The increasing complexity and frequency of malware attacks pose significant challenges to cybersecurity, as traditional methods struggle to keep pace with the evolving threat landscape. Current malware classification techniques often fail to account for the heterogeneity of malware data and models across different clients, limiting their effectiveness. In this chapter, we propose a distributed model enhancement-based malware classification method that leverages federated learning to address these limitations. Our approach employs generative adversarial networks to generate synthetic malware data, transforming non-independent datasets into approximately independent ones to mitigate data heterogeneity. Additionally, we utilize knowledge distillation to facilitate the transfer of knowledge between client-specific models and a global classification model, promoting effective collaboration among diverse systems. Inspired by active defense theory, our method identifies suboptimal models during training and replaces them on a central server, ensuring all clients operate with optimal classification capabilities. We conducted extensive experimentation on the Malimg dataset and the Microsoft Malware Classification Challenge (MMCC) dataset. In scenarios characterized by both model heterogeneity and data heterogeneity, our proposed method demonstrated its effectiveness by improving the global malware classification model’s accuracy to 96.80%. Overall, our research presents a robust framework for improving malware classification while maintaining data privacy across distributed environments, highlighting its potential to strengthen cybersecurity defenses against increasingly sophisticated malware threats.

In next-generation wireless networks, distributed clients collaborate to achieve data perception, knowledge discovery, and model reasoning. Generally, Federated Contrastive Learning (FCL) represents an emerging approach for learning from decentralized unlabeled data while upholding data privacy. In FCL, participant clients collaborate in learning a global encoder using unlabeled data, which can serve as a versatile feature extractor for diverse downstream tasks. Nonetheless, FCL is susceptible to local data leakage risks, such as membership information leakage, stemming from its distributed nature, an aspect often overlooked in current solutions. This study delves into the feasibility of executing a membership information leakage on FCL and proposes a robust membership inference methodology. Our objective is to determine if the data signifies training member data by accessing the model’s inference output. Specifically, we concentrate on attackers situated within a client framework, lacking the capability to manipulate server-side aggregation methods or discern the training status of other clients. We introduce two membership inference attacks tailored for FCL: the passive membership inference attack and the active membership inference attack, contingent on the attacker’s involvement in local model training. Experimental findings across diverse datasets validate the effectiveness of our method and underscore the inherent local data risks associated with the FCL paradigm.

The increasing use of energy conversion facilities has highlighted the interdependence between energy systems. Current literature has paid enormous attention to integrated heat-power networks and particularly focused on a holistic optimization framework. However, distribution networks and district heating networks belong to different entities and hold varied interests. In this paper, we propose a market model of integrated heat-power systems and a distributed algorithm for solving the equilibrium point. In the market model, energy trading between the heat market and the power market are taken into account, as well as the response of heterogeneous energy consumers to price signals from the markets. Besides, electricity and heat consumers are separately modeled based on corresponding features to make the modelling more accurate. In the solution method, a distributed algorithm based on fixed-point iteration is implemented after the model reformulation, to seek the market equilibrium with sensitive information of both networks well preserved. Numerical experiments are conducted on a testing system to validate the proposed model and the solution method.

Federated Contrastive Learning (FCL) represents a burgeoning approach for learning from decentralized unlabeled data while upholding data privacy. In FCL, participant clients collaborate in learning a global encoder using unlabeled data, which can serve as a versatile feature extractor for diverse downstream tasks. Nonetheless, FCL is susceptible to privacy risks, such as membership information leakage, stemming from its distributed nature, an aspect often overlooked in current solutions. This study delves into the feasibility of executing a membership inference attack on FCL and proposes a robust attack methodology. The attacker's objective is to determine if the data signifies training member data by accessing the model's inference output. Specifically, we concentrate on attackers situated within a client framework, lacking the capability to manipulate server-side aggregation methods or discern the training status of other clients. We introduce two membership inference attacks tailored for FCL: the passive membership inference attack and the active membership inference attack, contingent on the attacker's involvement in local model training. Experimental findings across diverse datasets validate the effectiveness of our attacks and underscore the inherent privacy risks associated with the federated contrastive learning paradigm.

