Abstract
A constant melt-front velocity during the filling of an injection mold cavity is commonly believed to bring about more uniform part quality. To maintain a constant melt-front velocity, injection velocity can be set proportional to the melt-front-area which is, however, not directly measurable. An on-line soft-sensor scheme is developed through neural network to correlate on-line measurable process variables to the melt-front-area. Simulations indicate that the soft-sensor developed for the melt-front-area works reasonably well for some selected molds.
| Original language | English |
|---|---|
| Publication status | Published - 2000 |
| Event | ANTEC 2000: SOCIETY OF PLASTICS ENGINEERS TECHNICAL PAPERS, CONFERENCE PROCEEDINGS, VOLS I-III - Duration: 1 Jan 2000 → 1 Jan 2000 |
Conference
| Conference | ANTEC 2000: SOCIETY OF PLASTICS ENGINEERS TECHNICAL PAPERS, CONFERENCE PROCEEDINGS, VOLS I-III |
|---|---|
| Period | 1/01/00 → 1/01/00 |
ISBNs
['1-56676-855-1']Keywords
- Combined neural network
- Injection velocity
- Melt-front-area
- Soft-sensor