**4. Results and discussion**

Once the full model of the vehicle side-impact is defined and implemented in LS-Dyna™ tests were made considering the vehicle door with the paddings made from PUF, MAC and IMP cellular materials. An additional test was made with the door with no padding, for the sake of comparison. The resultant values of kinetic energy, loads, accelerations and absorbed energy were registered. From these results it is possible to evaluate the relative performance of each material used in the side-impact padding. A detailed discussion of these results is presented in the following paragraphs.

## **4.1. Load distribution**

The evolution of the reaction force on the moving rigid wall (the propellant of the deformable barrier in the impactor vehicle) is computed and its dependence with time is registered and plotted on Figure 6. When the rigid wall first contacts the deformable barrier, a force peak is registered at instants *t* ≈ 0.5 ms for all the simulations performed. This occurs in an early stage of the crash test when no intrusion has yet happened, not even contact with the vehicle door. Thus, this event shall not be considered relevant for the analysis and should be considered a side-effect of the modelling approach used. After *t* ≈ 8 ms, it can be verified that the the reaction force evolution is, as expected, considerably more unstable for the simulation with no padding in the lateral door than for the remaining tests (using paddings with cellular materials). This fact suggests that there is a significant improvement on the behaviour of the vehicle in terms of protection in side-impact collisions when a lateral padding is applied.

Furthermore, from *t* ≈ 90 ms on the load almost goes back to zero on all numerical simulations. This is a consequence of the separation of the rigid wall from the impactor. During this stage, however, the system exhibits rigid body motion and, thus, the results obtained for times *t* > 90 ms are not considered relevant for the scope of this research.

It can also be clearly observed that the average load during the considered time interval is considerably lower (≈ 50kN), one order of magnitude, for all the crash tests including a lateral padding when compared the one observed for the test with no padding (≈ 500kN). This leads to the conclusion that the implementation of a padding, either PUF, MAC, IMP or ALF, leads to a much lower and smoother distribution of the load for the whole duration of the impact. The average force obtained for the simulations with the protection padding was around 85% lower than the ones obtained without padding and the maximum force was up to 79% lower.

**Figure 6.** Evolution of the reaction force on the moving rigid wall with time for the simulations of side-impact crash tests.

#### **4.2. Kinetic energy**

10 Will-be-set-by-IN-TECH

approach in order to describe all three different cellular materials. This material is generally adequate for honeycomb and foam materials with anisotropic behaviour [11, 25, 27]. This modelling approach assumes zero value for the Poisson ratio and considers a variable elastic modulus, increasing linearly from the initial value as a function of the relative volume (i.e. the ratio of the actual volume to the initial volume) up to the fully compacted material modulus. According to the EuroNCAP standards [38] the impacting vehicle (see Figure 4) must be modelled considering deformable 3030 and 5052 aluminium honeycomb blocks. These blocks should be attached to a mobile structure that is to be considered rigid. Within the scope of this work the deformable blocks were modelled using eight-node hexahedral finite elements. The material behaviour was once again described by the \*MAT\_HONEYCOMB constitutive model, for the same reasons stated for the padding materials. The necessary material properties were determined considering both honeycombs [14]. The propeller structure is, however, fairly complex according to EuroNCAP regulations and it is not fully described on the side-impact protocols available. Nonetheless, the total weight of the impactor system is known and, ultimately, the geometry of the impactor moving structure is not significantly relevant for the conditions of the performed tests. For this reason this moving structure was modelled as a moving rigid wall. The most determinant feature is the part of the impactor colliding with the vehicle's lateral door, i.e., the deformable aluminium honeycomb blocks. Hence, mass was added to the impactor's anterior part and the initial velocity of the crash tests (*v*<sup>0</sup> = 13.89 m/s)

Once the full model of the vehicle side-impact is defined and implemented in LS-Dyna™ tests were made considering the vehicle door with the paddings made from PUF, MAC and IMP cellular materials. An additional test was made with the door with no padding, for the sake of comparison. The resultant values of kinetic energy, loads, accelerations and absorbed energy were registered. From these results it is possible to evaluate the relative performance of each material used in the side-impact padding. A detailed discussion of these results is presented

