Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 International Journal of Applied Mechanics and Engineering

1 2010

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: modelowanie sieci neuronowych

Sortuj według:

Ogranicz wyniki do:

Modeling of magnetic properties of bulk amorphous alloys using an artificial neural network tool

Szota M., Borkowski S., Nabiałek M.

International Journal of Applied Mechanics and Engineering

2010

Vol. 15, no 2

571-578

This paper presents the possibility of using neural networks model for designing magnetic properties of bulk amorphous alloys. In recent years large quantity of data has been published about production methods of bulk metallic glasses (BMG). The most popular methods are: suction casting (Inoue and Tao, 1995; Ma et al., 2005) and injection casting (Park and Kim, 2004). From the microstructure investigation it was found that samples obtained as plates in contrast to rods have full amorphous microstructure. Samples as rods were partially crystallized. This process is complex and difficult as multi-parameter changes are non-linear. This fact and lack of mathematical algorithms describing this process make modeling properties of bulk amorphous-alloys by traditional numerical methods difficult or even impossible. In this case, it is possible to use an artificial neural network. Using neural networks for modeling is caused by several net features: non-linear character, the ability to generalize the results of calculations different from the learning data set, lack of need of mathematical algorithms describing influence of input parameters changes on modeling materials properties. The neural network structure is designed and specially prepared by choosing input and output parameters of the process. The method of neural network learning and testing, the way of limiting net structure and minimizing learning and testing error are discussed. Such a neural network model, after putting desirable values of bulk amorphous alloys properties in the output layer, can give answers to a lot of questions about production process. The practical implications of the neural network models are the possibility of using them to build control system capable of on-line process control and supporting engineering decision in real time. The originality of this research is a new idea to obtain desirable bulk amorphous alloys properties after crystallization process. The specially prepared neural network model could be a help for engineering decisions made in real time.