Integrated Extruder Plant Automation with Learning Control

Integrated automation in a manufacturing plant demands control of manufacturing and material and product flow processes. Efficient data archiving and retrieval, quality control and process automation facilities are essential components which should be deployed to form one coherent system. As managers of extruder plants running look for ways and means for meeting market requirements, the demand for such integrated automation systems is rising. Integrated Process automation involves: - Overall production and process control - Data acquisition, archiving evaluation with feed-back to production planning and process control - Decentral control in local control loops for achieving isothermal extrusion - On-line visualisation of process variables for tighter man-machine interaction Providing the facility for prescribing the ranges of relevant process variables viz. billet temperature, die temperature, profile temperature at die exit, rate of cooling of the profile and extrusion speed is part of the integrated automation system. At the heart of the system are processes for mastering the core task of management of the thermal processes in the extruder by employing sensors and control techniques. Goal is to minimise the time to extrude a billet and simultaneously keep the process variables within prescribed limits. Automation in extrusion plants is made possible by progress in sensor technology and enhanced possibilities of data processing and organisation, advanced control methods which require simple models and facilities of integration of hardware and software employing state-of-the-art networking and protocols. Crucial is the non-contact temperature measurement of the billet and the profile at die exit. Since about two decades there has been a continuous progress in temperature measurement technology based on multi-spectral radiation pyrometers. They perform well for a large number of alloys and surface finishes and are reliable for industrial use. Disturbances can be suppressed using adequate signal processing