CAD/CAE/CFD/CAO/HPC новости 8 Июня 2004 года
Данная новость была прочитана 3218 раз

Союз MSC.Nastran и ADAMS оправдал ожидания

 In April 2002, MSC.Software purchased Mechanical Dynamics Inc., thus marrying the former's MSC.Nastran structural finite element analysis software with the latter's ADAMS kinematic analysis package. Like any marriage, this one depends on communication—the ability of one package to talk to the other. To find out how well Nastran and ADAMS communicate, I tried them out on an analysis of an ATV (all-terrain vehicle) design, which engineers at MSC provided. My specific goal was to see if I could send files from Nastran into ADAMS without going through a translator, which had previously been required. The translator can add hours if not days to an analysis project, and can lose details in the model.

FEA and kinematics analysis are inherently different. FEA depends on material properties, nodes/elements, external loads, and applied boundary conditions. Kinematics depends on mass, damping properties of the rigid bodies, and external transient and static loads. Without FEA, engineers can't get the stiffness and natural frequency of a part. Without kinematics, they can't get overall dynamic behavior. Having all that data can be important, especially in automotive applications.

The Nastran files represented the deformable body of the lower control arm of the ATV. I wanted to put it into the ADAMS file so it could replace the equivalent solid part in ADAMS. That ADAMS part model had no nodes or elements, and therefore no stiffness. Without that stiffness, it would be impossible to check for actual deformation of the component under varying loads from the road so I could see where the stresses were.

The first part of the simulation was strictly for creating the bulk data file (.bdf) from the solid model of the lower control arm. Once I meshed the part, using multi-point constraints and attachment points, I created a superelement to help in extracting the normal modes of the lower control arm. That information went into Nastran.

The Nastran solver extracts the normal modes and outputs the model neutral file that will be read in ADAMS. The elimination of the translator in this stage is the real time saver.

The ATV suspension at this point has the rigid lower control arm in place and I ran the complete vehicle dynamic simulation in ADAMS with it. That exercise extracted the frequency mode shapes for all the components of the ATV that are rigid bodies. Once the part with the rigid lower control arm was done, I replaced it with the flexible lower control arm that I analyzed in Nastran with its mode shapes and imported into ADAMS. When I transferred the joint information of the rigid lower control arm to its flexible counterpart, I deleted the rigid lower control arm. No, my task was to run the ADAMS vehicle simulation once more with the flexible lower control arm mounted on the ATV assembly and measure its performance under the same conditions.

Conclusion

The software passed the test. I was able to easily import the Nastran model of a flexible part into a multi-body assembly in ADAMS. Calculating the part's performance was more accurate since I was able to take into account the actual flexibility of the part. The software engineers at both companies worked together well to write the algorithms necessary to accomplish the integration.

ADAMS View: At top is the ATV. The post processor is configured with three window graph areas (one of 12 possible layouts). The animation plot below left will be in sync with the cursor tracing force plots shown on the right. Since we ran the simulation twice, the force results with the rigid lower control arm can be overlaid with the ones of hte flexible lower control arm. One can appreciate the value of the Nastran/ADAMS integration by observing the differences in the force repsonse between the models. With the default settings, only the deformation can be shown as a contour plot, while entities like stresses and strains can be set in the MNF file.