A real time neural net estimator of fatigue life

A neural network architecture is proposed to estimate, in real time, the fatigue life of mechanical components. Component loading values are used as input by a two-hidden-layer feedforward neural network which has been trained to estimate the fatigue life resulting from an arbitrary load history. The ability of the network to learn the mapping based on a local strain approach between load sequence and fatigue life has been demonstrated for a uniaxial RQC-100 component. Because of its performance, the neural computation can be extended to load/fatigue mappings from experimental data in complex cases where the loads are biaxial or triaxial and where the geometry of the component is complex. In addition, the parallel network architecture allows real-time life calculations, even for high-frequency vibrations. Owing to its distributed nature, the neural implementation is robust and reliable, enabling it to be used in hostile environments, such as reusable rocket engines