基于因素空间的油气SCADA系统信息安全评价方法研究

发布时间：2018-04-24 02:25

本文选题：油气SCADA系统 + 安全评价　；参考：《西南石油大学》2016年硕士论文

【摘要】：目前油气SCADA系统正面临严峻的信息安全形势,而针对现有SCADA系统进行信息安全评价是重要应对措施。而现阶段国内外还没有一个系统的对油气SCADA系统信息安全进行评价的方法。本课题是在工控系统信息安全评价规范标准规范很不完善,评估对象比较模糊,评估方法也没有形成的时候,对油气SCADA系统信息安全进行评价的积极的尝试和探究。文章首先分析油气SCADA系统信息安全的具体含义,通过和信息技术系统的比较,并在系统和资产识别,区域划分等方面进行分析,解析油气SCADA系统信息安全的真正内涵；其次通过具有指导意义的工控系统信息安全标准、规范,明确了评价对象、了解评价要求、并进一步挖掘出评价指标(条件)因素。在此基础上引入事前、事中、事后的防御安全评价技术路线,将评价指标分别映射到定事前事中事后因素上,并制定油气SCADA系统信息安全事前事中事后评分表,并将评价指标因素用因果图的形式进行表示,可以清晰的展示影响系统信息安全评价的不同层次的条件因素；然后文章用因素空间理论对油气SCADA系统信息安全的评价进行理论建模、并对模型形式化描述,对于评价模型中因素神经元的评价方法,采用了传统模糊隶属度评价法,云模型评价法和基于因素空间的因素分析Mamdani推理评价法；在评价二级因素资源占用时采用云模型推理法进行评价,其它的由原始评分表中的因素值Fv向上层二级因素进行评价时采用隶属度评价法,当涉及到由定性到定性的评价(或由初始评分表的定量到一级因素即事前、事中、事后因素的定性评价)时采用因素分析法进行评价。文章编程实现了因素分析法(是实现大数据分析的非常有前途的一种方法),给定一张因素分析表作为输入可以得到关联规则,并将因素分析法关联出的规则用代码加入到模糊逻辑工具箱中,从而可以实现推理功能。然后用建立起来的因素分析法Mamdani推理模型和决策树算法在平均准确度和混淆矩阵两个方面进行比较,得到因素分析法的12次预测平均准确度为92.06%,决策树算法为91.45%。由混淆矩阵得出的四个综合指标也要优于决策树算法,比较说明基于因素空间的因素分析法是可以应用于系统评价预测的。最后文章以初始评分表为基础,分别采用三种方法,进行评价的例子,说明了文章提出的评价模型进行评价的一般过程。需要指出的是,文章提出的基于因素空间的因素分析法是一种比较新的评价方法(2014年相关文献出现),因此在多方面还有改进的地方。此外应该在本文基础上开发自动化的系统扫描评价工具,能够自动采集系统数据,能实现较智能的、客观的进行系统评价。
[Abstract]:At present, the oil and gas SCADA system is facing a severe information security situation, and it is an important measure to evaluate the information security of the existing SCADA system. At present, there is not a systematic method to evaluate the information security of oil and gas SCADA system at home and abroad. This subject is an active attempt and exploration to evaluate the information security of oil and gas SCADA system when the information security evaluation standard and specification of industrial control system is not perfect, the evaluation object is relatively fuzzy, and the evaluation method is not formed. This paper first analyzes the specific meaning of the information security of oil and gas SCADA system, through the comparison with the information technology system, and analyzes the real connotation of the information security of the oil and gas SCADA system in the aspects of system and asset identification, regional division and so on. Secondly, through the information security standard of industrial control system with guiding significance, this paper defines the evaluation object, understands the evaluation requirement, and further excavates the evaluation index (condition) factor. On this basis, the technical route of defense safety evaluation is introduced, and the evaluation indexes are mapped to the pre-incident and post-event factors respectively, and the score table of pre-incident and post-event of oil-gas SCADA system information security is worked out. And the evaluation index factors are expressed in the form of causality diagram, which can clearly show the different levels of conditional factors that affect the evaluation of system information security. Then the paper uses the factor space theory to model the information security evaluation of oil and gas SCADA system, and formalizes the model. For the evaluation method of factor neuron in the evaluation model, the traditional fuzzy membership degree evaluation method is adopted. Cloud model evaluation method and factor analysis based on factor space based Mamdani reasoning evaluation method, in the evaluation of secondary factor resource occupation, cloud model reasoning method is used to evaluate, Other factors from the original scoring table FV to the upper secondary factors are evaluated by membership evaluation method. When it comes to the evaluation from qualitative to qualitative (or from quantitative to first-order factors in the initial scoring table, that is, in advance), The qualitative evaluation of the factors after the event) was evaluated by the method of factor analysis. The method of factor analysis is a promising method to realize big data's analysis. Given a factor analysis table as input, the association rules can be obtained. The rules associated with factor analysis are added to the fuzzy logic toolbox by code, so that the reasoning function can be realized. Then the Mamdani inference model and decision tree algorithm are compared in average accuracy and confusion matrix. The average accuracy of 12 times prediction is 92.06 and the decision tree algorithm is 91.45. The four synthetic indexes derived from the confusion matrix are also superior to the decision tree algorithm. The comparison shows that the factor analysis method based on factor space can be applied to system evaluation and prediction. Finally, based on the initial scoring table, three methods are used to evaluate the model, and the general process of evaluation is illustrated. It should be pointed out that the factor analysis method based on factor space proposed in this paper is a relatively new evaluation method. In addition, an automatic system scan evaluation tool should be developed on the basis of this paper, which can collect system data automatically and realize more intelligent and objective system evaluation.
【学位授予单位】：西南石油大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TE48

【相似文献】