Abstract
Filling plays a key role in determining part quality in injection molding. On-line measurement of the melt-flow-length, a key melt-flow status in mold cavity, is of great importance in both the understanding and control of the process. In most cases, a hardware measurement of such a variable is not available. A soft-sensor measurement scheme is proposed taking online measurable variables as the model inputs. With the experimental data obtained from a set of purposely designed molds with basic feature geometry, a soft-sensor based on a recurrent neural network has been developed to predict the melt-flow-length. Experiments show that such a developed soft-sensor can predict well the melt-flow-length for filling of molds, which have not been used in the training, as long as the basic features of the mold geometry have been included in the training mold set. © 2004 Published by Elsevier B.V.
| Original language | English |
|---|---|
| Pages (from-to) | 245-254 |
| Journal | Materials Science and Engineering A |
| Volume | 384 |
| Issue number | 1-2 |
| DOIs | |
| Publication status | Published - 25 Oct 2004 |
| Externally published | Yes |
Bibliographical note
Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].Research Keywords
- Filling
- Injection molding
- Melt-flow-length
- Recurrent neural network
- Soft-sensor