Reliability Graph Solver for Aircraft Current Return Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current reliability estimation methods for large networked systems, such as aircraft current return networks, are inadequate due to the complexity and computational limitations of fault tree and reliability block diagram approaches, which are error-prone and unable to handle large models effectively.
Innovation Solution
A computer-based method using an improved reliability graph solver within the SHARPE software package that employs heuristic algorithms to select important paths and cutsets, calculating upper and lower bounds for reliability estimation, allowing for efficient approximation of reliability values in large networked systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fault tree method is used for reliability estimation, then reliability assessment can be performed, but the method becomes prohibitively complex and error-prone for large aircraft current return networks
Solution Approach 1:
The patent segments the large aircraft current return network into smaller sub-networks or modules, each analyzed independently using reliability graph methods. This segmentation reduces the overall model complexity while maintaining reliability assessment capability, avoiding the exponential growth of fault tree complexity in large systems.
Solution Approach 2:
The patent replaces the traditional fault tree mechanical construction method with an automated reliability graph solver that uses graph theory algorithms. This substitution eliminates manual model building errors and reduces complexity by using computational algorithms instead of manual fault tree construction and analysis.
2Device complexity
If reliability block diagram method is used, then reliability estimation is simplified, but the method cannot handle large network models effectively due to size and computational limitations
Solution Approach 1:
The patent implements a dynamic reliability graph solver that can adapt to large network sizes by using efficient graph algorithms and data structures. The solver dynamically adjusts its analysis approach based on network size and complexity, enabling it to handle large aircraft CRN models that exceed the capacity of static reliability block diagram methods.
Solution Approach 2:
The patent changes the fundamental parameters of the reliability analysis by using graph theory representations (nodes, edges, paths) instead of traditional block diagram parameters. This parameter transformation enables the system to handle large-scale networks efficiently through automated graph algorithms, overcoming the size limitations of reliability block diagram methods.
3Productivity
If automated translation to fault tree model is attempted, then translation speed increases, but current solvers cannot handle the resulting large fault tree models
Solution Approach 1:
The patent substitutes the fault tree solver mechanism with a reliability graph solver that is specifically designed to handle large-scale networks. This substitution maintains the benefit of automated translation while overcoming the solver capability limitation, as the graph-based approach scales better with network size.
Solution Approach 2:
The patent changes the mathematical representation from fault tree Boolean expressions to reliability graph path-based calculations. This parameter change enables the system to process large automated models efficiently, as graph algorithms can handle the scale of aircraft CRN networks without the exponential complexity that plagues fault tree solvers.
4Measurement precision
If exact solution methods are used for large reliability graphs, then solution accuracy is maximized, but computational intractability prevents practical application
Solution Approach 1:
The patent implements a hybrid approach that uses exact solution methods for critical sub-networks and approximation methods for less critical portions. This partial application of exact methods maintains solution accuracy where needed while reducing overall computational time, making large aircraft CRN analysis practically feasible.
Solution Approach 2:
The patent segments the large reliability graph into smaller sub-graphs that can be analyzed exactly, then combines the results using graph theory principles. This segmentation enables exact solution methods to be applied to manageable portions of the system, maintaining accuracy while avoiding the computational intractability of analyzing the entire large network at once.
Data Source
AI summary
A computer-based method for determining a probability that no path exists from a specified starting node to a specified target node within a network of nodes and directional links between pairs of nodes is described. The nodes and directional links form paths of a reliability graph and the method is performed using a computer coupled to a database that includes data relating to the nodes and the directional links The method includes selecting a set of paths, from the set of all paths, between the starting node and the target node that have been determined to be reliable, calculating a reliability of the union of the selected path sets, setting an upper bound for the unreliability of the set of all paths, based on the calculated reliability, selecting a set of minimal cutsets from all such cutsets that lie between the starting node and the target node, calculating the probability of the union of the minimal cutsets, and setting a lower bound for the unreliability of the set of all cutsets.