Vertical Loading Phase

Vertical loading of the K275 ROPS model resulted in an under-prediction of the experimental response。 The response of the FE ROPS model appeared to be fairly linear, as seen from Figure 6, with no residual vertical deflection upon load removal and was significantly stiffer than the corresponding experimental response。 The stiffness variation between the two models can easily be distinguished from this figure by examining the gradients of each curve。 The ROPS had undergone plastic deformation with formation of plastic hinges at the top and base of each post during the lateral loading phase。 This would have caused a reduced stiffness in the vertical direction。 It is evident that the numerical model seems to be unable to adequately simulate this reduced stiffness。 Reasons for this lack of correlation were attributed to the inability of the FE plasticity material model to handle (i) influence of load reversal effects such as the Baushinger effect,  (ii)  possible inherent residual stresses in the members from the heavy welding of the members during fabrication and (iii) residual stresses from the previous loading phase。

The Von Misses stress distribution during the vertical loading and unloading stages indicated that high stresses had occurred locally in the beam zone, however, these stresses were still within the elastic limitations of the material。 In addition to this, some localised yielding   had

taken place in the corner region at the connection zone between the beam and the post。 This localised yielding appeared to have had only a minor influence on the ROPS deformation in the vertical direction。

The longitudinal Loading Phase

The load deflection profile for the ROPS under the applied longitudinal load did not compare well with that from the corresponding experimental loading phase。 The reasons for this are probably similar to those for the vertical loading phase, as explained above。 The Von Mises stress distribution showed the presence of distinct yielding at the base of the front and back faces of each post at the point of maximum longitudinal load。 This result was expected, as the ROPS behaved as a two post cantilever which would therefore give rise to high stresses at the base of each post。

Further results from all three loading phases, both experimental and numerical, can be found in Clark (2005)。

Similitude Verification

The scaling laws and relationships between pertinent variables that were established earlier were verified using FEA for the full and ½ scale ROPS models for the K275 bulldozer。 The results from the response of both models when subjected to the loading requirements of the Australian standard confirmed the similitude ratios of ½, ¼ and 1/8 for deflections, loads and energy absorbed。 These values for the ½ scale ROPS model were: lateral load = 188kN, lateral deflection = 70mm, energy absorbed = 12148J, vertical load = 244kN, vertical deflection = 2。55mm, longitudinal load = 96kN and longitudinal deflection = 7。3mm。 The lateral load versus deflection profiles for both models are shown in Figure 12。 The ratio of the maximum deflection of the model to that of the prototype is clearly ½ (70:140), while the ratio of the maximum loads is ¼ (190:760)。 It is evident that the response of a full scale

ROPS can be accurately predicted by using the principles of similitude modelling and implementing an appropriate scaling law。

Post Yield Behaviour of ROPS

As current ROPS performance standards such as AS2294-1997 give limited assistance, designers usually specify overly stiff member sizes to ensure that the prototype ROPS will pass the requirements without premature failure。 Once a successful design has been developed, it may be reproduced without further test and put into production。 Based on this common day approach, it is clear that there is a lack of knowledge on the fundamental principles required for successful ROPS design and it is therefore essential that design guidelines be established to ensure that an optimum level of safety is provided to heavy vehicle occupants during rollover accidents。