Roberto Mosca
Institute of Technology and Mechanical Plants University of Genoa via Opera Pia 15, 16145 Genoa, Italy
Pietro Giribone
Institute of Technology and Mechanical Plants
University of Genoa via Opera Pia 15 16145 Genoa, Italy
Agostino G. Bruzzone
Institute of Technology and Mechanical Plants University of Genoa via Opera Pia 15, 16145 Genoa, Italy
Sinopsis
This paper proposes neural networks together with plant simulations as the right tool for solving complex industrial problems. In this field, the large number of analysis ranges sometimes makes traditional approaches rather difficult to use. Therefore, since it was necessary to use this innovative Simulator/Neural Network combination, a method had to be developed which could match each result with a statistical evaluation and a correct estimate of its specific reliability. The proposed application of this approach is very effective in determining the critical elements of a real plant.