How to Collect and Interpret Extrusion Process Data; Part 3

The trend plot shows a number of temperatures with a regular sinusoidal variation with the same frequency. The cycle time of the variation is about 6 min. The amplitude of the temperature variation varies considerably from 2° C up to 15° C (3.6 to 27° F). The IR measurement of tubing temperature after the die shows the largest variation (15° C/27° F). The melt-temperature variation measured with the immersion probe shows significantly less variation (2-3° C/4-5° F). The variation in tubing temperature is unacceptably large and indicates a problem.

The die temperature at the discharge end varies about 4°C (7° F), and it would be tempting to conclude that the die- temperature variation is causing the tubing-temperature variation. However, this is not possible because the tubing- temperature variation is about three to four times greater than the die-temperature variation. Clearly, the die-temperature variation is caused by the melt-temperature variation and not the other way around. A similar argument can be used for the variation in barrel temperatures.

There is little variation in motor load and no variation in screw speed. As a result, the tubing-temperature variation cannot be caused by changes in viscous dissipation. This leaves one other possible cause of the problem: variation in the feed material entering the extruder. Unfortunately, this extruder was not equipped with a pellet temperature sensor in the feed port. However, there was strong indirect evidence of pellet temperature variation. The polymer was dried before it was conveyed to the extruder. It turned out that the feed hopper was filled with hot, dry material every 6 min.

The feed hopper itself was not heated. Therefore, the hot pellets from the drier would cool down in the feed hopper as they made their way to the feed opening of the extruder. The hopper was filled at 6-min intervals. This created a 6-min cycle in the temperature of pellets entering the extruder. Once the cause of the problem is known, the solutions are obvious. One possible solution is to install a hopper drier on the extruder so that the pellets stay dry and at constant temperature in the hopper.

It should be noted that the problem would have been noticed immediately if the extruder had been equipped with a pellet temperature sensor at the feed opening of the extruder. Such a temperature sensor provides important information. Unfortunately, extruder manufacturers generally do not provide a temperature sensor in this part of the machine.

The variation in screw speed and motor current are quite severe. The screw speed varies between 80 and 160 rpm with a variation in motor current from 400 to 800 amps.

Calcium deposits in the cooling channels of the feed housing will reduce heat transfer and lead to higher temperatures. It is important to measure the inlet and outlet temperatures and the flow rate of the water. With these three data points, the amount of cooling can be quantified. If the amount of cooling declines over time, this is likely caused by calcium deposits.

For water cooling of the extruder feed housing it is good practice to use a closed-loop circuit. This reduces the chance of getting hard water in the cooling system. This applies not only to the water-cooled feed housing but also temperature zones along the extruder barrel that employ water cooling.

The barrel-temperature trend plot shows that the variation in barrel temperature was much too large—about 10-15° C (18-27° F). This suggested poor tuning of the temperature controllers for the barrel zones. When the tuning was corrected, the barrel-temperature variation reduced to normal levels and the pressure variation reduced from ±10% to about ±0.7%.