The application of computer network has mushroomed with the networking manufacture, networking instruments and etc. The ability of the production or system that can endures the breakage and the fault is a very important performance index. Especially today, the problems by reason of information security including hidden failures and defects in the system directly threaten the security and performance of system. This performance index to indicate these cases is the vulnerability. Networks vulnerability is defined as the abilities of network which to endure attacks, destroy and faults, also being called invulnerability of network. Vulnerability is often measured by the degree of vulnerability, which directly indicates the magnitude of invulnerability. This paper firstly introduces the conception of networks vulnerability, and distinct the analysis methods of networks vulnerability. The results of this paper are being looked forward to giving much direction on the system analysis and synthesis too.
In this paper, we mainly aim at D-S theory of evidence and the network intrusion detection these two fields. It discusses the method how to apply this probable reasoning as an AI technology to the Intrusion Detection System (IDS). This paper establishes the application model, describes the new mechanism of reasoning and decision-making and analyses how to implement the model based on the synscan activities detection on the network. The results suggest that if only rational probability values were assigned at the beginning, the engine can, according to the rules of evidence combination and hierarchical reasoning, compute the values of belief and finally inform the administrators of the qualities of the traced activities -- intrusions, normal activities or abnormal activities.
Stator Winding Bar Hollow Strand Blockage (SWBHSB) is one of the main faults for large turbo-generators with water and hydrogen cooling system. It will lead to increasing water temperature at the bar exit which may cause hidden troubles for turbo-generator's security. According to a three-layer-structural model of data fusion, this paper presents a fault diagnosis method for turbo-generators based on data fusion technology. Firstly, a bp network on pixel level fusion is set up, in which several temperature parameters at the bar exit are accurately computed. Then in feature level fusion, the fingerprints are distilled from the result of pixel level fusion. Finally, decision level fusion gives a fault diagnosis for the measuring channels and thermometric components. This method can effectively avoid problems such as misinformation and fake report.
This paper presents a precise and credible pressure calibration instrument. To improve the precision of the SF6 density measured, the gas density in the cabin is sampling, and then the density is calibrated and current density in the cabin and the real density value to be calibrated are displayed.
In this paper we present a modified target recognition data fusion module. Knowledge filter and knowledge reconstruction based on Rough sets optimize the knowledge acquisition, simplify the following inference machine's structure and reduce the complexity of calculation. The blackboard frame makes the mutual expert systems cooperate well. The method had been used in fault diagnosis of water turbogenerator in fire-power-plant and wined the satisfactory effect.