Investigation of torque fluctuations in extrusion through monitoring of motor variables

The extrusion process is highly prone to fluctuations in throughput and melt viscosity due to feed material variations and conveying, mixing, and thermal inefficiencies of the process. These fluctuations can result in poor product quality and higher scrap rates, hence early detection & diagnosis is important to minimize possible product defects and process inefficiencies. Usually extruders are equipped with melt pressure and temperature instrumentation at the screw tip. These give only point based measurements at the latter stage of the process but not an overall picture of the process inside the barrel. Therefore, more advanced methods are required for early diagnosis of process fluctuations. Screw torque variations offer an integral measurement of the polymer state along the length of the extruder. Such measurements have the potential to yield earlier and more sensitive indication of process instability than measurements at the die. Strain gauges have been used to investigate dynamic torque signals in a specially-designed solids conveying device, but are not practical or cost-effective for retrofit in a production environment. Also, the installation of a torque sensor in between screw and motor shaft is not easily achievable due to several constraints. However, there is potential to track screw load torque dynamics through the analysis of extruder motor current and voltage signal variations. In this work, process fluctuations in extrusion are investigated by analysis of melt pressure and high frequency motor signals. Experimental studies were carried out to investigate process fluctuations with a virgin and a recycled material under different operating conditions. The estimated torque signal is shown to be a useful complement to pressure measurement in the identification and diagnosis of process issues.