DYNAMICS - AGENDA

09:00

Welcome address

Helmut Dannbauer, Magna Powertrain Engineering Center Steyr

09:30

Alexander Lion, University of the Federal Armed Forces in Munich

Since several years, car manufacturers of the old school develop vehicles with hybrid or purely electric drive systems. To develop well-adapted engine mounts for such vehicles, it is important to consider the propagation of structure-borne sound in the elastomer part of the bearing. Its transfer path leads from the vibrating engine via the elastomer to the body of the car. Since most of the electric vehicles possess neither manual nor automatic transmissions, the rotational speed of the engine is proportional to the driving velocity of the car. Since typical driving velocities are between 2m/s and 60m/s and the wheel circumferences of passenger cars are about 2m, the rotational speed of the wheels is between 1Hz and 30Hz. Using a constant gear reduction with a ratio of 1/10, the rotational frequency of the engine is between 10Hz and 300Hz. If the electric engine is mechanically unbalanced, which is often the case, excitations with excitation frequencies between 10Hz and 300Hz develop. The internal design of electric engines causes cogging torques and torque ripples whose frequency content contains multiples of the rotational frequency. In the case of synchronous electric motors, the frequency of the cogging torque is linked to the least common multiple of the numbers of poles and slots of the electric motor. If the motor has two poles and six slots on the stator, the least common multiple of both numbers is six such that the cogging frequencies are between 60Hz and 1800Hz. Consequently, the delivered engine torque becomes ripple. For synchronous electric engines with permanent magnets, a main source of noise and vibration is of electromagnetic nature. Some authors show that the frequency range even reaches values of more than 6000Hz. Mechanical imbalances whose characteristic frequency is equal to that of the rotor of the engine lead to inertia force-induced vibrations. Their frequency is thus much lower than that of the electromagnetic excitations but their amplitude increases quadratically with the rotational speed of the engine. In the talk, two mechanical models for elastomer mounts are proposed which are applicable in such situations. To consider the eigenmodes of the rubber part, point masses connected via viscoelasticity elements represent the deformable part of the mount. The frequency range of interest defines the number of the mass points and the dynamic mechanical behaviour of the elastomer the material parameters of the viscoelasticity elements. For comparison, a continuous model of viscoelasticity based on fractional derivatives is also developed. Both approaches are formulated in the time domain and subsequently transferred to the frequency domain. Simulations show the applicability of both approaches to understand and to enhance the dynamic behaviour of rubber bearings for electric vehicles. The point mass model with only one single mass element representing the inertia of the elastomer is a better approach than a model, which ignores the inertia completely. The model with the continuous mass distribution or the lumped mass model with 200 masses lead the best results. ​​​​​​​

10:00
Break

The automotive industry is experiencing disruptive changes with electric, autonomous & connected mobility. Engineers are entering a new phase in powertrain systems – ramping down internal combustion engine design and production and ramping up design and production of electric vehicles. One of the challenges in electric vehicle engineering is to ensure acoustic passenger comfort. Electric drives, comprised of the electric machine and a gearbox, are complex mechatronic systems and their performance in different domains needs to be evaluated to improve and optimize their design. Electric machines and gearboxes can produce tonal noises that are perceived as unpleasant. To analyze the noise and vibration behavior of electric drive systems, multi-physics analysis with linked simulations of the entire system are generally necessary in order to achieve accurate results. In this presentation, a simulation workflow to analyze noise and vibration of an electric drive is presented, including parametric CAD, structural FEM, electromagnetic simulation and focus on multibody system simulation (MBS).

