PSI - Issue 19

Thorsten Voigt et al. / Procedia Structural Integrity 19 (2019) 4–11

9

Dr.-Ing. Thorsten Voigt/ Structural Integrity Procedia 00 (2019) 000 – 000

6

Vertical load 1kN

Bottom view

Figure 13: failure modes of control arm

Figure 14: half axle test, FEM analysis for control arm

4.2. Test examination Test Specimen #1: In the first quasi static loading of the half axle assembly the rear control arm failed under a static vertical load of 6.4 kN . This vertical load corresponds to a load factor of less than 3, related to a static wheel load of 2,33 kN . Reason for this is the torsional moment due to the distance in y direction (see Figure 12) of the wheel to the connecting point of the spring/damper unit at the control arm. Test Specimen #2: After the failure of the first test specimen the load program was modified applying two measures: 1. removing of all vertical loads higher than 4 kN from the load program. These values represent special events with low incidence. These events should be run at the very end of the test, 2. Reduction of the estimated static wheel load from 2.33 kN to 1.4 kN (60%). Doing so, the amplitudes of all cycles were reduced to 60% of the prior load. Specimen #2 failed after 12% of the modified test programme ( 4.5 h test run of 38.5 h required). The failure mode is very simi lar to the one observed at test specimen #1. Identically to specimen #1 cause of damage was the torsional moment in the control arm. Test Specimen #3: To get more to the bottom of the early component failure in test #1 and test #2 a finite element calculation of the control arm was made. In this calculation a vertical unit load of 1 kN was applied to the wheel hub. The highest strains under this load are identified at the side of the spring/damper suspension, at the bottom side of the control arm’s flange ( Figure 14). This area is assumed to be the starting point for component failure and coincides with the fracture behavior of specimens #1 and #2. To investigate these theoretical results in practice strain gauges were applied at the highly stressed regions of the control arm. With quasi-static strain-matching with slow harmonic vertical load it could be determined that the measurements validate the results of the FEM calculations. For the last specimen the estimated static wheel load was further reduced from 1.4 kN (specimen #2) to 1 kN . The test was run with the further reduced load signal. Component failure occurred this time after 44% of the test run. The failure mode is similar to those observed at test specimen #1 and #2 (see Figure 13). 4.3. Comparison to service like loading and damage accumulation calculation The initial load program for the tests was derived basing on a first weight distribution calculation of the URBAN-EV vehicle for a maximum static rear wheel load of 233kg. Due to the critical results of the fatigue test at the rear half axle a new weight distribution calculation was carried out with the result, that the maximum static rear wheel load in the end configuration of the vehicle will be 200 kg. The load program was therefore adjusted from 233 kg to 200 kg static wheel load. The following analyzes are based on this reduced value. In Figure 15 (left axis) the endured vertical load test spectra of both tests (#2 and #3) are compared to the vertical load design spectra with and without considering the damage increase due to corrosion influence. More over the safety factor F is given as the quotient of damage sum test spectra and damage sum design spectra. With a safety factor clearly below 1 it becomes evident that the structure cannot bear the loads of the nominal load program for this vehicle. Reason for this is the torsional moment which is introduced into the control arm under vertical load. To validate the tests and to get information on the material cyclic fatigue strength at the failure area numerical and additional experimental stress analysis were examined . With a maximum strain of ε 1 = 1.25·10 -3 at 1 kN vertical load, compare Figure 14 and a Young’s modulus of 71000 GPa this would mean a maximum principle stress of around 89 MPa at 1 kN vertical load. With this value the force spectra of both test were transfered into stress spectra, see Figure 15 (right axis). With an estimated Woehler curve (R=-1) for AlSi7Mg with a slope of k=6 and a knee point at a = 60 MPa / = 5 ∙ 10 6 cycles and with an estimated mean stress