Skip to the main content

Original scientific paper

https://doi.org/10.17559/TV-20240524001658

Anomalous Network Traffic Detection: An Application of Deep Learning Algorithms

Liming Lin ; State Grid Information & Telecommunication Group Co., Ltd., Beijing, 102200, China *
Ang Xia ; State Grid Information & Telecommunication Co., Ltd., Beijing, 100032, China
Fangfang Dang ; State Grid Henan Information & Telecommunication Company (Data Center), Zhengzhou, Henan, 450000, China
Qin Yin ; State Grid Information & Telecommunication Group Co., Ltd., Beijing, 102200, China
Xirong Lv ; Xiamen Great Power Geo Information Technology Company Ltd., Xiamen, Fujian, 361000, China

* Corresponding author.


Full text: english pdf 601 Kb

page 1238-1248

downloads: 575

cite


Abstract

With the development of the electric power system, the safety of electric power industrial control network is increasingly prominent. Abnormal traffic detection is one of the important tasks in network security. To detect anomalous traffic quickly and accurately, the sine-cosine function and Levy flight are introduced into the locust algorithm to optimize its reduction factors. The improved locust algorithm is used to optimize the parameters of the anomalous network traffic model built on the basis of long short-term memory networks. Meanwhile, the optimized model is applied to the electric power industrial control system. The results showed that the accuracy of both the test and training sets was above 90%, proving that the designed model achieved good generalization capability and robustness. In the Dos attack type, the detection rate of the designed model reaches 98.15%, which is about 6% and 10% higher than that of other algorithms, which proves its high accuracy. These results prove the high efficiency and accuracy of the designed model in detecting network traffic attacks. The above results prove that the designed model can effectively identify anomalous flow and contribute to maintaining the safety of the power grid.

Keywords

abnormal network traffic detection; grasshopper optimization algorithm; levy flight; long short-term memory networks; reduction factor

Hrčak ID:

332813

URI

https://hrcak.srce.hr/332813

Publication date:

29.6.2025.

Visits: 946 *