Research On Neural Network Model Of SCADA Security Defense Analytic Factor

With the increasing development of industrial automation, automation and information technology is closely integrated, and an increasing number of information technologies are widely applied in the field of industrial control. The integration of automation and information technology brings enormous economic benefits to enterprises, while the safety of the industrial control system is facing a serious threat. Especially the security of oil gas gathering pipeline is particularly important which is taken as national economic artery. SCADA (Supervisory Control and Data Acquisition) system which is data acquisition and monitoring system as an important part of the industrial control system is widely used in the oil gas gathering field. Due to frequent security incidents happening in recent years, many SCADA systems rely on upgrading and rebuilding to avoid danger attack which results in unnecessary waste of production resources and also makes enterprises pay huge cost.The research content is derived from the National Natural Science Foundation of Project based on FNN (factor neural network) large-scale oil and gas gathering and transferring SCADA security defense system modeling theory and simulation method research (serial number:61175122). Belonging to the sub topic of the fund, this research mainly studies the model of system relationship and formal methods of modeling theory. Combined with the factor neural network theory, the author puts forward a SCADA security defense system model (Factor Neural Network Based SCADA Security And Defense Model, shorted for FSDM). The Petri net is applied to get a formal description of the model and the fuzzy Petri net theory is used to gain a formal inference on factors of program behavior. The global Petri net formal model is described to represent the relationship between the various factor neurons in SCADA security defense system, and then stateflow is used to simulate the neuron state transition in the system.The relationship of the factor neural network is described, and the relationship and distribution of each factor in the system are clarified, and the security defense prototype system of SCADA system is designed according to the C/S architecture. The distributed structure is adopted to meet the requirements of the structure of SCADA system. And defending client-side and server-side of SCADA system are respectively designed. Finally, data is sent to the client-side via server-side to test the prototype system communication, which contributes to build the initial network model for the development of SCADA security defense system and to describe the interaction between the implementation neurons and other non implementation neurons in SCADA security defense system.Based on the theory of factor space, SCADA security defense analytical factor neural network is constructed, and it lays a certain theoretical basis for the SCADA security defense system modeling and for the network formal description of the relationship between neurons in the factors. A formal reasoning method based on fuzzy Petri net is proposed and the initial interactive relationship model of SCADA security defense system is realized by adopting different platforms to develop the prototype system.