电力专网与公用互联网的互动可深度挖掘虚拟电厂的灵活性潜力,然而虚拟电厂在提供灵活性服务时易受到公网侧的网络攻击致使系统出现灵活性缺额,进而影响新型电力系统的运行安全.针对虚拟电厂遭受协同网络攻击,首先根据虚拟电厂的运行机制和通信网络架构,给出了公网侧潜在的网络攻击途径;其次,建立了电动汽车的混合逻辑动态模型来推演其后续时间断面的灵活性调节容量;然后,建立了虚假数据注入和拒绝服务两种典型网络攻击的事件驱动模型,在此基础上提出了基于事件驱动-融合流模型的电网安全状态动态推演方法;最后,对改进的IEEE 30节点系统进行了仿真计算,结果表明所提方法能够刻画跨空间故障的传播路径并量化协同网络攻击对电网的危害,为电网运行人员开展动态风险评估与信息安全防护奠定基础.

False data injection attacks (FDIA) destroy the integrity of information transmission by evading the bad data detection mechanism, and thus affects the stability of power cyber-physical systems (PCPS). Existing studies simply introduce complex neural network models for FDIA detection, ignoring spatial-temporal correlation and interpretability of neural networks. As a result, the accuracy and reliability of false detection may be negatively affected. To address the challenges above, this paper proposes an interpretable deep learning framework based on the spatial-temporal attention mechanism. Firstly, based on the gated recurrent unit (GRU), a dual attention mechanism is designed by combining spatial and temporal features of deep neural network to dynamically mine the potential correlations between the FDIA detection and the input features. Besides, the quantification of attention weights is introduced to interpret the spatial-temporal correlations between normal and attack data, which can effectively enhance the interpretability and reliability of detection results. Finally, based on the IEEE 14-bus test system and real operation data, simulations are conducted and the results show that the proposed STAGN model can detect FDIA effectively, has higher accuracy and stability than the latest detection models, and also has reasonable interpretability.

Renewable energy, such as wind power, has witnessed rapid development recently on account of increasing public awareness of environmental protection along with sustainable power supply. In this background, wind speed prediction plays an important role in helping ensure stable and economic operation of modern power systems with a high renewable penetration. However, the stricter data regulation rules and competitive demands in power markets set a limit to employing datasets from different wind farms to train a more accurate prediction model via machine learning techniques. Therefore, a privacy-preserving wind speed prediction framework based on federated deep learning is proposed to tackle the challenge, which utilizes multi-input inner-product functional encryption to offer extra data protection. A case study on the Wind Integration National Dataset proves that our method excels isolated training scenes and approaches the ideal centralized one in prediction performance without leakage of sensitive data.

为实现拓扑数据缺失及不准确情况下的低压配电网理论线损计算,提出基于虚拟用户拓扑等值法的台区理论线损计算方法.首先构建台区单一虚拟用户等值网络;然后基于台区终端数据和用户曲线数据进行等值电阻计算,并将每种运行方式类型下的等值电阻计算结果进行K-means聚类分析,形成各类运行方式下的等值电阻典型参数库;在此基础上,根据待计算日的运行方式与典型参数对应运行方式的相似程度选择等值电阻参数,分固定线损和可变线损进行理论线损计算.算例结果表明,该方法能够在台区拓扑数据缺失及不准确的情况下提供较准确的理论线损计算结果,满足实际线损排查和治理的需求.

实现电力大数据的有效共享是充分利用数据的基础环节,但这一过程中存在大量安全隐患.针对传统大数据共享体系中存在的数据稳定性、安全性与通信可靠性等问题,提出基于区块链技术的电力数据共享方案.通过建立可靠的点对点网络拓扑结构,区块链技术可以有效提升数据共享环节的可信性与安全性.在总体框架中,区块链作为数据共享体系的核心,保障了不同数据中心与终端之间的可信通信.同时,通过将区块链通信过程分解为多个层次,提升了数据共享机制的可扩展性与易用程度.

