System deployment
The key to the deployment of edge computing lies in how edge nodes obtain data, what data needs to be obtained, and how to use smart technologies to meet the business needs of all links of the power grid. The power system edge computing data communication framework is shown in Figure 4. As can be seen from the figure, under the original communication framework and standard of the power system, the source- side data can be exported from the D5000 platform of the dispatch center to the data collected by the pMu at the outlet of the power plant. Deploy edge computing servers in power plants, especially for more complex new energy models for power model parameters, using deep learning, reinforcement learning and other intelligent algorithms for online identification and real-time equivalent parameter modeling. The grid-side data and the charge-side data are collected by the
intelligent terminal and merger unit in the power grid, and are uploaded to the substation comprehensive system in the form of messages through optical fiber, wireless transmission, etc. Since the edge computing node is arranged in the substation, it can directly use communication to send the message to the edge computing platform, which is analyzed and stored by the edge computing platform. Edge computing can use the data after MMS (Manufacturing message specification) message and GOOSE (Generic object oriented substation events) message parsing to perform substation station domain parallel simulation, Infer and infer the future working conditions of the substation. We can also use the load-side data and user- side comprehensive energy data to generate a load model for the subordinate power grid of the substation and evaluate its load regulation capability. By parsing SV (Sample value) and GOOSE messages containing station domain data, using deep learning, reinforcement learning and other intelligent algorithms for fault feature extraction, stability discrimination, control strategy formulation, the station domain of the substation can be protected Control. Eventually, the dispatch center and the edge computing layer will carry out cloud-side collaborative communication, the dispatch center delivers simulation data, station domain analysis tasks, algorithm models, and the edge computing layer uploads station domain analysis results, power generation model parameters, load models, and comprehensive energy information.
V. Conclusion
To solve the problems mentioned in summary, we propose a solution to edge intelligence. This method uses advanced ideas such as attention mechanism, graph neural network, and topology network to be able to find hidden safety hazards in time and give alarm prompts. We conducted experimental verification on a real dataset, and the results show that the image recognition accuracy of our method reaches 84.60%, the stability of the system reaches 80.52%, and the failure rate is only 18.13%, which are better than other methods. in addition, we also verified the effectiveness of our proposed method in the actual system.
Acknowledgements
This work is supported by the Project of State Grid East china Power Grid Shanghai Municipal Electric Power company (520940180029).
References
H. Lu, Y. Li, M. Chen, H. Kim, and S. Serikawa, “Brain intelligence: go beyond artificial intelligence,” Mobile Networks and Applications, vol. 23, no. 2, pp. 368-375, 2018.
Q. Han, Q. Meikang, L. Zhihui, and G. Memmi, “An efficient key distribution system for data fusion in v2x heterogeneous networks,” Information Fusion, vol. 50, pp. 212-220, 2019.
K.-C. Chou, “The ploc_bal-mgpos is a powerful artificial intelligence tool for predicting the subcellular localization of gram-positive bacterial proteins according to their sequence information alone,” Glo J of Com Sci and Infor Tec, vol. 2, pp. 1-13, 2020.
S. Wang, X. Zhang, Y. Zhang, L. Wang, J. Yang, and W. Wang, “A survey on mobile edge networks: Convergence of computing, caching and communications,” IEEE Access, vol. 5, pp. 6757-6779, 2017.
H. Qiu, M. Qiu, and R. Lu, “Secure v2x communication network based on intelligent pki and edge computing,” IEEE Network, 2019.
A. Nikoobakht, M. Mardaneh, J. Aghaei, V. Guerrero-Mestre, and J. Contreras, “Flexible power system operation accommodating uncertain wind power generation using transmission topology control: an improved linearised ac scuc model,” IET Generation, Transmission & Distribution, vol. 11, no. 1, pp. 142-153, 2017.
E. Li, Z. Zhou, and X. Chen, “Edge intelligence: On-demand deep learning model co-inference with device-edge synergy,” in Proceedings of the 2018 Workshop on Mobile Edge Communications, 2018, pp. 3136.
F. WANG, “Shadow systems: a new concept for nested and embedded cosimulation for intelligent systems,” Tucson, Arizona State, USA:
Do'stlaringiz bilan baham: |