(DINO-STR-Test-FEA)10/9/18
Structure Test Plans
Subsystem:Structures______
Test Plan Subject: Finite Element Analysis______
Document Name:___(DINO-STR-Test-FEA)___
Author:Grayson McArthur______
Date:December 8, 2003______
Revision Log
Revision / Description / DateSignatures
Document Originator: ______
NameSignatureDate
Subsystem Team Lead: ______
(if different than above) NameSignatureDate
Systems Lead: ______
NameSignatureDate
Table of Contents (Delete this if the study is short)
1.Introduction
2.Analysis Setup
2.1.Loads and Restraints
2.2.Studies
2.3.Settings
3.Design Options
4.Future Work and Trouble Areas
4.1.FEA for Whole Structure
4.2.Box Modeling
4.3.Displacement
5.Conclusion
1.Introduction
Finite element analysis (FEA) is a method of decomposing a structure into a system of bars connecting nodes. By doing so this enables one to do calculations for a finite number of bars and nodes at a localized position. A bar element is essentially a spacer connecting the two end nodes and represents the material properties of the system. This spacer can also be thought of as a spring between the nodes, because the displacement of each node depends on the properties of the bar. A node represents the ends of a bar element and joints where two or more bar elements come together. All calculations are done to find results at each node of a bar. FEA is not an exact analysis but, the accuracy can be controlled by the number of nodes used to model the structure. The FEA program used for this report was Cosmos Works. There is a useful tutorial in Cosmos Works that should be completed before analysis if one does not have any experience with FEA or Cosmos.
The main objective of analyzing the side panel of the DINO structure is to optimize the mode frequency while minimizing mass. One idea on how to accomplish this was to add a solid plate onto the back of the Isogrid panel. Four cases were considered, each a variation on the thickness of the panel and the use of a plate.
2.Analysis Setup
2.1.Loads and Restraints
The first and probably most important step in preparing the model for analysis was choosing the location of the restraints and loads. Each of the eight mounting holes should be selected as the locations of all the restraints. The inside face of each hole was selected as the surface that the restraint will act over, because this ensures that the part is held steady and does not rotate about the edge of the hole. These restraints should be selected as unmovable, no translation, because each end of the panel will be bolted to either the top or bottom plate of the structure. As stated in the design requirements for DINO, the structure must withstand a 20G force in the direction of each axis. This load is described as acting through the center of gravity. Cosmos allows for a gravity load to be applied in the three directions simultaneously. Doing so is not practical, because it is not asked for in the requirements, such a load would greatly complicate the analysis, and may not represent the environment the structure will be in.
2.2.Studies
Cosmos Works can analyze the structure as a frequency test, static, buckling, thermal, or optimization. The only two studies concerned with for this report are the frequency and static. To test how the structure performs during vibrations is done using the frequency study. A predetermined vibration can be set as frequency the panel will be tested at, but for the purposes of the DINO project the mode frequency needed to be determined. The program was set without an upper or lower frequency and told to calculate the first 5 modes. The static study was used to calculate the stress, strain, and displacement of the part under the 20G load. Even though Cosmos is capable of factoring in the thermal effects and friction, these options were not used for the analysis of the panel. A static study was done for each 20G loading case as well.
2.3.Settings
After starting the studies, the material properties must be assigned and a mesh created. Assigning the material properties is done by selecting the desired material from the library provided with the program and setting the load type as linear elastic isotropic using SI units. The material used was aluminum 6061. Meshing creates the bar elements and nodes that model the structure. For the three dimensional structure, the nodes and bars are arranged in a three dimensional triangle called a brick element. The mesh basically controls the accuracy of the analysis by determining the number of bars and nodes used to model the panel. With more nodes the program will do more calculations to determine the forces acting on the structure, increasing the accuracy. The global size, set at 3.0328mm, controls the distance between each node, therefore the smaller the global size the more nodes used to make the model. The tolerance was set at 0.15164mm. The finer the mesh is the more calculations the program has to do to evaluate the model, which can take a considerable amount of time. One thing to take into consideration is that the mesh should be three brick elements across at the parts thinnest section. The analysis done using the mesh described above is pretty reasonable, but could be finer. For this preliminary analysis this mesh was decided to be adequate. The Isogrid panels and the outside plate were created in separate Solid Works files and assembled to represent the final product. In the cases that did require an outside plate, the two touching faces had to be set as bonded. This setting models the two parts as if they are made out of a single piece of material.
