PSI - Issue 58

J.R. Steengaard et al. / Procedia Structural Integrity 58 (2024) 61–67 J.R. Steengaard et al. / Structural Integrity Procedia 00 (2024) 000–000

62

2

a

b

Fig. 1. (a) Tractor with a disc mower mounted; (b) Cutterbar used in agricultural disc mowers. Pictures kindly provided by Kverneland Group Kerteminde A / S.

model the mower cutterbar accurately such that the model can be used for fatigue assessment. This can replace long and costly fatigue experiments. In this paper, a quasi-static experiment has been performed on a cutterbar. Experimental strain values are used as basis for validating and updating the finite element (FE) model. Results from FE analyses can be inconsistent with experimental results. This constitutes a problem, when using the FE model for fatigue assessment. These inconsistencies can be caused by incorrect modelling of boundary conditions, joints, material properties, and geometry (Arora (2011)). The errors can be reduced using model updating methods. These methods are either direct or iterative. Direct methods reproduce the experimental results exactly and are less computational demanding. Direct methods also violate structural connectivity, which complicates the interpretation of the results (Arora (2011)). Iterative methods yield easily interpretable results as the engineer chooses the parameters to be updated, but they are more computationally demanding (Arora (2011)). A parameter based model updating method is used, and the updated FE model is shown to be accurate. Following the guidelines by the International Institute of Welding (IIW) (Hobbacher (2016)), the hot spot approach is used on the updated FE model. Fatigue assessment is performed using the IIW fatigue criterion (Hobbacher (2016)). Of the three weld locations considered, the back weld is shown to be most exposed. A quasi-static experiment has been performed on a cutterbar at Kverneland Group Kerteminde A / S. The purpose of the experiment was to check if the cutterbar acts linear and to obtain measurements to validate the FE model. The experimental setup can be seen in Fig. 2. The cutterbar is mounted on two special brackets. These brackets are welded to a large and thick steel table, which ensures a high degree of rigidity. A hydraulic actuator is mounted in the middle of the cutterbar. The actuator is capable of exerting a transverse force on the cutterbar beam that resembles the direction of the upward pressure on the beam from the ground. A total of 20 strain gauges and two wire potentiometers have been attached to the cutterbar to record strains and deflections during the experiment. The placement and numbering of the strain gauges and wire potentiometers can be seen in Fig. 3. Three of the strain gauges are rosette strain gauges to capture the entire strain state at given locations. This includes strains in other directions than the uniaxial strain gauges, which are all in the same direction. The 17 uniaxial strain gauges, three rosette strain gauges, and two wire potentiometers make a total of 28 channels for data acquisition. The strain gauges are from Tokyo Measuring Instruments Laboratory (TML) part of the WF waterproof series with a gauge length of 3 mm. Most strain gauges are placed on the bottom side of the cutterbar, as higher strains are expected here, and due to simpler attachment. Strain gauges 7, 16, 17, and 18-19-20 (see Fig. 3) are located near welds of interest to measure the strain in these areas. These gauges are called local strain gauges. During the experiment, the load from the actuator is increased stepwise until a certain large deflection, see Fig. 4. This approach is used to show that the cutterbar is acting linear as expected. Furthermore, a preliminary FE analysis showed no yielding at the applied load. The strain and deformation measurements are averaged over a short duration of time at the maximum load to reduce noise, which is present mainly in the strain channels. As seen from the deflection 2. Experimentation