الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 124 |  |  |
| | | from | 124 | | |
Abstract
This thesis presents an approach to improve the characteristic of manufactured products using a performance prediction model developed by artificial neural networks technique. The proposed approach consists of two phases. The first phase explains how to determine the factors affecting the performance of the manufactured parts by designing experiments and to derive a model for measuring performance using artificial neural networks. The model is trained using data obtained from metrology measurements. Only those measurements are used that have effect on product performance. The second phase of the approach explains how to take advantage from this predicted model to get the largest number of manufactured products, which have better qualities through parts allocation of the matched assembled parts. This approach is explained and evaluated through a real case study for the manufacturing of hermetic reciprocating compressors. Through illustrative example, the results showed that the proposed approach achieved an improvement by transferring 15% of products that have the low performance level into which have the higher performance levels