To investigate the therapeutic effects of PN on intestinal inflammation and microvascular injury and its mechanisms, dextran sodium sulfate- (DSS-) or iodoacetamide- (IA-) induced rat colitis models were used. After colitis model was established, PN was orally administered for 7 days at daily dosage of 1.0 g/kg. Obvious colonic inflammation and mucosal injuries and microvessels were observed in DSS- and IA-induced colitis groups. DAI scores, serum concentrations of VEGFA121, VEGFA165, VEGFA165/VEGFA121, IL-6, and TNF-α, and expression of Rap1GAP and TSP1 proteins in the colon were significantly higher while serum concentrations of IL-4 and IL-10 and MVD in colon were significantly lower in the colitis model groups than in the normal control group. PN promoted repair of colonic mucosal injury and microvessels, attenuated inflammation, and decreased DAI scores in rats with colitis. PN also decreased the serum concentrations of VEGFA121, VEGFA165, VEGFA165/VEGFA121, IL-6, and TNF-α and increased the serum concentrations of IL-4 and IL-10, with the expression of Rap1GAP and TSP1 proteins in colonic mucosa being downregulated. The constituents of PN were identified with HPLC-DAD. To sum up, PN could promote repair of injuries of colonic mucosa and microvessels via downregulating VEGFA isoforms and inhibiting Rap1GAP/TSP1 signaling pathway.

AIM To investigate the effects of Panax notoginseng (PN) on microvascular injury in colitis, its mechanisms, initial administration time and dosage. METHODS Dextran sodium sulfate (DSS)- or iodoacetamide (IA)-induced rat colitis models were used to evaluate and investigate the effects of ethanol extract of PN on microvascular injuries and their related mechanisms. PN administration was initiated at 3 and 7 d after the model was established at doses of 0.5, 1.0 and 2.0 g/kg for 7 d. The severity of colitis was evaluated by disease activity index (DAI). The pathological lesions were observed under a microscope. Microvessel density (MVD) was evaluated by immunohistochemistry. Vascular permeability was evaluated using the Evans blue method. The serum concentrations of cytokines, including vascular endothelial growth factor (VEGF)A121, VEGFA165, interleukin (IL)-4, IL-6, IL-10 and tumor necrosis factor (TNF)-α, were detected by enzyme-linked immunosorbent assay. Myeloperoxidase (MPO) and superoxide dismutase (SOD) were measured to evaluate the level of oxidative stress. Expression of hypoxia-inducible factor (HIF)-1α protein was detected by western blotting. RESULTS Obvious colonic inflammation and injuries of mucosa and microvessels were observed in DSS- and IA-induced colitis groups. DAI scores, serum concentrations of VEGFA121, VEGFA165, VEGFA165/VEGFA121, IL-6 and TNF-α, and concentrations of MPO and HIF-1α in the colon were significantly higher while serum concentrations of IL-4 and IL-10 and MVD in colon were significantly lower in the colitis model groups than in the normal control group. PN promoted repair of injuries of colonic mucosa and microvessels, attenuated inflammation, and decreased DAI scores in rats with colitis. PN also decreased the serum concentrations of VEGFA121, VEGFA165, VEGFA165/VEGFA121, IL-6 and TNF-α, and concentrations of MPO and HIF-1α in the colon, and increased the serum concentrations of IL-4 and IL-10 as well as the concentration of SOD in the colon. The efficacy of PN was dosage dependent. In addition, DAI scores in the group administered PN on day 3 were significantly lower than in the group administered PN on day 7. CONCLUSION PN repairs vascular injury in experimental colitis via attenuating inflammation and oxidative stress in the colonic mucosa. Efficacy is related to initial administration time and dose.

This paper compares the major technical differences in intelligent substation secondary system among several standards of State Grid Corporation of China. And base on the demonstration projects and investigation results,the key techniques of secondary equipment integration are analyzed and expounded. Thereby,a scheme is proposed for the holistically-lowering of indoor substation bay level in intelligent substation.