The present study deals with the life correlation of axle housings of Heavy Commercial Vehicles for different loading patterns like vehicle level road loads and in-door bench testing loads with experimental results. Investigation has been carried out to understand the underlying cause for failure before acceptance in bench testing and improvements in life by enhancing the surface finish of housing plates has been proposed.
Design of the axle housing depends upon the vehicle usage pattern and its payload. Therefore, to have the realistic life estimates, the road roughness was captured by measuring the micro strain of the axle housing at vehicle level. Life calculation of housings for transient forces is carried out using FEMFAT. Similar life estimates for the equivalent damage using In-door bench testing loadings has been carried out. Both the design life estimates doesn’t reveal any failure before acceptance, which is contrary to the experimental testing’s results for bench loadings as housing developed crack before acceptance. Failure investigation reveals surface roughness levels increase from design intended 10 μm (max) to 22 μm at press forming bending radius zones created micro notches leading to initiation of cracks. In addition, the tensile strength of the housing plate also lowered from 610 to 500 MPa due to hot working. Life re-estimates carried out by modifying the SN curve to account for the losses reveals a close correlation between different loading patterns captured at vehicle level and bench loadings with experimental results. Life calculation iterations has been carried out by varying the surface finish factor ‘fsr’ but maintaining the tensile losses due to hot working in the component to be constant. It is observed that surface finish enhancement up to 3 μm will make the axle housing plate to meet acceptance life even though the tensile values of plates has been lowered to 500MPa from 610MPa by hot forming process. Enhanced surface finish can be achieved by using abrasive or pressurized sand blasting of housing plates at the required zones. Alternatively, if plate tensile strength is maintained up to 620 MPa after forming, then the surface finish can be lowered as low as 78 μm and still life meets acceptance.
The case study helps the manufacturer to achieve a trade-off between surface finish and the tensile strength as higher surface finish requirement increases the cost of manufacturing. The correlation between different loadings patterns and life enhancement by surface finish helps the designer to improve the fatigue life without undergoing for major design modifications, which adds cost and weight penalty to the component.

11:15
Break

Durability of passenger cars mainly depend on the integration of BIW structure with its sub-components against the road load excitation. To ensure this integration, dynamic response of sub-components need to be accounted during design/validation. In general, the road load excitations generated from Proving Grounds are random cyclic loads in time domain because of different road profiles which the passenger car undergoes. To simulate the dynamic response of sub-components in time domain is a biggest challenge in CAE as it requires a lot of solving time and memory usage. FEMFAT Harmonic is one of the alternative approaches to capture dynamic responses in Frequency domain with good accuracy, less solving time and memory usage. In this study, Frequency response analysis is carried out to determine the natural modes of each sub-component and FEMFAT Harmonic is used to convert the time based road loads into modal based response loads. These modal response loads are then mapped with model stresses and the durability cycle of each sub-component mounting is calculated using FEMFAT Channelmax. As an Output, for a full vehicle CAE simulation accuracy of 100% can be achieved with ~80% reduction in solving time and memory usage using FEMFAT Harmonic. ​​​​​​​

The presentation deals about one question:

Can we win the challenge to develop and validate a next generation forage harvester with…

  • higher load capacity
  • high new content in the structure concept
  • and the requirement to increase the lifetime by 25 %

…while just having one shot for a physical test?” ​​​​​​​

The main project aim is to ensure the fatigue lifetime of the next generation forage harvester‘s main structure for a physical roller test bench in the first shot.

The following process steps are planned:

  • Design review
  • Simulate existing concept and correlate with historical knowledge.
  • Extrapolate the measured loads from a known vehicle to the not yet existing one of the next generation.
  • Develop the structure using virtual loops.
  • Showdown „Roller Test Bench“

Content of the presentation contains the following topics:

  • How did we design the structural concept?
  • Idea & process definition of the validation concept.
  • Which argumentation was used to convince the management to approve the budget for the validation concept?
  • Preparing steps for a collaborative project between CLAAS and ECS Magna Steyr.
  • Current status, first results and outlook of the ongoing process using Virtual Iteration and FEMFAT.
12:30
End

Because of the large motor torque and rapid response of electric vehicles, the mount force is generally large. Therefore, it is necessary to pay more attention to its durability during the mount development process. The mount rubber is difficult to verify the durability due to its super-elastic properties. At present, the durability of mount bushes is mainly verified by test methods, because most of the mount bushing parts tests are based on the GM load cases and related classic load cases, the correlation with the vehicle verification condition is poor, so in previous projects, there were many cases of the bush that passed the component verification, but cracked in the vehicle durability road test, Therefore, there is an urgent need for a set of effective mount verification conditions to fully verify the mount bush in the early stage In this example, the mount load is solved by the virtual iterative method, based on the measured signal of the mount road spectrum, and the iterated load is used in the mount component test, thus effectively identify and avoid risks in the early stage of the project, and avoid failure problems such as cracking of the mount bushing in the later vehicle durability test.

