First, car body model is established by using CATIA software and then using the software Hyper Mesh reads it and converts into hm format, at the same time goes geometry repair and meshing by the software Hyper Mesh [4] .

The right material properties for the simulation result have important influence in the vehicle collision experiment. The vast majority of material data are obtained through experiments in the vehicle finite element model [5] . The main part of car is composed of sheet metal parts, so elastic-plastic mechanics characteristics material MAT340 is adopted.

Another type of collision components is almost not deformable, such as engine, transmission etc. They can be chosen MAT20 material model [6] . SDMAT1 material model can be used to simulate elastic damping. MAT_ SOPTWELD material model can be used to simulate solder material.

In the process of collision force transfer is through the connection between parts and the parts [7] . In the finite element model the common connection ways mainly have the following: The connection between the sheet metal parts are spot welding connections in body parts; In the model, the bolt connection mainly adopts rigid connection unit, rigid panel function is used to make the two bolt holes coupled together; In order to keep the body outer surface smooth and complete, the connection between the parts and the outer plates uses glue and edging treatment; Part of the chassis, a lot of parts are moving contact, there are several common connections: Ball connection, universal joint connection etc.

In order to prevent the penetration phenomenon, overall vehicle finite element model is defined as a general- sided contact [8] . Reference “C-NCAP Management Rules (2012 Edition)” the initial velocity is set to 50 km/h. Gravity is set to 9.8 m/s². Car collision time is set 120 ms. Finite element model is shown in Figure 1.

From Figure 2, the total energy line keeping level indicates that energy is conserved in the process of collision. Hourglass energy with the ratio of the total energy is 1.46% less than 3% of the value of industry regulations. The increasing quality with the ratio of the initial vehicle quality ratio is 3.46% less than 5% of the value of industry regulations. Therefore, the credibility of the entire simulation meets the requirement. The smooth transition curves meet simulation requirement.

In Hyper View post-processing software, vehicle deformation can be seen at the different time after d3plot file importing. Car collision deformation in the end is shown in the Figure 3.

The front energy absorption components of the car deforming sufficiently can be seen from the Figure 3. Hood is completely bent, tire contacts with fender, bumper and the front longitudinal beam deformation fully. The above phenomenon can guarantee energy absorption in the process of the collision.

On both sides of the door and a column don’t have obvious deformation that can help the crew to open the door to escape. The tank doesn’t have large deformation that can prevent the car getting fire. The floor and the exhaust pipe are slightly bent down. The back of the car body has obviously risen.

The front longitudinal beam is the main kinetic energy absorber during crash simulation of the vehicle. The ideal front longitudinal beam deformation is that the front of the deformation sufficient and the rear of the deformation slight, so that the car can efficiently absorb the kinetic energy, avoid high acceleration and prevent dash panel invading largely. It can provide safe space for driver and front passenger. Figure 4 shows front longitudinal beam collision deformation during different times.

As can been seen from Figure 4 that the front longitudinal beam deform from the front to the rear with the passage of the time. The front of the front longitudinal beam first happening completely crushing is helpful for

the absorption of energy. Over time the center and rear the front longitudinal beam appearing different degree of bending is harmful to the dash panel intrusion.

The dash panel displacement contours can show the direct invasion of the amount of the dash panel to the cab. The amount of the invasion is related to living space in the collision process, so the smaller amount of the invasion is beneficial to provide enough space for the front passenger. Figure 5 is the dash panel displacement contours relative to the cab. From the picture we can know that the amount of the invasion in the middle of the dash panel is 38.1 cm that has greatly threatened the survival of the front passenger.

The body after B-pillar deformation basic does not occur and B-pillar acceleration deceleration can be a good representative of the true passenger feel in the process of collision. So B-pillar acceleration deceleration is usually used as an evaluation index of the vehicle acceleration response. The acceleration curves of the right and left B pillar during collision shown in Figure 6. The car collision lasts up to 87 ms, the right B pillar maximum acceleration is 36.3 g at the time of 54.8 ms, and the left B pillar maximum acceleration is 40.1 g at the time of 71.1 ms.

The main energy-absorbing structure is front longitudinal beam in the collision. The deformation of front longitudinal beams is mainly to the mixture of crushing and bending mode. Figure 7(a) is the picture of the front lon- gitudinal beams deformation before improvements. The picture shown that collision deformation is serious, total energy absorbed 31,520 J, 19.7% of the total energy.

During the collision, the front of the front longitudinal beams crush deformation and the rear of it bending deformation occurs, so that energy absorption efficiency is low. In order to prevent bending deformation, the thickness of the front longitudinal beam increase from 1.8 mm to 2.5 mm. Figure 7 shows that bending deformation is weaken and crush deformation is enhanced after improvement. Meanwhile, the energy absorption increased to 37,920 J, accounting for 23.7% of the total energy, there has been a significant improvement. In reducing the front longitudinal beams bending deformation and enhancing crushing deformation, at the same time, in order to effectively reduce the dash panel invasion and increase the front passenger living space the thickness

of the dash panel increases from 1.0 mm to 1.5 mm. As shown in Figure 8, the amount of the dash panel invasion deduces from 38.1 cm to 27.2 cm, increasing the front passenger living space.

Although the local collision performance improved, but not the car collision don’t completely improve. As shown in Figure 9, A-pillar has a slight deformation at the end of collision. It impedes opening the front door for escaping, while the amount of the dash panel invasion is still too larger.

1) Absorbing box structure can be added at the back of the bumper, so that the absorption performance can be further improved. It can reduce the amount of the dash panel invasion.

2) The strength of reinforcing ribs which inside the front door is increased, which the strength of the A-pillar is increased, so that the chance of escaping is increased.