On the issue of developing the diesel engine neural network controller

The paper considers issues of using a neural network in the thermal engine controller in order to improve the quality of its control and setting conveniences. The study object was a speed controller for the D500 promising locomotive diesel engine. Tasks for design and development of a neural network controller were formulated, and the network input signals were determined. To adjust the neural network, the reinforcement learning method was introduced, where it interacted with the diesel engine computer model in a closed system. The criterion in setting up the network was the accuracy of the control program execution. A system of rewards was assigned, according to which the network was learning. Based on the results of studying the neural network controller influence on the quality of the control system operation, the network minimum possible composition for solving the problem presented was determined. Study results are presented in the form of graphs of rewards alteration during the learning process for various options of the neural network controller structure, as well as the control system simulated transient processes over the entire range of the diesel engine speed.

References

[1] Kobyakov D., Khvatov O. Modeling transient processes of diesel-generator set with variable rotation frequency on the basis of active voltage rectifier. Vestnik Astrakhanskogo gosudarstvennogo tekhnicheskogo universiteta. Ser. Morskaya tekhnika i tekhnologiya [Vestnik of Astrakhan State Technical University. Ser. Marine Engineering and Technologies], 2019, no. 3, pp. 94–104, doi: https://doi.org/10.24143/2073-1574-2019-3-94-104 (in Russ.).

[2] Ouladsine M.B., Gérard D.X. Neural modelling and control of a Diesel engine with pollution constraints. J. Intell. Robot. Syst., 2005, vol. 41, no. 2–3, pp. 157–171, doi: https://doi.org/10.1007/s10846-005-3806-y

[3] Palaguta K.A., Chirkin S.Yu., Kuznetsov A.V. Synthesis of systems for internal combustion engine control using hybrid and neural networks. Mashinostroenie i inzhenernoe obrazovanie, 2009, no. 4, pp. 59–66. (In Russ.).

[4] Ryzhov V.A. Native engines of D500 new generation. Novyy oboronnyy zakaz. Strategii [New Defense Order Strategy], 2015, no. 5, pp. 40–41. (In Russ.).

[5] Voronov K.E., Grigoryev D.P., Telegin A.M. Application of the direct propagation neural network for localization of the impact site of microparticles on the surface of the spacecraft. Trudy MAI, 2021, no. 118, doi: https://doi.org/10.34759/trd-2021-118-10 (in Russ.).

[6] Fedotov M.V., Klimenko Yu.I., Sharapov A.L. et al. Assessment and prediction of changes in diesel oil viscosity using neural network models. Zheleznodorozhnyy transport, 2022, no. 7, pp. 39–42. (In Russ.).

[7] Fedotov M.V., Grachev V.V. Predictive analytics of the technical condition of diesel locomotive systems using neural network predictive models. Byulleten rezultatov nauchnykh issledovaniy [Bulletin of Scientific Research Results], 2021, no. 3, pp. 102–114, doi: https://doi.org/10.20295/2223-9987-2021-3-102-114 (in Russ.).

[8] Velichko E.P., Sokolchik P.Yu. Neural network controllers in the system position control. Vestnik PNIPU. Khimicheskaya tekhnologiya i biotekhnologiya [PNRPU Bulletin. Chemical Technology and Biotechnology], 2015, no. 2, pp. 7–20. (In Russ.).

[9] Lysenko D.S. Intelligent system of adaptation typical laws automatic regulation with the function of identification structure and parameters object. Vestnik Samarskogo gosudarstvennogo tekhnicheskogo universiteta. Ser. Tekhnicheskie nauki [Vestnik of Samara State Technical University. Technical Sciences Series], 2022, vol. 30, no. 1, pp. 46–62, doi: https://doi.org/10.14498/tech.2022.1.3 (in Russ.).

[10] Ecoffet A., Huizinga J., Lehman J.J. et al. First return, then explore. Nature, 2021, vol. 590, art. 7847, doi: https://doi.org/10.1038/s41586-020-03157-9

[11] Kuznetsov A.G., Kharitonov S.V. A study of diesel engine operation with the focus on locomotive characteristics. Izvestiya vysshikh uchebnykh zavedeniy. Mashinostroenie [BMSTU Journal of Mechanical Engineering], 2017, no. 12, pp. 11–19, doi: https://doi.org/10.18698/0536-1044-2017-12-11-19 (in Russ.).

[12] Kuznetsov A.G., Kharitonov S.V., Vornychev D.S. A mathematical model of a diesel engine for simulation modelling of the control system. GJPAM, 2016, vol. 12, no. 1, pp. 213–228.

[13] Sharma S., Athaiya A. Activation functions in neural networks. IJEAST, 2020, vol. 4, no. 12, pp. 310–316, doi: https://doi.org/10.33564/IJEAST.2020.v04i12.054

[14] Feng J., Lu S. Performance analysis of various activation functions in artificial neural networks. J. Phys.: Conf. Ser., 2019, vol. 1237, art. 022030, doi: https://doi.org/10.1088/1742-6596/1237/2/022030

[15] Malyshev A.N., Danilov E.A. Overview of some activation features math neuron. Nauchnoe obozrenie. Pedagogicheskie nauki [Scientific Review. Pedagogical Science], 2019, no. 3–2, pp. 39–43. (In Russ.).