Elsevier

Computer Communications

Volume 181, 1 January 2022, Pages 80-89
Computer Communications

Reliability evaluation of Markov cyber–physical system oriented to cognition of equipment operating status

https://doi.org/10.1016/j.comcom.2021.10.004Get rights and content

Abstract

With the rapid development of computing, communication, and control technologies, cyber–physical systems that integrate physical space, information space, and social space have emerged and are widely used in various important infrastructures. This article introduces the composition and basic algorithm of the hidden Markov model, and gives the mathematical description of the hidden Markov model. Since the hidden Markov model can deduce the hidden state of the observation object through the observed feature values, a device operating state cognition scheme based on the hidden Markov model is proposed. A method for analyzing cascading failures is proposed, and the critical threshold value of cyber–physical system under random attack is obtained. It is verified by simulation experiments, and the changes of system critical thresholds under different network parameters are compared and analyzed. We mainly use several sets of simulation experiments to verify the reliability of the critical threshold, and then verify near the critical threshold. Before simulating the cascading failure process, we first construct two random networks based on the average degree and the number of nodes. According to the previous description of the cyber–physical system model, a node in network B is randomly connected with three nodes in network A, so that the two networks are connected together to form a coupled system. Random attack or failure is represented by randomly deleting nodes. In the simulation experiment, we will simulate the process of cascading failure at each step, and after each step of cascading failure, we output and save the number of remaining nodes. When no nodes in the two networks are deleted, the cascading failure will stop, and then we will verify the critical threshold through the data obtained from the analysis. This provides the support of related theories and methods for the design of stable and reliable cyber–physical systems.

Introduction

Cyber–physical systems are systems that integrate the dynamic characteristics of physical processes with the software and communication technologies involved in these physical processes [1], [2], [3]. These systems demonstrate the connection between network information technology and the physical world. According to different application backgrounds, this type of system is also called process control system, monitoring control and data collection system or network control system [4], [5]. The application of cyber–physical system in real life determines its safety is very important. Once its safety has problems, it will cause huge irreversible damage to the corresponding controlled system [6]. For example, in the distributed management and control system in the power grid, water network, and natural gas network system, it is the core part in the health care, weapon system and transportation management system [7]. Once these cyber–physical systems have problems, they will have a major impact on public safety, health and even the national economy [8].

Condition-Based Monitoring (CBM) essentially provides a basis for fault diagnosis, performance evaluation and life prediction through real-time monitoring and processing of equipment operating conditions, so as to formulate corresponding maintenance strategies to prevent failures while avoiding unnecessary maintenance costs. CBM systems usually contain multiple functional modules, such as data acquisition, signal processing, feature extraction and selection, condition monitoring and health assessment, fault diagnosis, life prediction, and decision reasoning. The collection of real-time operating data such as vibration, pressure, and temperature reflecting the operating status of the equipment is completed by various sensors, but the data collected by the sensor not only reflects the useful information of the equipment status, but also includes transmission errors, measurement errors and environmental noise. Therefore, the data needs to be processed before the actual diagnosis and analysis. This processing method is also called signal processing. Various signal processing technologies can analyze and interpret the collected data, remove noise interference, and improve the signal to noise of the original data. Using these characteristics, the establishment of equipment failure models and performance prediction models can diagnose the operating status of the equipment in real time, evaluate its performance status and predict the remaining life of the equipment. Fault diagnosis and prediction are two important components of CBM: fault diagnosis realizes the fault detection of monitoring equipment, the determination of the fault location when the fault occurs, and the identification of the fault type; and the fault prediction determines the equipment based on the current monitoring data of the equipment.

This article introduces the basic concepts of Markov chain and HMM, and discusses the three basic problems of HMM: evaluation problem, decoding problem and learning problem. The basic algorithms to solve these three problems are introduced: Forward-Back algorithm, Viterbi algorithm and Baum–Welch algorithm. Specifically, the technical contributions of this article can be summarized as follows:

First: This article discusses the problems that may occur in the actual use of the HMM algorithm and provides solutions, and analyzes the reliability of the cyber–physical system under the random attack strategy.

Second: This article studies and analyzes the cascading failure process of three types of cyber–physical systems. Based on the theory of percolation in network science, the cascading failure process of the cyber–physical system under random attacks is systematically analyzed, and the iterative relationship of the cascading failure process is obtained, and the critical threshold of the cyber–physical system is obtained.