12:30
Break
13:30

Non-linear dynamic simulation of body structures for fatigue analysis

Markus Breitfuss, Magna Powertrain Engineering Center Steyr

The following presentation covers the fatigue life calculation of a swing arm based on a multibody simulation. Strain gages are being used to measure the occurring strains during field studies. Via virtual iteration statically applied forces are being calculated to reconstruct the measured strains. The current calculation process uses those time dependent forces in combination with a Femfat-Max calculation to compute the damage values. Due to the static calculation dynamic effects are being neglected. Therefore, a dynamic approach is needed to take those effects into account. The main goal of this thesis is to compare the existing calculation with a dynamic approach. If possible, the new approach should be used for other components as well. The modal damage analysis is based on X-form files calculated with Adams, the corresponding modal stresses, and the number of repetitions for each file. The results showed that the best solution according to a compromise between computation time and output quality is accomplished by using 80 Hz filtered measurements in combination with the Craig-Chang reduction method and other filter methods to speed up the calculation. Compared to the static analysis no disadvantages were brought forward.

14:30
Break
14:45

Electrification of drive train and autonomous driving are increasing the number of electronic devices in cars and related PCB (Printed Circuit Boards). Due to the dynamic road excitation of such boards and other excitations any damages of the electronics should be avoided. To this end, an increase of the eigenfrequencies of PCB could help to overcome such vibration problems.

One way to increase such eigenfrequencies is to use a sufficient number of bolts to fix the board. These bolts have to be positioned in an optimal way. These tasks lead to a multi-modal optimization, which simultaneously solves both a shape optimization problem to position the supports and a sizing optimization of the support stiffnesses to assess the need of single supports.

The paper explains how the modeling of PCB supports is done and how the optimization task is set up. An example is used to demonstrate the process and to evaluate the results. The analysis and the multi-modal optimization is done in one single job using the industrial software PERMAS.

15:15

Dynamic simulation of electronics and fatigue of solder joints

Harald Ziegelwanger, Magna Powertrain Engineering Center Steyr

15:45
Break
16:00

Complex mechanical systems usually consist of subsystems that interact dynamically with each other. If the subsystems are separated from each other and partly exist as numerical model and as real hardware, then one speaks of hardware in the loop (HIL), cyber-physical or hybrid systems. In this presentation, two recently developed methods for the coupling of experimental and numerical subsystems into an overall system are presented. In both methods, no explicit mathematical models are required and the task of parameter identification is completely omitted.
In the first method, the hardware is characterized with impulse response functions. This leads to a linearized description of the component around an operating point. In addition to the presentation of the basic idea, some examples (e.g. rubber bearings) are shown.
In the second approach (non-simultaneous iterative hybrid simulation/testing), the test bench and the numerical simulation are driven repeatedly in a loop. After each loop, the compatibility conditions are checked. If they are fulfilled, both systems behave according to the cutting force principle as if they were actually coupled. No dynamics are neglected and nonlinearities are mapped. The data exchange is not critical, a communication via EMail would be possible. In addition to the basic idea, a wheel suspension is shown as an example, in which the shock absorber is present as hardware.

16:30

Farewell

Helmut Dannbauer, Magna Powertrain Engineering Center Steyr

16:45
End
08:30

WORKSHOP: Non-linear dynamics of built up structures - simulation approach and software solutions

Markus Breitfuss, Magna Powertrain Engineering Center Steyr

09:30
End