Multi-agent system for monitoring objects of the energy complex

DOI: 10.34759/trd-2023-130-23

Аuthors

Kasatikov N. N.^*, Fadeeva A. D., Brekhov O. M.^**

Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia

*e-mail: Nick925@yandex.ru
**e-mail: obrekhov@mail.ru

Abstract

The purpose of the presented article consists in developing a multi-agent system capable of effective management and control of various objects of a power complex. The system will employ artificial intelligence (AI) and machine learning (ML) techniques to analyze data from sensors and other sources to detect and prevent potential problems, optimize energy consumption and improve overall efficiency. The multi-agent system will enable communication and cooperation between the various agents involved in the management of the energy complex, such as operators, engineers and maintenance personnel, as well. The final goal of the article consists in creation of the system, capable of reducing the dead time, enhancing productivity and increasing safety in the energy complex. The work is of a review character, about multi agent system of a «smart city» and its application at the energetics and space-rocket industry in the first place. The results of this work are proposed steps on the multi-agent system general development for its implementation at the power and space-rocket enterprises. The article describes how the «smart city» multi-agent system may be applied at the space-rocket enterprises and power objects in various ways. These systems may be helpful while controlling complex and interrelated processes, associated with space missions, form the spacecraft start and to its operation. Multi-agent systems may enable connection and coordination between different agent, such as ground-based control and satellite systems. They may as well be helpful in energy consumption optimizing and reduce the number of waste due to the energy consumption and storing control. Besides, these systems may help with the maintenance service, revealing potential problems prior they become critical, as well as increasing the space mission overall security and effectiveness. The multi-agent system application at the power objects and space-rocket enterprises may present several advantages such as efficiency, reliability and cost effectiveness enhancing. One inference that can be drawn consists in the fact that the multi-agent system would help optimizing energy production and consumption by coordination and control of the separate agents’ behavior. Each agent in the system may have its certain functions such as energy consumption monitoring, equipment operation control or decision making on the energy storage and distribution.

One more inference consists in the fact that the multi-agent system is able to increase the energetics objects reliability by ensuring redundancy and fault tolerance. In case of one of the agents’ failure, other agents will be able to keep on running and compensate the failure, reducing the risk of dead time and increasing the overall reliability of the system. Besides, the multi-agent system may increase economic efficiency by the energy losses reduction based on the real time data. The system may help as well to control the energy demand reducing peak loads and demand for the extra power for energy production, which will result in savings on costs for both energy suppliers and consumers.

As a whole, the multi-agent system employing on the space-rocket objects and power objects would lead to significant increase in energy efficiency, reliability and economic efficiency, which makes it an up-and-coming approach to control and optimization of power systems.

Keywords:

Internet of things, multi-agent system, energy, smart city, artificial intelligence

References

1. Baldi P. Autoencoders, Unsupervised Learning, and Deep Architectures, JMLR Proceedings of Machine Learning Research, 2012, vol. 27, pp. 37–49.
2. Wang Y., Yao H., Zhao S. Auto-encoder based dimensionality reduction, Neurocomputing, 2016, vol. 184, pp. 232–242. DOI: 10.1016/j.neucom.2015.08.104
3. Murphree J. Machine learning anomaly detection in large systems, IEEE Autotestcon, DOI: 10.1109/AUTEST.2016.7589589
4. Araya Daniel Berhane, Katarina Grolinger, Hany F. El Yamany, Miriam A. M. Capretz. An ensemble learning framework for anomaly detection in building energy consumption, Energy Buildings, 2017. DOI: 10.1016/j.enbuild.2017.02.058
5. Bouabdallaoui Y., Lafhaj Z., Yim P., Ducoulombier, L., Bennadji B. Natural Language Processing Model for Managing Maintenance Requests in Buildings, Buildings, 2020, vol. 10 (9), pp. 160. DOI: 10.3390/buildings10090160
6. Su Y., Zhao Y., Niu C., Liu R., Sun W., Pei D. Robust Anomaly Detection for Multivariate Time Series through Stochastic Recurrent Neural Network, In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Anchorage, AK, USA, 4–8 August 2019, vol. 1485, pp. 2828–2837. DOI: 10.1145/3292500.3330672
7. Fan C., Xiao F., Yan C. A framework for knowledge discovery in massive building automation data and its application in building diagnostics, Automation Construction, 2015, vol. 50, pp. 81-90. DOI: 10.1016/j.autcon.2014.12.006
8. Moreno M.V., Dufour L., Skarmeta A.F., Jara A.J., Genoud D., Ladevie B. et al.
   Big data: The key to energy efficiency in smart buildings, Soft Computing, 2015, vol. 20, pp. 1749–1762. DOI: 10.1007/s00500-015-1679-4
9. Minoli D., Sohraby K., Occhiogrosso B. IoT Considerations, Requirements, and Architectures for Smart Buildings — Energy Optimization and Next-Generation Building Management Systems, IEEE Internet Things Journal, 2017, vol. 4, pp. 269–283. DOI: 10.1109/JIOT.2017.2647881
10. Akkaya K., Guvenc I., Aygun R., Pala N., Kadri A. IoT-based occupancy monitoring techniques for energy-efficient smart buildings, In Proceedings of the 2015 IEEE Wireless Communications and Networking Conference Workshops (WCNCW), New Orleans, LA, USA, 9–12 March 2015, pp. 58–63. DOI: 10.1109/WCNCW.2015.7122529
11. Bourdeau M., Zhai, X., Nefzaoui E., Guo X., Chatellier P. Modeling and forecasting building energy consumption: A review of data-driven techniques, Sustain Cities and Socierety, 2019, vol. 48. DOI: 10.1016/j.scs.2019.101533
12. Gunay B., Shen W. Newsham G. Data analytics to improve building performance: A critical review, Automion in Construction, 2019, vol. 97, pp. 96-109. DOI: 10.1016/j.autcon.2018.10.020
13. Marmo R., Nicolella M., Polverino F., Tibaut A., Marmo R. A Methodology for a Performance Information Model to Support Facility Management, Sustainability, 2019, vol. 11, pp. 7007. DOI: 10.3390/su11247007
14. Mekki K., Bajic E., Chaxel F., Meyer F. A comparative study of LPWAN technologies for large-scale IoT deployment, ICT Express, 2019, vol. 5, pp. 1-7. DOI: 10.1016/j.icte.2017.12.005
15. Zheng A., Casari A. Feature Engineering for Machine Learning, O’Reilly Media, Inc.: Sebastopol, CA, USA, 2018, 218 p.
16. Miller C., Arjunan P., Kathirgamanathan A., Fu C., Roth J. et al. The ASHRAE Great Energy Predictor III competition: Overview and results, Science and Technology for Built Environment, 2020, vol. 26 (1), pp. 1427-1447. DOI: 10.1080/23744731.2020.1795514
17. Sagheer A., Kotb M. Unsupervised Pre-training of a Deep LSTM-based Stacked Autoencoder for Multivariate Time Series Forecasting Problems, Scientific Reports, 2019, no. 9, pp. 19038. DOI: 10.1038/s41598-019-55320-6
18. Semenov M.E., Solov’ev A.M., Popov M.A. Trudy MAI, 2017, no. 93. URL: http://trudymai.ru/eng/published.php?ID=80231
19. Sudarenko D.A., Lyutov A.V. Trudy MAI, 2020, no. 120. URL: https://trudymai.ru/eng/published.php?ID=161428. DOI: 10.34759/trd-2021-120-14
20. Brekhov O.M., Tin M.A. Trudy MAI, 2015, no. 84. URL: https://trudymai.ru/eng/published.php?ID=63151

Download