Fatigue Design Load Identification Using Engineering Data Analytics

Abstract

Selecting an appropriate concept design in the early stage of the product development process is crucial for successful machine development. As one of the important design requirements, a structural design needs to validate the structural fatigue life against physical tests of actual field operations. Traditionally, the fatigue design loads for the concept design evaluations are generated by hand calculations based on past experience to capture envelopes of expected system responses. But such an approach often does not capture actual loading behaviors in the field. An alternative approach of leveraging observed data from physical tests is highly desired, and a new method of this approach is introduced in this study. Our goal is to develop a new methodology of identifying fundamental fatigue load (FFL) contents from observed data by using engineering data analytics (EDA) techniques. The proposed methodology is applied to determine a new sensor layout which enables us to capture fundamental structural damage load patterns. Numerical demonstrations and case studies of the proposed method are presented with a common structural component, an I-section cantilever beam, and an industrial large-scale structure, a front linkage of a hydraulic excavator.