FEMFAT LAB Agenda​​​​​​​​​​​​​​​​​​​​​

Introduction of Structural Analysis Department
​​​​​​​Dipl.-Ing. Helmut Dannbauer, Magna Powertrain Engineering Center Steyr

Introduction of VTM & KULI Software
Christian Rathberger, Magna Powertrain Engineering Center Steyr

Introduction of Fatigue Testing and FEMFAT LAB
Johann Traunbauer, Magna Powertrain Engineering Center Steyr

The presentation is about…

the big question: “How to choose measurement sensors and their positions to determine the load data required for the dimensioning of our products?”
A semi-automated process will be presented to define a measurement concept and strain gauge positions to determine the required load data out of measured strains

Knowing the loads that occur in field condition is essential to properly dimension any product. Obtaining this data for the use in a simulation or a test bench, however, is not an easy task.

Therefore, at CLAAS, a process was developed to determine the most relevant loads a given structure experiences during its use which can then be used in the development process for a fatigue analysis or a test bench.

This is not a straightforward task, and the approach can vary based on the subject. Nonetheless, there are some process steps to define and check a possible measurement concept. In iterative steps of defining and checking sensor definitions a feasible measurement concept can be developed.

Apart from the general methodology, two practical examples (Pictures above) how we measure loads in field conditions and their specifics towards the process will be shown.

Accurate load case definition plays a pivotal role in the virtual development of tractors. To effectively predict dynamic behavior and perform strength and mechanical fatigue analysis, it is essential to establish a reasonable correspondence between simulation conditions and realworld field conditions. Forces in the wheel hubs could be investigated with measurement rims. However, this procedure is expensive, and the determined forces refer to only one special tractor parameter set. This means, even a change of a ballast mass makes the measured forces unusable for further simulations. Therefore, developers seek input parameters that are independent of the specific tractor configuration, whereby the field path profile emerges as a promising solution.

The method presented here allows the identification of the road profile traversed by a tractor based on measuring accelerations. The objective is to obtain a profile that, when simulated, produces accelerations in the tractor as close as possible to those generated by the original profile. Starting with a complex nonlinear tractor model in ADAMS, a simplified model with a reduced number of degrees of freedom is derived. 

The simplified model is linearized to a linear time-invariant (LTI) system, whereby the wheel hub forces are the input variables. Under certain conditions, especially in view of number and positions of the acceleration sensors, the LTI system can be inverted, resulting in forces as outputs, and the measurement signals are used as inputs. Knowing forces, state vector, tire stiffness and damping, a differential equation is applied to deduce the actual field path profile under consideration of the tire geometry.

The original accelerations refer to a complex model, whereas the identification is carried out with a simplified LTI model. As a result, the identified profile deviates from the original one. Therefore, a virtual iteration process in time domain was developed so that the identified profile converges to the original one. Finally, three examples will be presented. The first one is the field path detection of the Merry-Go-Round test (MGR). Based on accelerations measured close to the wheel hub centers, the obstacle shape can be identified. The accelerations achieved with the iterated path profile correspond sufficiently accurately with those ones referring to the original profile.

A further test track is the ISO 5008 rougher road. In this case, too, the accelerations owing to the iterated path profile match with those caused by the original one.

The last example is the four-poster test rig, which consists of four hydraulic actuators, each positioned under a wheel of the vehicle. Objective is to achieve the same vertical accelerations as measured in the field. The presented algorithm provides the input signals for the actuators, and the target of reproducing the original acceleration on the test rig is met very well.

A cost-effective and sufficiently accurate method has been established for field path detection based on virtual iteration in time domain. Therefore, field path detection is a good basis to define accurate load cases for finite element analysis.

tbd

FEMFAT LAB virtual iteration, 3D road and model improvement – interfaces to dynamic simulation

Dynamic simulation results depend mainly on the used excitation data and model accuracy. Based on these two inputs, the goal is often to achieve results which correlate with road load data measurements (RLD).

FEMFAT LAB virtual iteration is based on the determination of the excitation of a model in the time domain using multi-body simulation. Using the iteration process with simulation analogous to the real test bench allows adjusting external loadings on a structure in such a way that internal measurements (RLD), i.e., proper load flow, can be reproduced with desired accuracy (solution of a non-linear inverse problem).

FEMFAT LAB 3D road can additionally generate the virtual road based on RLD signals by full vehicle simulations using a tire model.

FEMFAT LAB model improvement allows improving model parameters of an MSC.ADAMS model automatically based on RLD. Typical used signals of RLD in such a process are accelerations, relative displacements or angles, strain gauges, load cells, or wheel force transducers. Important parameters which can be optimized are mass, mass moment of inertia, location of center of gravity, or stiffness characteristics of the connection elements.

Creative invention is the daily challenge and practice of engineering. AI tools seem to replace human creativity in many areas. Will human creativity soon be obsolete? In the talk, I will present three practical creative techniques for human invention and discuss them also in the frame of the AI hype.