Characterization of Superalloys by Artificial Neural Network Method

In this study, the use of artificial neural networks in classification a superalloys whose chemical analysis is performed quality process investigated. general, spectro method alone not sufficient to determine which class steel belongs to. addition method, tests such as tensile test, hardness test or notch impact are applied. The both take time and destroy material. fact that an algorithm classifies only according results used has made destructive mandatory. Artificial (ANNs), usually simply called (NNs), computing systems inspired by biological constitute animal brains. An ANN based on collection connected units nodes neurons, loosely model neurons brain. Each connection, like synapses brain, can transmit signal other neurons. our total 34 superalloy materials belonging 6 different classes were used. Chemical composition values determined for each sample. appropriate network was values. A predict material value been created. Weka 3.9.5 package program create model. high success rate prediction gave hope determination with method.