RT Journal Article ID 1f7a4012681992be A1 Abrukov, S. V. A1 Karlovich, E. V. A1 Afanasyev, V. N. A1 Semenov, Yu. V. A1 Abrukov, Victor T1 CREATION OF PROPELLANT COMBUSTION MODELS BY MEANS OF DATA MINING TOOLS JF International Journal of Energetic Materials and Chemical Propulsion JO IJEMCP YR 2010 FD 2011-10-07 VO 9 IS 5 SP 385 OP 396 K1 burning characteristics K1 artificial neural networks K1 computational models K1 generalization K1 predicting AB The possibilities of data mining tools in particular artificial neural networks (ANN) for modeling and prediction of propellant burning characteristics are presented and discussed. In this paper ANN computational models are created for this purpose. They allow modeling temperature profiles in propellant combustion waves, predicting burning rate of various propellant mixtures for different ranges of pressure and initial temperature, determining propellant mixture providing a necessary burning rate for various pressures, etc. Methods for the creation of such models are discussed and illustrated. The results obtained show that ANN can be considered as a good approximation tool for experimental functions of several variables, for the generalization and prediction of the connections between variables of combustion experiment and theory, as a fast engineering calculator specialized to combustion tasks, for the study of combustion behaviors, as a more affordable way to receive "new" experimental results, and as a good tool to show the experimental results obtained. PB Begell House LK https://www.dl.begellhouse.com/journals/17bbb47e377ce023,6c0548345417f04e,1f7a4012681992be.html