The evolution of the reaction force on the moving rigid wall (the propellant of the deformable barrier in the impactor vehicle) is computed and its dependence with time is registered and plotted on Figure 6. When the rigid wall first contacts the deformable barrier, a force peak is registered at instants *t* ≈ 0.5 ms for all the simulations performed. This occurs in an early stage of the crash test when no intrusion has yet happened, not even contact with the vehicle door. Thus, this event shall not be considered relevant for the analysis and should be considered a side-effect of the modelling approach used. After *t* ≈ 8 ms, it can be verified that the the reaction force evolution is, as expected, considerably more unstable for the simulation with no padding in the lateral door than for the remaining tests (using paddings with cellular materials). This fact suggests that there is a significant improvement on the behaviour of the vehicle in terms of protection in side-impact collisions when a lateral padding is applied.

was assigned to the structure.

**4. Results and discussion**

in the following paragraphs.

**4.1. Load distribution**

The evolution of the kinetic energy of the whole system with time is represented on Figure 7. A sudden increase of the kinetic energy can be observed for *t* ≈ 3 ms for the crash test with no padding. This corresponds to the instant at which the impactor structure initiates contact with the door of the vehicle. Additionally, for all the simulations performed the curves exhibit an inflection near the final stage of the impact (*t* ≈ 80 ms), increasing from this instant until the end of the numerical simulation. From the visualisation of the kinematic results it is possible to relate this inflection to the beginning of rigid body motion of the door (and vehicle) after impact, where incremental deformation ceases to exist [40]. Hence, for the purpose of this investigation the parameters studied will only consider the instants between these marks (that is 3 < *t* < *t*<sup>r</sup> ms), where *t*<sup>r</sup> is the time when rigid body motion initiates. This time instant will be considered different according to each simulation: (i) *t* < *t*<sup>r</sup> ≈ 80 ms for the simulation without padding, (ii) *t* < *t*<sup>r</sup> ≈ 90 ms when using the polyurethane foam padding, (iii) *t* < *t*<sup>r</sup> ≈ 85 ms when using the IMPAXX™ padding, (iv) *t* < *t*<sup>r</sup> ≈ 75 ms for the micro-agglomerated cork padding and (v) *t* < *t*<sup>r</sup> ≈ 90.5 ms when using the aluminium foam padding.

**Figure 8.** Dependence with time of the acceleration and displacement of the the door's interior structure in the direction of the occupant's pelvic area, obtained from the numerical simulations of the side-impact

On the Use of Polyurethane Foam Paddings to Improve Passive Safety in Crashworthiness Applications 349

The resulting evolution in time of the acceleration and displacement measured on the internal structure of the lateral vehicle's door, in the direction of the pelvic area of the occupant, is shown on Figure 8. A SAE-180 filter was used to refine the acceleration results as advised by the protocols used by EuroNCAP [38] for acceleration measurements on the occupants' pelvic area. Very high maximum values of acceleration can be observed at the initial instants of the simulation for all the tests that included a padding, as opposed to the ones observed for the simulations with no padding. This fact is most probably due to the added stiffness of the padding on the first instants of the crash. However, this happens for the early stages of the impact when the displacement of the lateral door's interior is still inexistent and, consequently, there is no contact between the door and the occupant. Thus, the maximum acceleration values were determined only starting from the moment when the displacement

The results obtained prove that the use of an interior padding can lead to a significant decrease in the maximum acceleration felt by the occupants. This decrease can be as high as 59%. While the maximum acceleration value for the simulations with no padding is *a*<sup>0</sup> = 6.1 mm/ms2, acceleration peaks of *aPUF* = 4.8 mm/ms2, *aIMP* = 3.9 mm/ms2, *aMAC* = 3.5 mm/ms<sup>2</sup> and *aALF* = 2.5 mm/ms<sup>2</sup> were observed for the crash tests with PUF, IMP, MAC and ALF padding, respectively. Thus, the inclusion of a polyurethane foam padding results in a reduction of roughly 20% even though it is not the best performing material among the ones investigated.

crash tests.

of the door initiates.

**Figure 7.** Evolution of the system's kinetic energy with time for the simulations of side-impact crash tests.

#### **4.3. Maximum acceleration**

An anthropomorphic dummy was not considered or modelled during the finite element analyses carried out during this research because of the added complexity and Central Processing Unit Time (CPU) time it would bring. Hence, for the purpose of evaluating the maximum acceleration (or maximum deceleration) resulting from the impact felt on the passenger's pelvis area, as well the intrusion level in the passenger compartment, numerical results on nodes on the pelvis direction, on the inner side of the vehicle door, were analysed and are discussed herein.

