半监督环境下基于AE-ELM模型的工业网络安全防御研究

Research on Industrial Network Security Defense Based on AE-ELM Model in Semi Supervised Environment

摘要:

针对工业网络的安全性问题,在半监督环境的支持下,利用AE-ELM模型优化设计工业网络安全防御方法。根据不同网络攻击程序的攻击原理,设置工业网络入侵攻击检测标准。采集满足质量要求的工业网络实时数据,在半监督环境下,利用AE-ELM模型提取工业网络数据特征,通过与设置检测标准的匹配,得出工业网络攻击类型的检测结果。根据网络攻击类型,从边界、传输通道、终端等方面进行防御部署,实现工业网络的安全防御。通过与传统防御方法的对比测试得出结论：综合考虑两种攻击场景,在优化设计方法的防御作用下,工业网络的数据丢失率和篡改率分别降低约4.36%和2.35%,即优化设计方法具有更高的安全防御效果。

Abstract:

To address the security issues of industrial networks, with the support of a semi supervised environment, the AE-ELM model is used to optimize the design of industrial network security defense methods. Set industrial network intrusion detection standards based on the attack principles of different network attack programs. Collect real-time data of industrial networks that meet quality requirements, and extract industrial network data features using the AE-ELM model in a semi supervised environment. By matching with the set detection standards, the detection results of industrial network attack types are obtained. According to the type of network attack, defense deployment is carried out from the aspects of boundaries, transmission channels, terminals, etc. to achieve security defense of industrial networks. Through comparative testing with traditional defense methods, it is concluded that, considering the two attack scenarios comprehensively, under the defense effect of the optimized design method, the data loss rate and tamper rate of the industrial network are reduced by about 4.36% and 2.35%, respectively, indicating that the optimized design method has higher security defense effect.