Abstract: The authors have undertaken a research task with a view to apply of evolutionary algorithms and artificial neural network to design of the course of a gas nitriding process, which guarantees to obtain the expected hardness profile in the nitrided layer. The gas nitriding process is widely used in industry in order to improve the functional properties of machine and tool parts. First of all, the artificial neural network was trained dependences between physical properties of steel and process parameters in relation to hardness profile, formed in gas nitriding process. Those data was collected experimentally. Such trained neural network was used as a mathematical model for the design approach of gas nitriding process. The aim of such designing was to predict the parameters of nitrinding industrial process, in which the required hardness profile of nitriding layer will be obtained. Prediction of manufactory conditions was realized with the help of evolutionary algorithms. In the approach developed, each chromosome includes encoded parameters of a single steel nitriding process. For each chromosome, a steel hardness profile related to it is determined by the model which is represented by neural network. On the further step, the chromosomes undergo a selection, a modification with the aid of crossing mutation operations and promotion to the next population. In this manner, by approach of a directed evolution of steel nitriding parameters, one is selected for which the hardness profile matches the best to profile sought.

Keywords: evolutionary algorithm, neural network, designing of nitriding process, nitriding layer, hardness profile, model, manufactory conditions prediction.

ACM Classification Keywords: D.2.2 Design Tools and Techniques - Evolutionary prototyping

Link:

THE APPLICATION OF ARTIFICIAL NEURAL NETWORKS AND EVOLUTIONARY ALGORITHM FOR THE DESIGNING OF GAS NITRIDING PROCESS

Mateusz Kosikowski, Zbigniew Suszyński, Roman Olik, Jerzy Ratajski, Tomasz Suszko

http://www.foibg.com/ibs_isc/ibs-13/ibs-13-p04.pdf