensitive plastics, such as PVC. It also reduces the downstream cooling required, since there is less heat to be removed. In a blown film application, for example, lower melt temperature means higher throughput for a given height of cooling tower.
· Much higher die pressures can be achieved, where they are required.
· The melt pump is simply a far more efficient pump!
So, don’t be a Luddite: use a melt pump wherever high throughput stability and close tolerances are required.

6. Measurement and control of melt pressure
When it comes to the placement of melt pressure sensors the extrusion engineer faces considerably fewer problems compared to the placing of temperature sensors. One of the basic principles of hydraulics is that the pressure in a fluid is equal in all directions. This may not be strictly true in the case of a very viscous, moving fluid, but it is close enough for our purposes. One of the few things that the extrusion industry was able to standardize is the shape and size of the mounting hole in the extruder for a melt pressure sensor. This mounting was diagramed in fig.3.
A typical pressure sensor (see fig.9) will have a thin diaphragm at the end, in contact with the plastic. This diaphragm which is a measure of the pressure.
The pressure is ultimately measured by a pressure sensitive transducer of some sort. Unfortunately, none of the practical pressure transducers can stand the elevated temperature of the plastic stream, so it is necessary to locate them in the remote end of the sensor body.
The problem of coupling the diaphragm to the transducer body has led to the diversity in the different pressure sensors available. The earliest ones simply used a rod to couple the two. The problem here was the unpredictable thermal expansion of the rod. One end was very hot, being in contact with the diaphragm, which in turn is in contact with the hot plastic. The other end is in contact with the pressure transducer, located in the other end of the sensor housing. The expansion of the rod in between, due to temperature, was very erratic.

---

pre：Improperly designed metering section for a two-stage screw
next：Rubber compound custom mix adds benefits to melt gear pumps

Name:	*
E-mail:	*
Tel/Cell:
country/area:
Message:
CAPTCHA:	mistiness?