© Vitalii Krivda, PhD, Associate Professor, ORCID: 0000-0002-8304-2016, е-mail: krivda.v.v@nmu.one; © Olha Sakno, PhD, Associate Professor, ORCID: 0000-0003-4672-6651, е-mail: sakno.o.p@nmu.one; © Valentyna Olishevska, PhD, Associate Professor, ORCID: 0000-0002-3098-1351, е-mail: olishevska.v.ye@nmu.one (Dnipro University of Technology)
JUSTIFICATION OF A NEURAL METHOD FOR CONTROLLING THE TYPE OF POWER SUPPLY OF AN INTERNAL COMBUSTION ENGINE DEPENDING ON OPERATING CONDITIONS DOI: 10.33868/0365-8392-2024-4-281-15-22
Abstract. The rapid development of the modern automotive industry and increasing requirements for environmental friendliness, energy efficiency, and economy necessitate the development of new control systems for internal combustion engines (ICEs). A topical issue for the strategic development of road transport is the design of new technologies: a combination of different fuel types and neural control of ICE power supply. The paper analyzes the combination of types of ICE power, namely motor gasoline, diesel, and gaseous fuels, and a neural method of controlling engine power based on operating conditions. A neural approach to optimizing fuel choice by operating conditions is presented. The neural system allows you to automatically adjust the fuel supply parameters using an adaptive PID controller (proportional-integral-differential controller), which adjusts the fuel supply in real time based on external conditions. The neural network can use temperature, load, and engine speed as input parameters. To consider more complex conditions of internal combustion engine operation, the model can be supplemented with vehicle speed, atmospheric pressure, wear of engine parts, and driver reaction. Using multi-layer neural networks allows you to adapt the fuel supply in real-time. The architecture of a neural network for this task may include several layers: an input layer that receives ambient temperature, engine load, engine speed, fuel type, atmospheric pressure, and wear level of parts; a hidden layer that performs nonlinear transformations and makes decisions based on the interaction of parameters; and an output layer that generates a fuel supply control signal. The neural network can be trained based on data from real or simulated conditions, which allows it to predict the most economical mode of operation of the internal combustion engine. Adaptive neural networks can be used to optimize the ratio of fuels. A graphical visualization is performed to help understand that fuel consumption depends on temperature and load for each fuel type. Keywords: internal combustion engine (ICE), fuel type, intelligent neural control systems, model for controlling the fuel supply process, operating conditions, optimization of fuel feed.
References
1. Nykyforuk, O. I., Chukayeva, I. K, Piriashvili, B. Z., Lyashenko, O. F., Karpov, V. M., Kudryts’ka, N. V. … Plyuta, I. Yu. (2017). Development of infrastructure sectors as a factor in the implementation of priority areas of Ukraine’s economic policy: a collective monograph [Rozvytok infrastrukturnykh sektoriv yak chynnyk realizatsiyi priorytetnykh napryamiv ekonomichnoyi polityky Ukrayiny: kolektyvna mono-grafia]. Kyiv: NASU, SI “Institute of Economics and Forecasting of the NAS of Ukraine” [in Ukrainian].
2. Olishevs’ka, V. Ye., Olishevs’kyy, H. S. (2024). Substantiating rational rolling stock at an enterprise in the transition to electric vehicles [Obgruntuvannya ratsional’noho rukhomoho skladu pidpryyemstva v umovakh perekhodu do elektromobiliv]. Avtoshlіakhovyk Ukrayiny. 2, 35–44. DOI: 10.33868/0365-8392-2024-2-279-35-44 [in Ukrainian].