Third: We verified the calculated theoretical threshold and analyzed the relationship between the reliability of the cyber–physical system and network parameters.

The rest of this article is organized as follows. Section 2 discusses the theoretical basis of the reliability assessment of cyber–physical systems. Section 3 establishes a hidden Markov model for cognition of equipment operating status. Section 4 analyzes the experimental results. Section 5 summarizes the full text.

Section snippets

Related work

In the early development of cyber–physical systems, there was a lack of security protection considerations in the information domain [9]. With the widespread application of CPS, there are more and more industry fields for data circulation at all levels of the system, and the importance of CPS information security has become increasingly prominent. Unlike traditional IT systems, CPS emphasizes confidentiality and integrity in the security of the information domain [10]. The security of the

Reliability assessment method of equipment operating state

The reliability of equipment operating status is mainly to evaluate the sufficiency of equipment operating status, and quantitatively evaluate by calculating reliability indicators [21], [22]. According to the different evaluation objects, the reliability indexes of equipment operation status can be divided into load point indexes and system indexes. The load point index is used to characterize the reliability of a single load point to analyze the reliability of users in the system; the system

Hidden Markov model

The Markov process is a random process without aftereffect, in which the random process X(t) under the condition that the “now” state is known, the “future” state of the process is only related to the current state, but is related to the “past” state.

Given a random process X(t), t T, for any n, the probability value satisfies: P(Sn>X(tn))=X(tn1)Sn2

Then we call X(t) a Markov process. The random sequence S is called a Markov sequence, and the conditional distribution function P is often

Simulation verification

Since the cascading failure process for the three types of cyber–physical systems is exactly the same, as long as it is verified that the critical threshold of one type of cyber–physical system is correct, the other two must be correct, so we only choose one type for simulation experiment verification. Next, we mainly verify the first type of cyber–physical system.

We assume that the cyber–physical system is composed of two networks with different numbers of nodes. The two networks that make up

Conclusion

This paper counts the duration of the system state, calculates the load point and system reliability indicators. Aiming at the supporting role of the information system on the operation control of the equipment operating state, the topology structure of the information system and the overall structure of the communication network are introduced. Based on the function of the information system to transmit data information and control commands, a network connectivity model and information

CRediT authorship contribution statement

Qin Zhang: Performed the validation. Yutang Liu: Wrote the manuscript.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (28)

  • VenkataramananV. et al.

    CP-SAM: Cyber-physical security assessment metric for monitoring microgrid resiliency

    IEEE Trans. Smart Grid

    (2019)
  • GiftyR. et al.

    Weibull cumulative distribution based real-time response and performance capacity modeling of cyber–physical systems through software defined networking

    Comput. Commun.

    (2020)
  • GuoJ. et al.

    Reliability assessment of a cyber physical microgrid system in island mode

    CSEE J. Power Energy Syst.

    (2019)
  • SharmaA. et al.

    Service level agreement and energy cooperative cyber physical system for quickest healthcare services

    J. Intell. Fuzzy Systems

    (2019)
  • Cited by (8)

    • Model of the information security protection subsystem operation and method of optimization of its composition

      2022, Egyptian Informatics Journal
      Citation Excerpt :

      Such models are more informative, but are still based on information, some of which was collected as a result of uncontrolled experiments. Models [18–22] were created with a focus on the description of the fourth performance indicator based solely on the results of controlled experiments. Considering that the causes of failures are usually interrelated, models of this type are based on the mathematical apparatus of Markov chains.

    • Retroreflective technology of highway traffic retroreflective value based on data mining

      2023, Chang'an Daxue Xuebao (Ziran Kexue Ban)/Journal of Chang'an University (Natural Science Edition)
    • Review of Interruption Control Technologies for Cascading Faults in Power Systems

      2023, Proceedings - 2023 2nd Asian Conference on Frontiers of Power and Energy, ACFPE 2023
    • Equipment Asset Management and Equipment Health Based on Fuzzy Algorithm Evaluation Model

      2023, Journal of Combinatorial Mathematics and Combinatorial Computing
    • Framework Design and Key Technology Research of Digital Twin Substation

      2023, Gongcheng Kexue Yu Jishu/Advanced Engineering Sciences
    • Digitized Status Evaluation of Secondary Equipment in Traction Substation

      2023, Proceedings of SPIE - The International Society for Optical Engineering
    View all citing articles on Scopus
    View full text