Probabilistic Fatigue Crack Life Estimation via Monte Carlo Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating fatigue crack life in components under cyclic stress are limited by their inability to account for material property and flaw size scatter, leading to either overly conservative designs or inaccuracies in probabilistic calculations, particularly due to neglecting individual material variations and scatter in properties.
Innovation Solution
A probabilistic fatigue crack life estimation method that uses a Monte-Carlo approach to simulate a large number of component representations, each defined by random material and flaw size conditions, to determine the probability of failure by calculating crack growth through Linear Elastic Fracture Mechanics and its extensions, incorporating scatter in material properties and flaw sizes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deterministic fracture mechanics life calculation based on minimum/maximum material properties and flaw-sizes is used, then safety is improved, but design conservatism increases and component potential is not fully utilized
Solution Approach 1:
The patent transforms the deterministic approach (using fixed minimum/maximum values) into a probabilistic approach by changing the parameters from single values to distributions. Material properties and flaw sizes are represented as probability distributions, allowing the calculation to account for the actual scatter in data rather than using overly conservative bounds, thus resolving the contradiction between safety and component potential utilization
Solution Approach 2:
The patent creates multiple virtual copies of the component, each with different material properties and flaw sizes sampled from their respective distributions. By analyzing many such copies through Monte Carlo simulation, the method captures the full range of possible outcomes without requiring extremely conservative single-value assumptions, thereby improving both reliability assessment and component utilization
2Reliability
If conservative assumptions are made in fatigue crack life estimation, then safety is improved, but measurement precision of actual life decreases
Solution Approach 1:
The patent changes the approach from using conservative fixed values to using probability distributions for material properties and flaw sizes. This allows the life estimation to reflect the actual precision of the input data, reducing unnecessary conservatism while maintaining safety through proper probabilistic analysis
Solution Approach 2:
The patent incorporates feedback from actual measured data by using Weibull distributions fitted to experimental results. The distribution parameters (shape and scale) are derived from actual measurements, allowing the model to continuously improve its accuracy as more data becomes available, thus resolving the contradiction between safety and measurement precision
3Measurement precision
If scatter in material properties and flaw sizes is accounted for using probabilistic methods, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent uses Monte Carlo simulation which creates numerous virtual copies of the component with randomly sampled properties from distributions. This approach improves measurement precision by accounting for scatter, while the computational complexity is managed through efficient sampling algorithms and the use of standardized distribution functions
Solution Approach 2:
The patent transforms the complex problem of accounting for scatter in multiple parameters into a more manageable form by representing all uncertainties through probability distributions. This parameter transformation allows the use of well-established statistical methods and Monte Carlo techniques, improving precision without excessive complexity
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
The present invention relates to probabilistic estimation of fatigue crack life of a component configured for being subjected cyclic stress. A plurality of representations of the component are defined from material property scatter data (90a-g) and flaw-size scatter data (90h-j)of the component, wherein each representation is defined by one possible material condition and flaw-size condition associated with the component. For each individual representation, a component location is selected (93) and a determination (96) is made whether said individual representation fails after a given number of cycles N, based on the calculation (95) of a crack growth in the selected location. The crack growth is calculated (95) on the basis of the material condition and the flaw-size condition in the selected location. Failure of the individual representation is determined (96) if the crack growth is determined to be unstable. The sum total of the number of the representations that failed after N cycles is determined (97). A probability of failure of the component after N cycles is then determined (99) as PoF(N) = Nf/S, wherein PoF(N) is the probability of failure of the component after N cycles, Nf is the sum total of the number of representations that are determined to have failed after N cycles, and S is the total number of representations.