Abstract
In this paper, a fuzzy logic-based embedded software, Fuzzy Expert System for X-ray Inspection (FESXI), for the inspection of soldering defects in an electronic manufacturing process of Nokia Networks, is presented. The FESXI was developed to improve the X-ray inspection results of signal transmission products by reducing false alarms to prevent unnecessary rework and increasing the throughput time of production.
Inspection results are inferred from a set of input measurement values. Each input value is fuzzified using automatically generated and continuously updated fuzzy membership functions, and fuzzy results are inferred with the aid of a compact rule base, where rules are presented by linguistic relations changed into matrix equations form.
The model produces more accurate decision-making support than the traditional dual logic-based approach of commercial X-ray machines. Experimental results showed that false alarms can be reduced by about 44% using the software, and since the first test, it has been in daily production use.
Inspection results are inferred from a set of input measurement values. Each input value is fuzzified using automatically generated and continuously updated fuzzy membership functions, and fuzzy results are inferred with the aid of a compact rule base, where rules are presented by linguistic relations changed into matrix equations form.
The model produces more accurate decision-making support than the traditional dual logic-based approach of commercial X-ray machines. Experimental results showed that false alarms can be reduced by about 44% using the software, and since the first test, it has been in daily production use.
| Original language | English |
|---|---|
| Pages (from-to) | 111-125 |
| Journal | Journal of Manufacturing Systems |
| Volume | 21 |
| Issue number | 2 |
| DOIs | |
| Publication status | Published - 2002 |
| MoE publication type | A1 Journal article-refereed |
Keywords
- decision making