Fault propagation identification is an indispensable task in complex system safety analysis. With the growing of system scale and complexity, it is hard for the traditional safety analysis techniques, which depend mainly on analysts' personal skills and experiences, to keep completeness and timeliness; moreover, some failure modes may be neglected and failure effects misjudged during the analysis. Formal science provides a new way to solve this problem, where formal verification method such as model checking can automatically validate whether the system design satisfies the given safety requirements, which can reduce an analysts' repetitive work and design cost, and improve the efficiency and quality of safety analysis. However, there is lack of a deliberate and reasonable way to build system models because of the diversity and flexibility of languages used for model checking, which results in that it is difficult to specify and model system quickly and accurately, and leads to some deviation in model checking. In this paper, a system modeling and safety property specifying approach using symbolic language SMV is proposed, including guidance on the mapping relationships between the formal language elements and system functions, architecture and failure modes; moreover, how to define system specifications and safety requirements using temporal logic formulas is discussed as well. Finally, a case study about airborne system safety analysis is provided, in which the counter-examples that do not meet system specifications can be identified automatically using model checker NuSMV to find out fault events and their propagation that can result in accidents.
System safety is analyzed from inherent network topology structure of integrated complex system. This paper discusses the influences of system inherent topology structure and the statistical property on fault propagation. Fault propagation model is established and the fault propagation intensity is defined as the weight of the edge among nodes considering the network statistical information. The highest risk path of fault propagation and critical nodes are got through qualitative fault propagation analysis. Finally an application example is illustrated to indicate the practicability and validity of the proposed method, which can be used as important basis for safety assessment and fault prevention.
Many real-world systems can be abstracted into network systems.They have made a great contribution to human daily life, however, risk and disadvantages of these network systems are also serious since a tiny fault may lead to a big disaster.So the ability of resilience that a system can response to an adverse disruption and recovery back to the normal condition after disruptions is needed for modern systems.This paper mainly proposed a resilience analysis framework based on the resilience engineering concept and numerical simulations are put forward based on a generated scale-free network.The simulation results reveal that factors of component reliability, failure propagation failure detection, and recovery strategy would indeed contribute to the resilience of network systems.
Many real-world systems such as multi-processor computing and power transmission systems have load sharing attribute where the whole system workload is distributed to different system elements. In these systems, when one element fails, other remaining elements would take over the workload which will change their failure behavior and cause a dynamic dependence of failure, thus a reliability evaluation work will be necessary for a high mission success probability of these systems. In this paper, the reliability evaluation problem of Linear Transformer Driver (LTD) system with typical attribute of load sharing and failure dependence is discussed. Firstly, the LTD system with reliability concerns is introduced. Then, the load sharing and failure dependence properties are specified with mathematic formulations. Finally, the key factors of initial work coefficient, load distribution strategy and failure model in the LTD development stage are analyzed from the reliability perspective. As a preliminary exploration, the result would provide a guidance for better industrial application.
Taking the features of complex system into account, the safety of complex system is researched from fault hazard, human error, dangerous characteristic and fatal environment based on entropy and dissipative structure theory in this paper. The concept of system safety entropy is proposed and the paper establishes the safety entropy model, which can be used in safety assessment for complex system.
Quantitative analysis of the fire system efficacy of is the trend of China's fire equipment supervision and management. One of the important work is to analyze the effects of failure on fire equipment, and formulate corresponding countermeasures of maintenance and management. Utilizing function analysis system technique (FAST), the mission levels of the fire equipment in building were divided, the analyzing process for fire equipment failure was established according to the failure information during the actual operation of fire equipments, and on the level of component unit, subsystem function and system mission, the method of ranking the effective factors of fire equipment failure was proposed based on the mission and function analysis. The method divided the effects of failure into 3 classes: minor (failure class III), general (failure class II), significant (failure class I), and corresponding management measures were proposed for each class. The method was used in a case study.
Integrated Modular Avionics (IMA), which is the novel concept of avionics architecture, can effectively improve the efficiency of system while reducing resource allocation. However, it also brings new types of risk such as fault propagation since the flexible mapping between logical configuration and physics configuration. In some way, IMA could be regarded as a network where interdependency relationship between function and resource denotes the edge, function and resource denotes the node. Vulnerability and criticality of nodes are characterized by scheduling frequency and degree, for simplicity, redundancy is not considered. As a safety-critical system, IMA structure itself provides a potential route for fault propagation with the need of flexible resource scheduling. Effective modeling and risk assessment would help for the development bargaining or reconfigure strategy implementing. This paper provides a formal specification and risk assessment approach for integrated complex system and a case study in IMA domain is illustrated.
Most traditional engineered systems are designed with a passive and fixed reliability capability and just required to achieve a possibly low level of failure occurrence. However, as the complexity at spatialtemporal scales and integrations increases, modern complex engineered systems (CESs) are facing new challenges of inherent risk and bottleneck for a successful and safe operation through the system life cycle when potential expected or unexpected disruptive events happen. As a prototype for ensuring the successful operation of inherently risky systems, resilience has demonstrated itself to be a promising concept to address the above-mentioned challenges. A standard multi-dimensional resilience triangle model is first presented based on the concept of the three-phase system resilience cycle, which can provide a theoretical foundation for indicating the utility objectives of resilience design. Then, the resilience design problem for CESs is proposed as a multi-objective optimization model, in which the three objectives are to maximize the survival probability, to maximize the reactive timeliness and to minimize the total budgeted cost. Furthermore, the proposed multi-objective optimization programming is solved based on the efficient multi-objective evolutionary algorithm NSGA-II. Finally, the effectiveness of the proposed models and solving procedure is illustrated with an engineered electro-hydrostatic aircraft control actuator resilience design problem, a comparative analysis on the case study is also carried out with respect to previous works. This work can provide an effective tradeoff foundation to improve the resilience of CESs.