Combining static and temporal process data in the modelling of FSW weld quality and mechanical properties using computational intelligence
- Panoutsos G. ,
- Mahfouf M. ,
- Beamish K. ,
- et al
- Panoutsos G. ,
- Mahfouf M. ,
- Beamish K. and
- Norris I.
Links
Repository
Description
A Granular Computing (GrC) Neural-Fuzzy (NF) modelling approach that can learn relational data descriptions from a minimum number of experimental data is proposed for finding optimal process operating windows in friction stir welding (FSW). Research work involving experiment, computer intelligence modelling and process optimisation is presented. An orthogonal experimental design approach, based on Taguchi arrays, was used in which samples of aluminium alloy AA5083-O of 5.8 mm thickness were subjected to FSW at rotational speed between 280-580 rpm and travel speeds between 0.60 and 1.40 mm per minute. Plates were then subjected to a tensile test to assess weld strength.
8th International Symposium, 18-20 May 2010, Session 5B: Modelling 2, Paper 04
Related Records
