Study of ammonia-nitrate composition combustion processes using neural network modeling and software module development

This paper presents the results of a study of combustion processes of an ammonia-nitrate composition with the following mass ratio of components: 72 % ammonia-nitrate (NH4NO3), 25 % epoxy binder and 3 % potassium dichromate (K2Cr2O7), used as a mixed solid fuel for enhanced oil recovery. This study predicts the fundamental ballistic characteristics of the energy material itself. In order to improve the efficiency and reliability of the ammonia-nitrate composition charges, a software module based on a multilayer neural network, implemented in the NeuroShell environment, was developed and verified. The module provides an accurate prediction of the combustion rate (average relative error up to 6 %) for various operating parameters: charge density (1,4-1,46 g/cm³), pressure (5-25 MPa) and charge diameter (36 mm). The possibility of modeling complex nonlinear dependencies is demonstrated, including a decrease in the combustion rate by 22-23 % with an increase in charge density and an increase in pressure by 75-79 %. A comparative analysis of the experimental data is carried out and key patterns are revealed, such as the dominant effect of pressure and structural changes on the combustion kinetics. A comparative analysis of the obtained dependencies of the combustion rate on pressure and charge density is carried out. It is established that the curves of the dependence of the combustion rate on pressure have a similar nonlinear character. The results obtained open up prospects for designing highly reliable solid propellant systems in conditions close to real oil production processes.

References

1. Mukhutdinov A.R., Vakhidova Z.R., Lyubimov P.E., Improving the efficiency of the TP-230 boiler using neural network technologies (In Russ.), Vestnik Kazanskogo tekhnologicheskogo universiteta, 2011, V. 14, no. 21, pp. 91–94.

2. Mukhutdinov A.R., Vakhidova Z.R., Efimov M.G., Modeling of the combustion process of solid fuel in a furnace device (In Russ.), Vestnik Kazanskogo tekhnologicheskogo universiteta, 2014, V. 17, no. 20, pp. 114–116.

3. Mukhutdinov A.R., Marchenko G.N., Vakhidova Z.R., Neyrosetevoe modelirovanie i optimizatsiya slozhnykh protsessov i naukoemkogo teploenergeticheskogo oborudovaniya (Neural network modeling and optimization of complex processes and high-tech thermal power equipment), Kazan: Publ. of Kazan State Power Engineering University, 2011, 296 p.

4. Mukhutdinov A.R., Efimov M.G., Safiullin R.I., Mefod’ev A.V., Neural network based software module for predicting combustion rate of mixed solid fuel (In Russ.), Vestnik tekhnologicheskogo universiteta, 2017, V. 20, no. 24, pp. 102–104.

5. Mukhutdinov A.R., Lubimov P.E., Application of a neural network model for revealing specific features and regularities of solid fuel burning process, Thermal Engineering, 2010, V. 57, no. 4, pp. 336–340, DOI: https://doi.org/10.1134/S0040601510040105

6. Mukhutdinov A.R., Okulin M.V., Development of a neural network programming module for predicting the strength properties of solid fuel, Chemical and Petroleum Engineering, 2011, V. 47, no. 3, pp. 266–269, DOI: https://doi.org/10.1007/s10556-011-9457-3