Naslov
Determination of the physical properties of heat treatable steels

Autori
Filetin, Tomislav ; Žmak, Irena, Markučič, Damir ; Novak, Davor

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Integration of heat treatment and surface engineering in the manufacture of engineering components / Liščić, Božidar - Zagreb : Hrvatsko društvo za toplinsku obradu i inženjerstvo površina, 2001, 399-406

Skup
8th Seminar of the International Federation for heat treatment and Surface Engineering

Mjesto i datum
Dubrovnik, Hrvatska ; Cavtat, Hrvatska, 12.09.2001. - 14.09.2001

Vrsta sudjelovanja
Poster

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
toplinska difuznost ; toplinska vodljivost ; toplinski obradljivi čelici ; predviđanje svojstava ; neuronske mreže
(thermal diffusivity ; heat conductivity ; heat treatable steels ; prediction of properties ; neural networki)

Sažetak
The following data: heat conductivity, specific heat, the coefficients of linear thermal expansion, the densities and other physical properties of steels versus temperature, for calculation and simulation of heating and cooling processes of different heat treatment technologies, are needed. In literature there could be found a restricted amount of data for steel grades in question - for defined chemical composition of steel and temperature. In the determination of relevant data for steel in question the statistical methods and/or methods of artificial intelligence for predicting of properties are applicable. An issue in this approach was the development of method for prediction of heat conductivity for known chemical composition of steel and defined temperature. This paper presents the results of predicting the heat conductivity depending on temperatures of different steels, using the regression analysis and by means of neural network. The regression equations for estimation of heat conductivities at different temperatures are defined. The model for statistical analysis is determined by genetic algorithm. Acceptable correlation between the input variables - sum of elements of steel and temperature and the heat conductivities are estimated. The static multi layer perceptron neural network is proposed to predict the heat conductivity of steels. To accelerate the convergence of proposed static error back propagation learning algorithm, the momentum method is applied. In the learning datasets, 41 different constructional, corrosion resistant and tool steels are used. The inputs for learning and testing were chemical composition and measured data from literature for heat conductivities at different temperatures (between 20 and 700  C). The mean error between measured and predicted data and standard deviation for testing steel types is found to be acceptable. The results direct to further testing of wider dataset of steel groups and of other physical properties.

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo

POVEZANOST RADA