12 Will-be-set-by-IN-TECH

end of the numerical simulation. From the visualisation of the kinematic results it is possible to relate this inflection to the beginning of rigid body motion of the door (and vehicle) after impact, where incremental deformation ceases to exist [40]. Hence, for the purpose of this investigation the parameters studied will only consider the instants between these marks (that is 3 < *t* < *t*<sup>r</sup> ms), where *t*<sup>r</sup> is the time when rigid body motion initiates. This time instant will be considered different according to each simulation: (i) *t* < *t*<sup>r</sup> ≈ 80 ms for the simulation without padding, (ii) *t* < *t*<sup>r</sup> ≈ 90 ms when using the polyurethane foam padding, (iii) *t* < *t*<sup>r</sup> ≈ 85 ms when using the IMPAXX™ padding, (iv) *t* < *t*<sup>r</sup> ≈ 75 ms for the micro-agglomerated

cork padding and (v) *t* < *t*<sup>r</sup> ≈ 90.5 ms when using the aluminium foam padding.

**Figure 7.** Evolution of the system's kinetic energy with time for the simulations of side-impact crash

An anthropomorphic dummy was not considered or modelled during the finite element analyses carried out during this research because of the added complexity and Central Processing Unit Time (CPU) time it would bring. Hence, for the purpose of evaluating the maximum acceleration (or maximum deceleration) resulting from the impact felt on the passenger's pelvis area, as well the intrusion level in the passenger compartment, numerical results on nodes on the pelvis direction, on the inner side of the vehicle door, were analysed

tests.

**4.3. Maximum acceleration**

and are discussed herein.

**Figure 8.** Dependence with time of the acceleration and displacement of the the door's interior structure in the direction of the occupant's pelvic area, obtained from the numerical simulations of the side-impact crash tests.

The resulting evolution in time of the acceleration and displacement measured on the internal structure of the lateral vehicle's door, in the direction of the pelvic area of the occupant, is shown on Figure 8. A SAE-180 filter was used to refine the acceleration results as advised by the protocols used by EuroNCAP [38] for acceleration measurements on the occupants' pelvic area. Very high maximum values of acceleration can be observed at the initial instants of the simulation for all the tests that included a padding, as opposed to the ones observed for the simulations with no padding. This fact is most probably due to the added stiffness of the padding on the first instants of the crash. However, this happens for the early stages of the impact when the displacement of the lateral door's interior is still inexistent and, consequently, there is no contact between the door and the occupant. Thus, the maximum acceleration values were determined only starting from the moment when the displacement of the door initiates.

The results obtained prove that the use of an interior padding can lead to a significant decrease in the maximum acceleration felt by the occupants. This decrease can be as high as 59%. While the maximum acceleration value for the simulations with no padding is *a*<sup>0</sup> = 6.1 mm/ms2, acceleration peaks of *aPUF* = 4.8 mm/ms2, *aIMP* = 3.9 mm/ms2, *aMAC* = 3.5 mm/ms<sup>2</sup> and *aALF* = 2.5 mm/ms<sup>2</sup> were observed for the crash tests with PUF, IMP, MAC and ALF padding, respectively. Thus, the inclusion of a polyurethane foam padding results in a reduction of roughly 20% even though it is not the best performing material among the ones investigated.

#### **4.4. Energy absorption**

The dependence of time of the energy absorbed by the structure of the vehicle is plotted on Figure 9. Analysing these results, it becomes clear that the inclusion of a polyurethane foam padding inside the door structure is the best way to increase the capability to absorb impact energy during a side-impact, when compared to both the crash test numerical simulations with no padding and those using paddings made from other cellular materials. The energy absorbed by the structure with the PUF padding exhibits a higher dissipation capability, leading to an increase in energy absorption of approximately 13% when compared to the structure to no padding.

as energy absorbers was confronted with the results with no padding. The results obtained show that the implementation of a foam like material — a cellular material — as a padding for energy dissipation in lateral doors can, in fact, lead to considerable improvements, mainly in terms of maximum values of deceleration (the direct consequence leading to injury levels)