3. Ponedilok, Yu. V. & Kryvoshapko, S. B. (2020, Lystopad). Problems and directions of development of the road transport system [Problemy ta napryamky rozvytku avtomobil’notransportnoyi systemy. Рaper presented at the mizhnarodnoyi naukovo-tekhnichnoyi konferentsiyi «Intelektual’ni tekhnolohiyi upravlinnya transportnymy protsesamy»]. Kharkiv, 186–188. Retrieved from https://sci.ldubgd.edu.ua/bitstream/123456789/7218/1/%D0%97%D0%91%D0%86%D0%A0%D0%9D%D0%98%D0%9A_%D0%9C%D0%90%D0%A2%D0%95%D0%A0%D0%86%D0%90%D0%9B%D0%86%D0%92.pdf [in Ukrainian].
4. Bazhynov, O. V., Smyrnov, O. P., Syerikov, S. A., Hnatov, A. V. & Kolyesnikov, A. V. (2008). Hybrid vehi-cle: a monograph [Hibrydni avtomobili: monografia]. Kharkiv: KhNADU [in Ukrainian].
5. Kubich, V. I. (2021). Hybrid powertrains for passenger vehicle: educational manual [Hibrydni sylovi ustanovky lehkovykh avtomobiliv: navchal’nyy posibnyk]. Zaporizhzhia: NU «Zaporizhzhia Polytechnic» [in Ukrainian].
6. Lyashuk, O. L., Plekan, U. M., Ts’on’, O. P. & Hevko, B. R. (2023). Development technologies of cars hybrid power plants [Rozvytok tekhnolohiy hibrydnykh sylovykh ustanovok avtomobiliv. Tsentral’noukrayins’-kyy naukovyy visnyk. Tekhnichni nauky], 8(39), I, 139–146. https://doi.org/10.32515/2664-262X.2023.8(39).1.139-146 [in Ukrainian].
7. Ziegler, J. G. & Nichols, N. B. (1942). Optimum settings for automatic controllers. Transactions of the ASME, 64, 759–768.
8. Shuklin, D. Ye. (2003). Semantic neural network models and their applications in artificial intelligence systems [Modeli semantychnykh neyronnykh merezh ta yikh zastosuvannya v systemakh shtuchnoho intelektu: (dys. … kand. tekhn. nauk) National University of Radio Electronics]. Kharkiv [in Ukrainian].
9. Tereykovs’kyy, I. A., Bushuyev, D. A. & Tereykovs’ka, L. O. (2022). Artificial neural networks: basic concepts: educational manual [Shtuchni neyronni merezhi: bazovi polozhennya: navchal’nyy posibnyk]. Kyiv: National Technical University of Ukraine «Igor Sikorsky Kyiv Polytechnic Institute». Retrieved from https://ela.kpi.ua/server/api/core/bitstreams/9fee52b6-83fc-4e99-8541-c2767f634c7c/content [in Ukrainian].
10. Subbotin, S. O. (2020). Neural networks: theory and practice: educational manual [Neyronni merezhi: teoriya ta praktyka: navchal’nyy posibnyk]. Zhytomyr: Publishing house O. O. Yevenok [in Ukrainian].
11. Krivda, V. V., Yerisov, M. M. & Maksymenko, M. O. (2023, kviten’). Artificial intelligence in vehicle control systems [Shtuchnyy intelekt u systemakh keruvannya transportnymy zasobamy. Рaper presented at the mizhnarodnoyi konferentsiyi «Innovatsiyni tekhnolohiyi pidhotovky kadriv dlya promyslovosti ta transportu 2023»]. Dnipro, 98–104. Retrieved from https://okmm.nmu.org.ua/ua/2023/ITPTIT2023.pdf [in Ukrainian].
12. Dobrovs’ka, L. M. & Dobrovs’ka, I. A. (2015). Theory and practice of neural networks: educational manual [Teoriya ta praktyka neyronnykh merezh: navchal’nyy posibnyk]. Kyiv: Publishing house «Polytechnic». Retrieved from https://ela.kpi.ua/server/api/core/bitstreams/c4dbb86a-d617-44df-9b7a-5b6e9f91cebf/content [in Ukrainian].