Stuck-at fault

About: Stuck-at fault is a research topic. Over the lifetime, 9707 publications have been published within this topic receiving 160254 citations.

Hybrid System State Tracking and Fault Detection Using Particle Filters

Abstract: When particle filters are used for fault detection, they have the problem of sample impoverishment, which means there are not enough particles that can transition to a rare-occurring faulty mode. The consequence is that the fault cannot be properly detected. This paper proposes a method to overcome this problem. Essentially, we develop an algorithm for tracking the states of hybrid systems where fault detection is modeled as a special case of the state tracking of a hybrid system. Extensive simulations are carried out to analyze the effects of various parameters on the performance of the algorithm. It is shown that our algorithm can detect both known and unknown faults using a very small number of particles

Method and system for processing alarm objects in a communications network

Abstract: A distributed method and system of controlling a communications network having a plurality of spans of interconnected network elements some of which include a network element processor distributes network topology information to respective span databases; stores original fault objects in the respective span databases; advertises fault objects to other network element processors in a local span when the original fault affects network elements other than a network element in which the fault occurred; advertises alarm objects to other network element processors that are respectively associated with a circuit affected by the original faults; stores the advertised fault and alarm objects in the respective span databases; and performs distributed processing of the advertised fault and alarm objects with the other network element processors and the respective span databases. Aggregation of other faults and alarms that may be occurring on the communications network due to other faults other than the received fault aids in determining causality of the fault. Causality may be determined by correlating other faults and alarms with the received fault. If not a root cause of another fault or alarm, the received fault is sympathetic to another fault or alarm. Sympathetic faults are suppressed while root cause faults are promoted to an alarm and reported to affected network elements. The number of alarms viewed by a network manager as well as the reporting of alarms and underlying faults are reduced by performing such distributed alarm correlation and fault reporting suppression.

Forward-looking fault simulation for improved static compaction

Abstract: Fault simulation of a test set in an order different from the order of generation (e.g., reverse- or random-order fault simulation) is used as a fast and effective method to drop unnecessary tests from a test set in order to reduce its size. We propose an improvement to this type of fault simulation process that makes it even more effective in reducing the test-set size. The proposed improvement allows us to drop tests without simulating them based on the fact that the faults they detect will be detected by tests that will be simulated later, hence the name of the improved procedure: forward-looking fault simulation. We present experimental results to demonstrate the effectiveness of the proposed improvement.

A Random and an Algorithmic Technique for Fault Detection Test Generation for Sequential Circuits

Abstract: Two procedures are presented for generating fault detection test sequences for large sequential circuits. In the adaptive random procedure one can achieve a tradeoff between test generation time, length, and percent of circuit tested. An algorithmic path-sensitizing procedure is also presented. Both procedures employ a three-valued logic system. Some experimental results are given.

Improved accuracy in quantitative fault tree analysis

Abstract: The fault tree diagram defines the causes of the system failure mode or 'top event' in terms of the component failures and human errors, represented by basic events. By providing information which enables the basic event probability to be calculated, the fault tree can then be quantified to yield reliability parameters for the system. Fault tree quantification enables the probability of the top event to be calculated and in addition its failure rate and expected number of occurrences. Importance measures which signify the contribution each basic event makes to system failure can also be determined. Owing to the large number of failure combinations (minimal cut sets) which generally result from a fault tree study, it is not possible using conventional techniques to calculate these parameters exactly and approximations are required. The approximations usually rely on the basic events having a small likelihood of occurrence. When this condition is not met, it can result in large inaccuracies. These problems can be overcome by employing the binary decision diagram (BDD) approach. This method converts the fault tree diagram into a format which encodes Shannon's decomposition and allows the exact failure probability to be determined in a very efficient calculation procedure. This paper describes how the BDD method can be employed in fault tree quantification.