A simple,flexible and universal solution using the Eclipse modeling framework(EMF) for developing systems based on the common information model(CIM) is proposed.The scheme contains three aspects.Firstly,the Java code of package,class and object relations defined in CIM are automatically generated.Secondly,the CIM extendable markup language(XML) file can be efficiently scanned through streams,objects are formed in the memory during the scanning process,the relations between objects are constructed after the scan is finished.The differences of different energy management system(EMS) suppliers and different CIM versions are shielded,so good compatibility is reached.Thirdly,the constraints of objects defined in CIM are validated.The validity of the solution is testified by test results of CIM data of several real systems.

Based on the overall framework of intelligent dispatching system for Shanghai metropolitan power grid,this paper expounds the application practices at the first construction stage,emphatically in the field of grid security prevention and control,coordinated dispatching management,new energy accessing management,and technical application and development.

The advanced energy management system(AEMS) is a new-type automatic system for performing multi-objective,near-optimal and closed-loop control of electric power grids.The data-sharing platform(DSP) is the fundamental platform of AEMS that provides data sharing service.The structure of DSP is proposed,and distribution memory cached technique is described.This technique uses string "key-value" mapping to store shared data in distribution memories,which makes things simple,open and efficient.Finally,serialization techniques for the common information model(CIM) data and results of state estimation are proposed.CIM data is compressed before transfer to network which makes data interchange more efficient.The effectiveness of the techniques of data sharing proposed is proved by test results and application.

In combination with the status quo of EMS real-time information release system of Shanghai grid dispatching organizations,this paper analyzes the necessity of constructing panorama real-time running information system for Shanghai power grid.The system standards and construction scheme are also elaborated.

The reactive power optimal control is a multi-objective optimization problem with discrete variables in nature. The conventional control schemes are usually “time-based” which lack the flexibility for online control. Based on “event-driven” strategy, a hybrid control approach is proposed. The principle of hybrid control, system design ideas and implementation methods are introduced. The application results in Songjiang District of Shanghai City show that the proposed system can effectively eliminate all kinds of events to achieve near-optimal system state while ensuring the reliability and quality of distribution power supply and coordinating with higher-level control center.

Traditional graphical visualization of the power system is with several deficiencies, such as heavy maintenance workload, difficulties in exchange of graphical information. The idea of automatic generation of diagrams characterized with automatic maintenance, full coupling of diagrams and models, and without dependence on graphics tools is put forward. Based on common information model (CIM), the improved method of self-generated topology diagrams including substation and single-line diagrams is proposed. Dynamic diagrams which visualize current state and monitoring indicators of the power system are introduced. Three-dimensional visualization technology is used to draw those diagrams. Then a platform of early-warning and surveillance for multiple indicators is formed. The platform has been put into real operation in the actual grid.

The uniform and operational function specification of batch remote control based on different supervisory control and data acquisition(SCADA)systems is described.The realization method in actual engineering is proposed.The proposed function is applied in Shanghai Power Grid,which shows that the function can extend the application of SCADA system and increase the level of SCADA system practicability.It provides effective technical measures for dispatchers to cut load in case of over supplying and emergent load curtailing.It can also improve the efficiency of switching load between different dispatching regions.As an expanding SCADA system application,the proposed function is worth spreading.

利用配电网络具有树状层次结构的特点,提出了一种逐层筛选寻优的配电网无功优化算法。文中将辐射网逐层分解为随电压线性变化的等值负荷,并推导了等值负荷参数递推计算的方法;对含电容器、变压器分接头的控制策略空间也进行逐层分解,将每一个子策略映射到一组等值负荷参数,通过淘汰每一层无功电压负荷参数相似而局部有功网损较大的子策略,有效压缩了策略筛选空间,实现了自底向上的逐层筛选寻优。通过选取合适的参数,能保证算法在实现快速计算的同时具备全局寻优能力。另外,该算法优化结果具有确定性,不存在初值选取和收敛性问题。文中给出了实际配电网18节点系统和IEEE69节点系统的算例仿真,并与传统遗传算法在计算速度和优化效果方面进行了对比,说明该算法适用于含离散变量的大规模辐射网无功优化问题的在线计算。