Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
406968 | Neurocomputing | 2014 | 10 Pages |
The present paper describes a neural network-based strategy for crack prediction aimed at improving the steel-casting process performance by decreasing the number of crack-generated failure cases. A neural system to estimate crack detection probability has been designed, implemented, tested and integrated into an adaptive control system. The neural system, consisting of two distinct neural networks, provides 0 or 1 probability values (1—high probability of crack occurrence, 0—low probability of crack occurrence). Also, a decision block, based on fuzzy logic (implementing an expert system), has been designed and implemented, triggering one or the other specific set of rules (according to 0 or 1 value of neural system) and tuning the set point of the control system.