Browse Prior Art Database

Efficient Computational Procedure for Fatigue Life Assessment Using DOE/DSA Techniques

IP.com Disclosure Number: IPCOM000008913D
Publication Date: 2002-Jul-23
Document File: 3 page(s) / 62K

Publishing Venue

The IP.com Prior Art Database

Abstract

During the early product design stage, it is very important to explore the design space for the optimal and robust design. The current analytical design of experiment (DOE) procedure is not efficient for durability fatigue applications in automobile industries. In the fatigue analysis, the unit-load stress coefficients are needed to compute the stress history and fatigue life. The usual practice is to run finite element analysis (FEA) to obtain those stresses. The FEA is performed using any commercial FEA software and, depending on the size of the model, it can be computationally very expensive. The DOE process works by initially laying out the design points in the space defined by the design parameters. For each design point, the finite element analysis must be performed. The serious drawback with the current DOE procedure is that hundreds of FEA evaluation may be needed to obtain a reasonably accurate fatigue life distribution, which makes the current DOE process undesirable to many engineers in automobile industries. Since automotive engineers are faced with vehicle Computer-Aided Engineering (CAE) models that have up to half million degrees of freedom, the current DOE process is impractical.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 50% of the total text.

Efficient Computational Procedure for Fatigue Life Assessment Using DOE/DSA Techniques

During the early product design stage, it is very important to explore the design space for the optimal and robust design.  The current analytical design of experiment (DOE) procedure is not efficient for durability fatigue applications in automobile industries.  In the fatigue analysis, the unit-load stress coefficients are needed to compute the stress history and fatigue life. The usual practice is to run finite element analysis (FEA) to obtain those stresses.  The FEA is performed using any commercial FEA software and, depending on the size of the model, it can be computationally very expensive.  The DOE process works by initially laying out the design points in the space defined by the design parameters.  For each design point, the finite element analysis must be performed.  The serious drawback with the current DOE procedure is that hundreds of FEA evaluation may be needed to obtain a reasonably accurate fatigue life distribution, which makes the current DOE process undesirable to many engineers in automobile industries.  Since automotive engineers are faced with vehicle Computer-Aided Engineering (CAE) models that have up to half million degrees of freedom, the current DOE process is impractical.

The clear advantage of the proposed DOE/DSA procedure is that it reduces the number of FEA to only two.  In other words, even if the engineer decides to perform 100-run DOE, only two FEA evaluations are needed in the entire process.  (In the current DOE process, the engineer would need to evaluate 100 finite element analyses.)  This is possible by utilizing the design sensitivities from the FEA code, such as MSC/NASTRAN, and updating the unit-load stress coefficients by using these sensitivities.  Therefore, by integrating the sensitivity information into the DOE process, the proposed DOE/DSA process is much more efficient and desirable for automotive engineers to use.

Procedure

When performing an analytical design of experiment (DOE) for fatigue life assessment, the following steps are currently taken (See Figure 1):

1.       Define the design space from the selected design parameters (gage, material, load, etc.).

2.       Select the DOE method.

3.       Spread the design points in the design space.

4.       For each design point;

4a.  Run the finite element analysis to obtain unit-load stress coefficients.

4b.  Calculate the stress history from the given load history.

4c.  Run the fatigue analysis

5.       Go to step 4 until all design points are evaluated.

The accuracy of fatigue life assessment rests on how well, and how many, the design points are spread across the design space.  Ideally, the engineer would prefer to spread many design points as evenly as possible to cover the entire space.  But due to the lengthy nature of DOE procedure, the engineer cannot afford to evaluate too many design point...