Superimposed Failure Model for Maintenance Interval Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lifetime distribution models for parts in manufactured products, such as aircraft, only account for the first instance of failure, leading to overly conservative maintenance estimates and inefficient maintenance scheduling due to lack of consideration for multiple failure instances across different locations.
Innovation Solution
A method is developed to generate a superimposed failure model (SFM) using failure data from multiple instances of a part located at distinct locations, applying lifetime distribution models like exponential or Weibull distributions to determine a more accurate maintenance interval by calculating survival functions and residual lifetimes, thereby refining maintenance scheduling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a traditional lifetime distribution model is used that only accounts for the first instance of failure, then the model is simple to implement, but the maintenance interval estimate becomes overly conservative and inaccurate
Solution Approach 1:
The patent segments the failure data by separating the first failure instance from subsequent failure instances. The traditional model handles the first failure, while the new superimposed failure model specifically handles subsequent failures at different locations. This segmentation allows the system to maintain simplicity for the baseline model while adding complexity only where needed to improve accuracy for multi-instance parts.
Solution Approach 2:
The patent merges the traditional lifetime distribution model with a new superimposed failure model to create a comprehensive maintenance scheduling system. The traditional model provides the baseline lifetime estimate, while the superimposed model adds the effect of multiple failure instances. By combining these models, the system achieves more accurate lifetime estimates without completely replacing the proven traditional approach.
2Productivity
If maintenance intervals are determined using traditional models, then the scheduling process is straightforward, but unnecessary maintenance activities increase leading to resource waste
Solution Approach 1:
The patent changes the key parameter from a single lifetime estimate to a superimposed lifetime distribution that incorporates multiple failure instances. By modifying the lifetime parameter to account for parts used in multiple locations with different failure patterns, the system optimizes maintenance intervals to reduce unnecessary maintenance activities while maintaining reliability.
Solution Approach 2:
The patent replaces the traditional mechanical approach of uniform maintenance scheduling with a data-driven statistical model. Instead of applying fixed maintenance intervals based on simple lifetime estimates, the system uses superimposed failure models to dynamically determine optimized maintenance intervals based on actual failure patterns across multiple locations, reducing unnecessary maintenance time loss.
3Reliability
If multiple failure instances are considered in the model, then the lifetime estimate becomes more accurate, but the computational complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing failure data to identify and categorize failures by location and instance number before applying the superimposed model. This preliminary organization of data simplifies the subsequent modeling process, allowing the system to handle multiple failure instances systematically without overwhelming computational complexity during the actual lifetime estimation phase.
Data Source
AI summary
A system is provided for maintenance of a manufactured product composed of a plurality of parts. A modeling engine may receive failure data for the plurality of parts in which the failure data indicates individual instances of failure of at least two of a multiple quantity of a part of the plurality of parts. Each of the at least two of the multiple quantity may be located at a respective distinct location in the manufactured product. The modeling engine may generate a superimposed failure model (SFM) for the part, including determining a lifetime distribution of the part based at least partially on application of a lifetime distribution model to the SFM. A maintenance engine coupled to the modeling engine may perform a maintenance activity including determining a maintenance interval determined for the part according to the lifetime distribution of the part.