ABSTRACT
This report investigates the stresses caused on a steering knuckle component. A physical test with strain gauge readings is compared to a Finite Element Analysis model in Ansys and also validated by hand calculations. The test and Ansys results were found to have a 5% error which is well within a 10% reasonable value. The hand calculations showed a 70% error which was attributed to the simplifications and assumptions made to make the calculations possible as well as the geometry complexity. The component was then redesigned to remove material from low stress area in order to reduce its mass.
Christos Kalavrytinos
Page i
CONTENTS
ABSTRACT ................................................................................................................................ I CONTENTS ............................................................................................................................... II 1.0 INTRODUCTION ................................................................................................................. 1 1.1 OBJECTIVES ...................................................................................................................... 1 2.0 DETERMINATION OF LOADINGS..................................................................................... 2 2.1 ROAD BUMP CASE .............................................................................................................. 2 2.2 BRAKING CASE................................................................................................................... 3 2.3 CORNERING CASE .............................................................................................................. 3 3.0 SOLID MODELING ............................................................................................................. 4 4.0 VALIDATION ....................................................................................................................... 5 4.1 PHYSICAL TEST .................................................................................................................. 5 4.2 HAND CALCULATIONS ......................................................................................................... 6 4.3 FINITE ELEMENT SIMULATION OF TEST ................................................................................. 8 4.4 RESULTS ......................................................................................................................... 11 5.0 FINITE ELEMENT MODELING ......................................................................................... 12 5.1 ROAD BUMP FEA ............................................................................................................. 13 5.2 BRAKING FEA ................................................................................................................. 15 5.3 CORNERING FEA ............................................................................................................. 17 5.4 RESULTS ......................................................................................................................... 18 6.0 DISCUSSION .................................................................................................................... 19 7.0 REDESIGN ........................................................................................................................ 20 8.0 CONCLUSION................................................................................................................... 21 9.0 RECOMMENDATIONS ..................................................................................................... 21 REFERENCES: ....................................................................................................................... 22 APPENDIX .............................................................................................................................. 22
Christos Kalavrytinos
Page ii
1.0 Introduction
The aim of this report is to redesign the steering knuckle (i.e. upright) of an off-road buggy in order to reduce its mass while retaining a satisfactory safety factor. In order for this to be achieved, the component will be subjected to structural analysis with different methods; finite element analysis (FEA) on Ansys, physical test and hand calculations. According to the analysis results, material can be added to parts that are subjected to higher stress than the safety factor permits. Material can also be removed from low stress areas, thus, helping to reduce the component mass. In order to allow for a validation with hand calculations, certain simplifications and assumptions have to be made and justified.
1.1 Objectives
In order to successfully complete the project, the following objectives have to be achieved:
Finite Element Analysis: Determination of loads in three different load cases (i.e. bump, braking, cornering) as provided in the brief. Simplify and make assumptions necessary to perform the finite element analysis. Refine mesh at points of interest and show converged results.
Experimental Test: Observe the physical testing conditions (e.g. supports used, loading arrangement, etc.) and relate that to the FEA model. Observe and record values of elastic strains and deflections.
General: Perform simplified calculations based on mechanics of solids theory to determine stress and strain values in order to compare them to the physical test and FEA model.
Christos Kalavrytinos
Page 1
Redesign the component to reduce weight where possible and ensure a safety factor of 5 on this critical component.
There is also an inertia force created by the unsprung mass which is calculated using only the rim and tyre weight (10kg). This might not be realistic since the unsprung
Christos Kalavrytinos
Page 2
mass consists of other suspension components, but in this case it is an assumption for the sake of simplification. This is how it is calculated.
This force will also be applied on the hub plate opposite to the direction of travel.
This force acts at a distance from the centre of the hub and is therefore a moment. The distance is the dynamic radius R of the wheel at approximately 0.305m.
This force is transferred to the two brake calliper fixing holes which are at a 94.7 mm distance from the centre of the hub and at a 40 angle from the vertical axis. The force is split in half to be applied equally on both holes and analysed to two components.
Christos Kalavrytinos
Page 4
4.0 Validation
The model is validated by hand calculations and strain gauge measurements during a load test. This allows for a comparison between the three sets of results to see if the FEA model is accurate.
