Search Results

Understanding total fatigue life of welded joints is crucial to developing durable products. Traditional fatigue analysis methods have focused independently on either crack initiation or crack growth. Each of these methods has strengths but neither method predicts the total life of the part from fabrication to fracture. Recently the SAE Fatigue Design and Evaluation committee evaluated and validated a fatigue analysis technique that can predict the total life of the weld, from microscopic crack initiation to crack growth and finally to fracture. This paper describes a finite element-based (FE) methodology for implementing this total life fatigue analysis in a CAE environment.

Finite Element Analysis (FEA)-based structural simulations are typically used to assess the durability of automotive components. Many parts experience vibration in use, and resonance effects are directly linked to many structural problems. In this case, dynamics must be included in the structural analysis. Dynamic FEA can be more realistic than static analysis, but it requires knowledge of additional characteristics such as mass and damping. Damping is an important property when performing dynamic FEA, whether transient or steady state dynamics, as it governs the magnitude of the dynamic stress response and hence durability. Unfortunately the importance of damping is often overlooked; sometimes a default damping value is erroneously assumed for all modes. Errors in damping lead to errors in the stress response, which in turn lead to significant changes in the fatigue life estimates.

As part of the design and validation of engine-mounted components, it is essential to define the vibratory mechanical environment in which these components will operate. This is required in order to optimize the reliability of such components subjected to loading from both the engine and road profile, while minimizing development costs and time scales. This paper presents a methodology that superimposes a swept sine on a power spectral density of acceleration in order to evaluate the mechanical durability of engine mounted or gear box mounted components. The first step in the process is to obtain the wave form of the dominant engine orders by extracting the deterministic signals from the random process using an order tracking method in the time domain. The second step is to assess the fatigue damage and extreme response spectra of a Swept-Sine-On-Random profile.

Ground vehicle components are designed to withstand the real operational conditions they will experience during their service life. Vibration tests are performed to qualify their endurance. In order to replicate the same failure mechanism as in real conditions, the test specification must be representative of the service loads. The accelerated testing method, based on fatigue damage spectra (FDS), is a process for deriving a synthesized power spectral density (PSD) representing a random stationary Gaussian excitation and applied over a reduced duration. In real life, however, it is common that service loading includes non-Gaussian excitations. The consequences of not using a representative test signal during product validation testing are a higher field failure rate and added warranty costs. The objective of this paper is to describe a method for synthesizing a PSD test specification with a given kurtosis value, which represents a nonstationary non-Gaussian signal.