Abstract
This paper describes sensor-based methodologies for novel real-time quality control strategy in manufacturing. Feasibility of the methodology is demonstrated in two manufacturing processes: turning and powder injection molding. In case of turning, the process is monitored using force and vibration sensors. The level of tool wear on the cutting tool is estimated using the information from these sensors. The quality of workpiece is maintained by adjusting the operating conditions and cutting tool position to compensate for the undesirable effects of the progressive tool wear. In case of powder injection molding, the product quality factors are obtained by visual inspection and process states are abstracted from the sensor data. The diagnostic knowledge is represented by a causality network and solved by network conversion method and abductive reasoning. The diagnostic solution provides a set of alternative hypothesis of disorders which explains the manifested quality factors and process states.
| Original language | English (US) |
|---|---|
| Pages | 81-87 |
| Number of pages | 7 |
| State | Published - 1994 |
| Event | 1994 AAAI Spring Symposium - Palo Alto, United States Duration: Mar 21 1994 → Mar 23 1994 |
Conference
| Conference | 1994 AAAI Spring Symposium |
|---|---|
| Country/Territory | United States |
| City | Palo Alto |
| Period | 3/21/94 → 3/23/94 |
All Science Journal Classification (ASJC) codes
- Artificial Intelligence