Christos Kalavrytinos
Page 5
Known data:
Beam length (L) = 0.222 m Point load at center (P) = 5000 N Distance of strain gauge from centre (x) = 0.0611 m (measured in Catia) Young's modulus (E) = 210 GPa Moment of Inertia (Iyy) =
Christos Kalavrytinos
Page 6
Bending moment:
As shown in the schematic in Fig. the bending moment at a particular point x where the strain gauge no. 6 is positioned, is:
Deflection:
Stress in Y direction:
, where y is the distance from centroid (60.063, 35.559mm from Catia) Christos Kalavrytinos Page 7
Strain in Y direction:
Christos Kalavrytinos
Page 8
The setup used in the validation model is shown in Fig. 8 where frictionless supports were used as the closest representation of the original supports used in the test. A force of 5000 N was applied at the same surface as in the test. Figure 9, shows the geared hole which added more geometrical complexity, thus, increasing the number of nodes to a point with no interest. The geometry was simplified and the mesh was refined in general and also at the position of the strain Christos Kalavrytinos Page 9
gauge as shown in Fig. 10. Mesh statistics show a total of 98439 nodes with 56221 elements. A close inspection of the mesh reveals that the elements are correctly placed at their boarders without any large differences in side lengths.
Christos Kalavrytinos
Page 10
4.4 Results
The model was solved and straight away it is visible (Fig. 11) that there is a high stress concentration at the ball shafts. However, in this case they are not of particular interested and the focus is on the knuckle only. The strain gauge number 6 reading was exported to Excel and the strain values were averaged with a result of -39 (Fig. 11). The strain component used was the normal elastic strain to the Y axis, since this is the direction that the gauge measures strain on. Table 2 shows the comparison between the strain gauge reading and the Ansys result. The percentage error at 5% is well within a reasonable 10% limit.
The theoretical result from calculations at 11.6 compared with the Ansys test at 39 is within a 70.3% error.
Christos Kalavrytinos
Page 12
Christos Kalavrytinos
Page 13
The mesh was then refined twice until result convergence was evident. The first refinement was a general Body Sizing method where element size was set to 3mm for the whole part. The second refinement (2mm face sizing) was done on the front face where the stress was found to be higher than the original mesh as well as at the point of the strain gauge where the first measurement was taken (10 MPa). The two refinements can be seen in Figures 15 and 16.
Christos Kalavrytinos
Page 14
Christos Kalavrytinos
Page 15
Figure 18, First mesh refinement (3mm body sizing). Christos Kalavrytinos Page 16
5.4 Results
Table 3 shows the results of stress at points of interest on the knuckle. It is clearly evident that the road bump case causes the highest stress on the knuckle. Figure 22 illustrates the stress measures in Ansys. Results seem to converge after two refinements.
Christos Kalavrytinos
Page 18
6.0 Discussion
The maximum tensile strength for the material as given by the brief is 430 MPa. The minimum safety factor is also given as a minimum of 5. Therefore the maximum stress in any of the cases must not be more than 86 MPa. Even at the road bump case, the stress does not exceed 86 MPa with the highest stresses being on the front plate where the hub boss is connected. The physical test results and FEA results are within a 5% error which means that the results for the three cases will be close in an actual driving test with road bumps, braking and cornering. However, the hand calculations are within a 70% error. This can be attributed to the simplifications made to make the calculations possible. Simple mechanics of solids approaches were used. Even though the calculations are simple, there are many things that could influence the stress results. The assumptions made also play a major role in the outcome of the calculations. The goal now is to redesign the part so that its mass is reduced by removing material from low stress areas.
Christos Kalavrytinos
Page 19
7.0 Redesign
The geometry was redesigned to reduce weight from where possible. A total of 60 grams were reduced from the part with the introduction of some holes illustrated in Fig. 23. The stress results can be seen in Fig. 24 and are within the safety factor. There could be further material removal to reduce mass without affecting the safety factor. However, the redesign process is an iterative task and can take a long time and simulations to ensure that each change does not increase the stress beyond the specified point, in this case, 86 MPa.
Christos Kalavrytinos
Page 20
8.0 Conclusion
The model was successfully validated with the use of strain gauges during the physical test. However the hand calculations showed a high error (70%). This can be attributed to the assumptions and simplifications made during the calculations. These were essential due to the complexity of the geometry. The redesigned component helps to show low stress areas where material can be removed from to reduce the mass.
9.0 Recommendations
This work can be improved further with an investigation into more accurate hand calculations. Moreover, using more realistic model constrains for the physical test in the FEA such as remote displacements could provide more accurate results. This is due to the fact that the supporting brackets used were made of aluminium which is elastic compared to steel and could have displaced during the test.
Christos Kalavrytinos
Page 21
The component mass can be further reduced if time is spend finding low stress areas and carefully and gradually applying and measuring design changes and how they affect the stress.
References:
Walker N., Mechanics of Solids Lecture Notes, Birmingham City University website
Appendix
Christos Kalavrytinos
Page 22
Christos Kalavrytinos
Page 23