Research On Intrusion Prevention Technology Of Industrial Control Network Based On Deep Learning

"Industry 4.0" is an era where information technology is used to promote industrial change.Meanwhile,with the introduction of the "Internet +" concept,it has triggered a major integration and transformation of information technology and industry.Therefore,IT technology is applied to industrial control systems(ICS:Industrial Control System),making it from a proprietary closed system to a highly open and interconnected system.With the convenience of work and the improvement of production efficiency,the networked development of industrial control systems has led to an increase in system security risks and intrusion threats.Due to the weak ability of network security protection in industrial control system network protocol,as the number of access devices increases,network intrusions are increasing.However,the industrial control network intrusion detection is still in an initial stage.Aiming at these problems,this article focuses on the industrial control network intrusion prevention technology based on deep learning for research.First,the architecture and characteristics of the industrial control network are analyzed in detail,and the principles of industrial control virus intrusion,common access routes and attack methods are explored.Understand the current status of intrusion detection in industrial control networks and analyze its existing deficiencies.On this basis,aiming at the problem that the low-layer selection feature classification leads to low detection accuracy,an industrial control network intrusion detection model based on AE-ELM is proposed.The proposed algorithm uses the deep learning method to extract the features of the industrial control network data and then effectively classify the extracted features.The experiment also verifies the theoretical judgment and proves that the model also improves the detection rate.Furthermore,aiming at the problem of low accuracy of multi-classification for attack types,an intrusion detection algorithm based on multi-scale convolution is proposed,which converts intrusion detection data into "image data",constructs a convolutional neural network model,performs feature extraction on data convolution and pooling,and then The softmax classification is used to prove the feasibility of the algorithm through simulation experiments.Finally,an intrusion prevention system for industrial control networks isdesigned and implemented.The intrusion detection algorithms based on deep learning are proposed in this paper are used as intrusion detection modules,and the data acquisition module,system log module,defense response module,and central control module forms a complete set of network behavior monitoring and anomaly detection systems with complete functions and superior performance.