Selected parameters of moulding sands for designing quality control systems

• Autorzy: Jakubski, J. , Dobosz, S. M.
• Języki publikacji: EN

Abstrakty

EN

One of the modern methods of production optimisation are artificial neural networks. Neural networks owe their popularity to the fact that they are convenient tools, which can be utilised in a wide scope of problems. They are capable of reflecting complex functions. Especially their non-linearity should be emphasised. They are gaining wider and wider application in the foundry industry, among others, to control melting processes in cupolas and arc furnaces, designing castings and supply systems, control of moulding sands treatments, prediction of properties of cast alloys as well as selecting die casting. An attempt of the application neural networks to the quality control of moulding sands with bentonite is presented in the paper. This is a method of assessing the suitability of moulding sands by finding correlations in between individual parameters, by means of artificial neural network systems. The presented investigations were performed with the application of the Statistica 8.0 program. The investigations were aimed at the selection of the proper kind of a neural network for prediction a sand moistness on the bases of certain moulding sand properties such as: permeability, compactibility and friability. These parameters – determined as sand moistness functions - were introduced as initial parameters. Application of the Statistica program allowed for an automatic selection of the most suitable network for the reflection of dependencies and interactions existing among the proposed parameters. The best results were obtained for unidirectional multi-layer perception network (MLP). The neural network sensitivity to individual moulding sand parameters was determined, which allowed to reject not important parameters when constructing the network.

Słowa kluczowe

PL

zarządzanie jakością; masa formierska; sztuczne sieci neuronowe

EN

quality management; green sand; artificial neural networks

• Czasopismo: Archives of Foundry Engineering
• Rocznik: 2010
• Tom: Vol. 10, iss. 3
• Strony: 11--16
• Opis fizyczny: Bibliogr. 7 poz., rys., tab.

Twórcy

autor

Jakubski, J.

• AGH - University of Science and Technology, Faculty of Foundry Engineering, al. Mickiewicza 30, 30-059 Kraków, Poland,

autor

Dobosz, S. M.

• AGH - University of Science and Technology, Faculty of Foundry Engineering, al. Mickiewicza 30, 30-059 Kraków, Poland

Bibliografia

• [1] Z. Ignaszak, R. Sika, System do eksploracji wybranych danych produkcyjnych oraz jego testowanie w odlewni. Archiwum Technologii Maszyn i Automatyzacji. 2008, Vol. 28, 1, pp. 61-72.
• [2] Hülya Kaçar Durmuş, Erdoğan Özkaya, Cevdet Meri Ç, The use of neural networks for the prediction of wear loss and surface roughness of AA 6351 aluminium alloy. Materials& Design. 2007, Vol. 27, pp. 156–159.
• [3] Mahesh B. Parappagoudar D.K. Pratihar, Datta G.L, Forward and reverse mappings in green sand mould system using neural networks. Applied Soft Computing. 2008, Vol. 8, pp. 239-260.
• [4] M. Perzyk, A. Kochański, Prediction of ductile cast iron quality by artificial neural networks. Journal of Material Processing Technology. 2001, Vol. 109, pp. 305-307.
• [5] J. Jakubski, St. M. Dobosz, The use of artificial neural networks for green moulding sands quality control. Transaction of the VŠB – Technical University of Ostrava Metallurgical Series. 2009, Vol. 2, pp. 109-114.
• [6] M. Perzyk, R. Biernacki, A. Kochański, Modeling of manufacturing processes by learning systems: The naïve Bayesian classifier versus artificial neural networks. Journal of Material Processing Technology. 2005, 164–165, 430–435.
• [7] Internetowy podręcznik Statystyki, StatSoft, Inc., 1984-2003.