3.Design Options
Four different cases were analyzed, two of which were just the Isogrid panel and the other two included the outer wall. Table 1 shows the mass and thickness break downs for each design option. Due to constraints on the structure the maximum total thickness of the side panel could not exceed 0.25 in. In an attempt to find a trade off between mass and stiffness two cases were considered were the total thickness of the panel was 0.125 in. The minimum thickness of the wall due to machining constraints is 0.05 in. To conserve mass this minimum thickness was selected. The plots generated with Cosmos Works for the design chosen are contained in the DINO Technical Manual.
Table 1: Design Option Specifications
Option 1 / Option 2 / Option 3 / Option 4Isogrid Thickness (in) / 0.25 / 0.125 / 0.20 / 0.075
Wall Thickness (in) / none / none / 0.05 / 0.05
Total Thickness (in) / 0.25 / 0.125 / 0.25 / 0.125
Mass of Isogrid (kg) / 0.339 / 0.170 / 0.272 / 0.102
Mass of Wall (kg) / none / none / 0.223 / 0.223
Total Mass (kg) / 0.339 / 0.170 / 0.495 / 0.325
4.Future Work and Trouble Areas
4.1.FEA for Whole Structure
At the time of this report several tasks still remain to be done for a complete FEA. The main reason for these not being completed is that the DINO structures team did not have any members with FEA experience, so all the analysis had to be started from square one, learning the basics of FEA. The biggest and most complex analysis left is to model the entire satellite structure as a whole. This should be done by assembling all of the side panels with the top and bottom plates and then meshing the structure as a whole and subjecting it to the three 20 G loads. Several attempts were made to accomplish this but none were successful. After speaking with people more experienced with FEA several suggestions were brought up. One being removing the horizontal mounting holes in the top and bottom plates because they are not drilled all the way through the material. Another but time consuming suggestion was to make the mesh as fine as Cosmos will allow. There also might be some problems with how the side panels are modeled where they connect to the top and bottom plates. In reality the side panels will be secured to the plates using bolts so they should not be modeled in the same manor as the touching faces between the Isogrid and outer wall.
4.2.Box Modeling
The structure will house multiple boxes of different sizes and masses. When the box arrangement is decided the possibility of modeling the boxes inside the structure should be explored. The load of the boxes will affect the stress, strain, displacement, and mode frequency. To what extent these will affect the performance should be determined to decide if there need to be any reinforcing of the structure. This also might reveal opportunities to cut mass from the system if the load from the boxes increases the mode frequency. One way to create this model might be to add point masses at the mounting holes locations that will be use to secure the box.
4.3.Displacement
Based on discussions and recommendations from those more experienced with FEA a displacement requirement should be created or at least examine the impact of structural displacement on the satellite. One concern might be that if a panel displaces to much it might cause components inside to collide. A way around this problem might be to determine the maximum expected displacement of the structure with the boxes placed inside the structure. Then the boxes could be arranged in such a way that they will not collide during this maximum displacement.
5.Conclusion
At the time of this report the FEA is not complete due to having to learn Cosmos and FEA from the ground up. The intent of this report is to provide any predecessors with the volume of knowledge obtained thus far about how to set up the structure for analysis. Much progress has been made in the field of understanding the appropriate methods to use for a comprehensive analysis of the entire structure. The Technical Manual contains all the results generated with Cosmos Works for the chosen side panel design. The Trade Point Study on the side panel design contains the frequency results for all of the design options.
1of 6