Fatigue Risk Analysis with Intelligent Digital Twins Based on Condition Monitoring

Fatigue mechanisms proceeding through the formation and growing of cracks are well-known but the progress is not easy to detect with measurements during the operation. The localised structural damage is caused by repeated loading and unloading when the load exceeded certain thresholds. The effect of the loading is highly nonlinear and structures fracture suddenly when a crack reaches a critical size. This research focuses on the advanced data analysis aimed at detecting effective stress impacts by using generalised norms and intelligent stress indices based on nonlinear scaling to provide good severity indicators. Digital twin type solutions can help in detecting changes in fatigue risk analysis. Contributions of the stress are calculated in each sample time, which is taken as a fraction of the cycle time. The Wöhler curve is represented by a linguistic equation (LE) model where the stress part is represented by the intelligent stress indices. The cumulative sum of the contributions indicates the deterioration of the condition, and the simulated sums can be used to predict failure time. Scheduling the maintenance actions can be extended to avoiding risky stress levels. The generalised statistical process control (GSPC) is a feasible solution to demonstrating in real time these risky levels during the operation. The analysis is adapted to changing operating conditions by updating recursively the parameters of the scaling functions. Algorithms of the model and digital twin remain unchanged. In a rolling mill, torque measurements are collected and analysed with a combination of two norms scaled with the nonlinear scaling approach.