Cascade failures

Failure of one element will cause an overvoltage across the others in parallel and may lead to a cascade failure. 

Figure 25.3 Cascade failure of healthy elements in the absence of an internal protection...

The Mars Polar Lander loss is a component interaction accident. Such acddents arise in the interactions among system components (electromechanical, digital, human, and social) rather than in the failure of individual components. In contrast, the other main type of accident, a component failure accident, results from component failures, including the possibility of multiple and cascading failures. In component failure accidents, the failures are usually treated as random phenomena. In component interaction accidents, there may be no failures and the system design errors giving rise to unsafe behavior are not random events. 

It is predictive and target setting (i.e. it determines the system s safety objectives without any architectural limitations). It should be used to identify design precautions necessary to ensure independence, to determine the required software level and to avoid common mode and cascade failures (see Chapters 6-8). 

Cascade failures A single electrical bus failure may overload the remaining channels, thereby increasing the probability of their failure. In a two-channel system, each channel with a failure rate of 1 in 1000 h, the probability of any one of the... 

Provide warning systems which allow the crew to immediately recognise the problem and take appropriate corrective action. Consider the implication of multiple simultaneous warning (e.g. from cascade failures) and their impact on crew workload and performance (see Chapter 10). Ensure that the pilot is kept in the loop . For instance, an autopilot may automatically cope with a series of developing problems without informing the crew. When it reaches a point where the autopilot cannot control the aircraft, it may then suddenly drop a complex situation in the lap of the pilot. 

While system-level effects are expressed, the technique provides for limited traceability between cascading failure modes and system hazards. 

Independence A concept that minimises the likelihood of common mode errors and cascade failures between aircraft/system functions or items [ARP4754A],... 

Carreras et al., 2005. Critical points and transitions in an electric power transmission model for cascading failure blackouts, Probab. Eng. Inf. Sci., 19 4, 475 88 Conrad, St. H et al. 2006. Critical national infrastructure reliability modeling and analysis, Bell Labs Technical Journal,. ll(3) pp. 57-71. 

The location-specific interdependencies are analogous to the geographical and physical categories of Rinaldi et al. (2001) and the spatial interconnectedness of Zimmerman (2001). These mainly cause common cause failures (CCF). Fimctional interdependencies focus on the functions of the systems in the scenario. This is similar to the cyber and logical interdependency categories of Rinaldi et al. (2001), and to the functional interconnectedness of Zimmerman (2001). These interdependencies mainly cause cascading failures. 

ABSTRACT In this work, original indicators are introdnced to characterize the criticality of components in a network system with respect to their contribution and participation to failure cascade processes. Three different models of cascading failures are considered, differing both on the failure load distribution logic and on the cascade triggering event. The criticality indicators are compared to classical measures of topological centrality, for identifying the centrality measure most characteristic of the cascade processes considered. 

In this study, some of these issues are looked at by means of an abstract modeling framework for analyzing the system response to cascading failures, which can be used to guide a successive detailed simulation focused on the most relevant physical processes and network components. The need for such an analysis tool is even stronger for systems in which the cascade dynamics is rapid and modifications are actuated onto the network in order to mitigate the evolution of the cascade. For example, a cascade of events leading to blackout usually occurs on a time scale of minutes to hours and is completed in less than one day (Dobson, Carreras et al. 2007). 

The paper is organized as follows three models of failure propagation in network systems are presented in Section 2 in Section 3, indicators of component criticality in cascading failures are introduced in Section 4, they are computed for a reference network in Section 5, the results obtained are compared with the topological centrality measures and a centrality measure relevant to the cascade process is identified. Conclusions are drawn in Section 6. 

The algorithm for simulating the cascading failures proceeds in successive stages as follows ... 

Propagation of cascading failures due to targeted intentional attacks... 

The importance of the cascade effect with respect to intentional attacks stems from the fact that a large damage can be caused by the attack on a single component. Obviously, in general more links render a network more resistant against cascading failures, but this increases the cost of the network. 

In this sense, the final size of the cascade is considered a direct indicator of the criticality of a component whereas the duration measure by itself does not allow drawing clear-cut conclusions about the critical contribution of components to the cascading failure process. 

Table 1 shows the results for the model relative to cascading failures due to targeted intentional attacks. This model describes a situation completely different from the previous ones since the load at a component is the total niunber of shortest paths passing through that component. Components 19,5 and 13 turn out to be the... 

In Table 1, the rankings with respect to the four criticality indicators introduced in Section 3 are displayed for each one of the three models of cascading failures. Considering each individual model, the results are consistent with a physical analysis of the network system, in highlighting the components which most affect the failure spreading. 

The results from the propagation of cascading failures due to targeted intentional attacks cannot be fully compared with those of the two previous random failure propagation logics since the former is a completely deterministic process, in which the components load is the number of shortest paths passing through it,... 

Table 1. Summary of the criticality indicators rankings for the three models of cascading failures. 

From the comparison between the results in Table 2 and those for the model relative to cascading failures due to targeted intentional attacks (Table 1), it can be said that the betweeimess centrality measure only partially accounts for the criticalities highlighted by the indicator Si, since not only the centrality of a component is relevant but also the fact that it is connecting central components (as do nodes 14, 16 and 11 in the network). 

When applied to models of local propagation of a fixed amount of load and of redistribution of the failure load, three of the proposed criticality indicators, namely/i, di and st, are consistent in their criticality ranking of the components. When applied to the model of cascading failures due to targeted intentional attacks fi and di are consistent in ranking the components according to their criticality. 

Little, R. G. 2002. Controlling Cascading Failure Understanding the Vulnerabilities of Interconnected Infrastmc-tmes. Journal of Urban Technology 9(1) 109 -123. 

Zheng, J.-F., Z.-Y. Gao, et al. 2007. Clustering and congestion effects on cascading failures of scale-free networks. EPL (Europhysics Letters) (5) 58002. 

It should be determined whether the remaining (intact) capacity of the safety systems is still sufficient for dealing with the situation that has occurred (i.e. the PIE together with the induced effects and cascade failures). If the adequate performance of the safety systems cannot be demonstrated, then either additional protection should be provided or the redundancy of the safety systems should be increased, or a combination of the two could be used, provided that adequate performance of the safety systems can be demonstrated. 

One of the benefits of the prismatic cell is that it has the higher capacity typical of laminate cell, but it also may integrate some of the safety features of the small cells such as vents and, occasionally, thermal fuses. The prismatic cell may also use a mechanical connection to the cell which reduces the need to weld the bus bars to make the cell to cell connections in a pack. The issues of this cell format include the risk of cascading failure (much more energy is released in a large cell failure). In addition, as these cell formats are still produced in relatively low volumes, ramping up the manufacturing to volumes and quality that match those of the 18650 format will take time. 

Security of industrial control systems (ICS) used to control critical infrastructure (Cl) has been extensively studied in the last few years by both industry and academia. Generally, the services offered by Cl are somewhat robust with respect to single component failures of the underlying network. The reaction to multiple and cascading failures, however, is much harder to predict. Dependencies and interdependencies are an important source of risk and a significant factor in our uncertainty of risk assessment, particularly the risk due to cascading failures in which the rate and size of loss is amplified. 

Look up cascading failure. To what part of the infrastructure does it generally apply Are there other parts of the anthrosphere or infrastructure to which cascading failure may apply ... 

Identify the major vulnerabUities and points of weakness in the supply chain for these chemicals and chemical processes that could lead to catastrophic consequences. Include vulnerabilities during transportation and any special vulnerabihty that could exist during national emergencies. Examine the possibility of cascading failures that could lead to catastrophic supply disruption. 

---