On the Use of Polyurethane Foam Paddings to Improve Passive Safety in Crashworthiness Applications 351

Reductions as high as 59% in terms of maximum acceleration values can be observed when comparing the results obtained with and without padding. This reduction was achieved by implementing an aluminium foam padding. This was followed by a cork micro-agglomerate padding, with an improvement of 43%, and IMPAXX™ with 36%. A padding of rigid polyurethane foam, even though it is the one leading to a smaller reduction, can result in

The average loads in the crash tests with padding are more than 85% lower than the ones from the tests with no padding and its distribution is more balanced. Additionally, the maximum load could also be reduced by up to 79% when including a protective padding, being the best results obtained with cork micro-agglomerate. Polyurethane foam padding enclosed inside the vehicle's lateral door resulted in reductions of 83 and 73% in terms of average and

Furthermore, in terms of energy absorbed by the vehicle's global structure, polyurethane foam was the material exhibiting the best behaviour. The inclusion of this padding, as well as

micro-agglomerated cork padding, resulted in improvements of approximately 13%.

*Faculty of Engineering & Industrial Sciences, Swinburne University of Technology, Australia*

[1] [n.d.]. LS-Dyna™ (971) [Software]. (2008). Livermore, CA, Livermore Software

[2] Anindya, D. & Shivakumar, N. D. [2009]. An experimental study on energy absorption behavior of polyurethane foams, *Journal of Reinforced Plastics and Composites*

[3] Automotive, D. [2006]. *Tech Data Sheet IMPAXX™ 300 Energy Absorbing Foam*, The Dow

[4] Avalle, M., Belingardi, G. & Montanini, R. [2001]. Characterization of polymeric structural foams under compressive impact loading by means of energy-absorption

[5] Belytschko, T., Lin, J. & Chen-Shyh, T. [1984]. Explicit algorithms for the nonlinear dynamics of shells, *Computer Methods in Applied Mechanics and Engineering* 42(2): 225–251.

diagram, *International Journal of Impact Engineering* 25: 455–472.

*Dept. Mechanical Engineering, University of Aveiro, Portugal*

maximum accelerations 21% lower when compared to tests without padding.

and loads transmitted to the occupants of the vehicle.

maximum load, respectively.

**Author details** Mariana Paulino

Filipe Teixeira-Dias

**6. References**

28: 3021–3026.

Chemical Company.

Technology Corporation.

**Figure 9.** Variation with time of the energy absorbed by the structure of the vehicle during the side-impact crash tests.

#### **5. Conclusions**

The introduction of a structural padding made from cellular materials inside the lateral doors of common vehicles is suggested as a passive safety mechanism in side-impact vehicle-to-vehicle collisions. In order to evaluate the viability and efficiency of this safety mechanism crash tests were performed using finite element analysis software LS-Dyna™. The EuroNCAP [38] standards and definitions were considered when the defining and implementing the crash-test models. Rigid polyurethane foam, IMPAXX™ , micro-agglomerated cork and aluminium foam paddings were tested and their performance as energy absorbers was confronted with the results with no padding. The results obtained show that the implementation of a foam like material — a cellular material — as a padding for energy dissipation in lateral doors can, in fact, lead to considerable improvements, mainly in terms of maximum values of deceleration (the direct consequence leading to injury levels) and loads transmitted to the occupants of the vehicle.

Reductions as high as 59% in terms of maximum acceleration values can be observed when comparing the results obtained with and without padding. This reduction was achieved by implementing an aluminium foam padding. This was followed by a cork micro-agglomerate padding, with an improvement of 43%, and IMPAXX™ with 36%. A padding of rigid polyurethane foam, even though it is the one leading to a smaller reduction, can result in maximum accelerations 21% lower when compared to tests without padding.

The average loads in the crash tests with padding are more than 85% lower than the ones from the tests with no padding and its distribution is more balanced. Additionally, the maximum load could also be reduced by up to 79% when including a protective padding, being the best results obtained with cork micro-agglomerate. Polyurethane foam padding enclosed inside the vehicle's lateral door resulted in reductions of 83 and 73% in terms of average and maximum load, respectively.

Furthermore, in terms of energy absorbed by the vehicle's global structure, polyurethane foam was the material exhibiting the best behaviour. The inclusion of this padding, as well as micro-agglomerated cork padding, resulted in improvements of approximately 13%.
