Injection molded part quality depends strongly on the key processing conditions such as injection velocity during filling phase, packing pressure during packing-holding phase, and melt temperature during plastication phase. Profiling methods of those key parameters are studied in this thesis....[ Read more ]

Injection molded part quality depends strongly on the key processing conditions such as injection velocity during filling phase, packing pressure during packing-holding phase, and melt temperature during plastication phase. Profiling methods of those key parameters are studied in this thesis.

Injection velocity is profiled to keep a constant melt-front rate throughout the mold filling to produce uniform parts. It can be implemented by controlling the average-flow-length following a ramp. To measure the melt-position and also to collect data for modeling and control, a capacitive transducer is designed and experimentally tested for different materials and molds. Potential applications of such a transducer for the detections of V/P transfer, gate freezing-off point, and over-packing, are also discussed. As such a hardware transducer may not be conveniently installed in all molds, a soft-sensor is developed via a neural network model to correlate the average-flow-length with other online measurable variables. With this soft-sensor model, a profiling method based on optimization is proposed and experimentally verified.

Two parameters defining a packing pressure profile, the gate freezing-off time, and the shape and level during packing-holding, are studied in this thesis. An online detection system for the gate freezing is developed, with results matching well with the established off-line method. Influences of the packing profiles on part weight, evenness, shrinkage, and flash are studied in details. Different types of packing profiles, including constant, ramp, and step change profiles, are compared using different mold inserts. Based on the rules concluded from this extensive experimental study, profiling methods are proposed.

The melt temperature is affected by several parameters in plastication phase. A transparent and instrumented barrel is set up to observe and to promote the understanding of the melting and transport phenomenon in the injection plastication. The melt temperature is correlated to other molding parameters via a neural network model. For a given material and a required melt temperature, proper settings of the plastication conditions are obtained through optimization.