基于函数挖掘的能源信息物理系统数据安全风险识别算法

Data Security Risk Recognition Algorithm for Energy Cyber Physics System Based on Function Mining

数据安全风险评估对于能源信息物理系统安全稳定运行至关重要。现有的从二次设备、信息等角度来分析数据安全风险已经无法满足能源信息物理系统广泛的能源接入和各能源之间的能量、信息交互需求。首先提出基于粗糙集的数据安全风险要素特征选择算法,对影响能源信息物理系统中数据的安全风险特征集进行特征选择,降低能源信息物理系统数据安全风险要素集的维度;在此基础上,利用基因表达式编程(gene expression programming,GEP)的函数挖掘特性,提出基于混合GEP的能源信息物理系统数据安全风险识别算法,通过设计小生境种群生成策略以及动态自适应变异概率动态调整策略来提高数据安全风险识别的准确率和效率。仿真实验结果表明,所提算法对于复杂高维的能源信息物理系统数据安全风险集的识别和预测具有较高的准确率和较强的实用性,可为下一步制定能源信息物理系统数据安全防护策略提供理论方法支撑。

Data security risk assessment is essential for the safe and stable operation of energy cyber physics system(CPS).The existing data security risk analysis from the perspective of secondary equipment and information cannot meet the requirements for extensive energy access as well as energy and information interaction between various energy sources in the energy CPS.Firstly,a feature selection algorithm for data security risk elements based on rough set(FSDSRF-RS)is proposed to select the data security risk feature sets in the energy CPS,consequently reducing the dimensions of the data security risk element sets of the energy CPS.And then,a data security risk recognition algorithm for energy cyber physics system based on hybrid gene expression programming(DSRR-HGEP)is proposed.In the DSRR-HGEP,a niche-based population generation strategy and a dynamic adaptive mutation probability adjustment strategy are designed to improve the accuracy and efficiency of data security risk identification.Simulation and experimental results show that the proposed algorithm in this paper has a high recognition and prediction accuracy for the complex and high-dimensional data security risk sets in the energy CPS,and can provide a theoretical support for formulating data security protection strategies of the energy cyber physical system in the future.