Method and system for generating a test coupon specification for predicting fatigue life of a component

A method for generating a test coupon specification for predicting fatigue life of a component, the component being a mechanical component, the method comprising:

• determining a load condition for the component;
• providing a model for a component design;
• performing a strength analysis on the component design under the load condition, such that a critical area of the component and at least one stress-related parameter of the critical area are determined;
• providing at least one material condition of the component at least for the critical area of the component;
• providing a material model, the material model being a machine learning fatigue model;
• providing the at least one stress-related parameter and the at least one material condition as an input to the machine learning fatigue model;
• generating, by a processor executing the machine learning fatigue model, as an output, the test coupon specification for being tested in a testing machine, wherein the machine learning fatigue model is provided by a method comprising:
• providing a set of test coupon specifications to the machine learning fatigue model, wherein the set of test coupon specifications comprises different material conditions of the at least one material condition of the component;
• collecting data points for maximum stress versus cycles to failure for the test coupon specifications;
• and
• training the machine learning fatigue model with the collected data points and selecting a test coupon specification from the set of test coupon specifications by applying the at least one stress-related parameter and the at least one material condition to the trained machine learning fatigue model as inputs;
• and
• producing a test coupon based on the generated test coupon specification, the produced test coupon being a physical device usable for fatigue prediction of the component.