%0 Journal Article %A Abrukov, Victor %A Troeshestova, D. A. %A Chernov, A. S. %A Pavlov, R. A. %A Smirnov, E. V. %A Malinin, G. I. %A Volkov, M. E. %D 2007 %I Begell House %N 6 %P 665-679 %R 10.1615/IntJEnergeticMaterialsChemProp.v6.i6.10 %T APPLICATION OF ARTIFICIAL NEURAL NETWORKS FOR SOLUTION OF SCIENTIFIC AND APPLIED PROBLEMS FOR COMBUSTION OF ENERGETIC MATERIALS %U https://www.dl.begellhouse.com/journals/17bbb47e377ce023,46ae61aa3e0cddf4,61f570e615399cb1.html %V 6 %X A goal of the paper is a presentation of possibilities of artificial neural networks (ANN) technologies for combustion modeling, experimental investigation, and diagnostics. A short introduction to ANN technologies and their applicability to scientific problems is included. For the first time, the use of sigmoid function for obtaining analytical solutions of combustion wave propagation differential equations has been suggested. Methods for the development of new computational models of combustion, as well as possible applications of ANN are discussed in this paper. A goal of the models is to solve the hard tasks involved with the experimental investigation of propellant combustion. An example of ANN being used for an investigation of deflagration-to-detonation transition (model of pulse detonation engine) under various experimental conditions is presented. It is shown that ANN can make up for gaps in experimental data. New ways for solving inverse problems of optics and their direct tasks by means of ANN are presented also. It is shown that ANN can be used for many optical and non-optical methods of combustion diagnostics as well as for testing and controlling combustion processes when the usual approaches cannot be used. Possibilities for the use of the only value of a function of signal distribution in a plane of the registration in order to determine full distribution of local characteristics in an object, as well as its integral characteristics, are shown. %8 2007-11-01