1. Introduction
The ongoing evolution of modern vehicle systems toward enhanced protection abilities, multi-functional capabilities, and high intelligence has led to a continuous increase in the overall weight of these systems. While developing heavy vehicles weighing over 2 tons, there is also an urgent need to research the accompanying airdrop stabilization and safety technology [
1]. In the past, a lot of previous studies have studied the stability of the projectile–parachute system before landing based on dynamic simulations and experiments [
2,
3,
4,
5,
6,
7,
8,
9]. Doherr et al. [
10] developed a 9-degree-of-freedom (DOF) computer program for the simulation of the trajectory and the dynamic behavior of the parachute system. Dobrokhodov et al. [
11] built up a 6-degree-of-freedom dynamic model of a guided circular parachute. Guglieri [
12] established the dynamics model of a parachute–payload system during the terminal deceleration phase. The results presented for a realistic case study confirm the validation of their simulation outputs. Neuhaus et al. [
13] developed a Generic Multibody Parachute Simulation Model to completely analyze the performance of the vehicle. Avanzini et al. [
14] conducted the multibody analysis of the system formed by an entry vehicle and a parachute. They also investigated the most important factors that characterize the final phase of an entry trajectory. Gao et al. [
15] developed a new fluid–structure coupling model to predict both the opening phase of parachute and the trajectory of payload during airdrop mission in low altitude. The opening performances of the parachute at different airdropping velocities are analyzed and compared with the experimental results. The results show that, in a finite mass inflation scenario, their coupling model is efficient enough to predict the dynamics behaviors of parachute–payload system. Gao et al. [
16] investigated the inflation behavior of a disk-gap-band parachute using the arbitrary Lagrangian–Euler (ALE) penalty coupling method. The results of the airdrop test demonstrate the validation of their model.
However, landing anti-rollover technology serving as a critical stability metric and directly characterizes the dynamic equilibrium of airdrop systems during touchdown has not triggered too much attention. Up to now, the current airdrop systems exhibit significant sensitivity to environmental variables. Furthermore, existing passive anti-rollover devices demonstrate insufficient attitude adaptability to match complex operational environments. According to statistics, the current probability of an airdrop rollover is about 6%.
In recent years, the most widely used method to improve the landing stability of the airdrop vehicle is the application of the airbag [
17,
18]. Some previous studies tried to analyze the airbag’s buffer characteristics through finite element simulation and experiment [
19,
20]. For example, Li et al. [
21] investigated the cushioning characteristics of self-inflating airbags through indoor drop tests. Results showed that the maximum internal pressure in the airbag increases proportionally with both landing velocity and payload mass while peak acceleration is directly proportional to landing velocity but inversely proportional to payload mass. Focusing on an omni-directional airbag with multi-chamber, Wen et al. [
22] built up a validated finite element model to simulate its soft-landing process. Key design parameters were analyzed for their effects on cushioning performance. Results indicated that reducing initial pressure and increasing airbag/vent diameters significantly lowered peak acceleration while diaphragm hole diameter had minimal impact. Wang et al. [
23] conducted the multi-objective optimization of an airbag landing attenuation system for heavy airdrop based on the multi-dimensional response surfaces method. The results show that the optimization method presented in their paper is a practical tool for the optimization of an airbag landing attenuation system for heavy airdrops. Lian et al. [
24] conducted a study on the cushioning performance during airdrop scenarios, utilizing LS-SDYNA to analyze the impacts on both rigid and soil ground surfaces. The findings revealed that when the airdrop system impacts the ground strictly in a perpendicular manner, the disparity in the maximum overload experienced by the cargo between the rigid ground and soil ground cases is minimal. Yang et al. [
25] addressed the challenges of high cost, long duration, and high risk in airborne vehicle airdrop testing by developing a dynamic finite element model to simulate the landing buffer process of a vehicle-airbag system. They found that under normal conditions, the vehicle’s top deck experienced the highest average peak acceleration (11.31 g) with structural stress meeting design requirements. Liu et al. [
26] proposed an approximation non-linear interval number programming (A-NINP) method to obtain the optimal buffering characteristics of a landing airbag in manned airdrop. The results indicate that their method can lead to better buffering characteristics of landing airbags and ensure the safety of astronauts. He et al. [
27] proposed a hierarchical updating method for finite element model of airbag buffer system under landing impact. They validated their method through experiment. Fu et al. [
28] developed a finite element (FE) model of an airborne armored vehicle-airbag system to simulate the landing impact process. Their findings indicated that when the vertical landing speed is less than 8 m/s, the vehicle body does not experience plastic strain and its strength satisfies the airdrop requirements. Zhang [
29] built up a thermodynamic analytical model as well as a finite element model (using the CV method) to study the landing buffer characteristics of an unmanned airdrop rescue vehicle. The results demonstrated that the airbag buffer effectively reduced system overload from 35 g to 22 g, achieving a significant 37.1% reduction in peak acceleration and a 45.2% decrease in cargo platform stress. He et al. [
30] evaluated the cumulative damage to an airdrop vehicle hull during landing by integrating virtual prototype technology, finite element analysis, and the Lemaitre damage model. Results revealed a linear relationship between the damage variable and the number of landings. Fu et al. [
31] built up a finite element method to analyze the landing buffering adaptability of a manned armored vehicle. They validated their method through a typical airdrop experiment. According to their simulation, the airbag buffer system can help manned armored vehicle realize an effective buffer when the vertical landing speed is less than 9 m/s at an altitude up to 3000 m.
Apart from evaluating the buffer characteristics based on the overload, there are also several studies that shed light on the dynamic stability of the vehicle with airbag buffering system [
32,
33]. Zhang et al. [
34] developed a 10-degree-of-freedom dynamics model using Kane’s method to analyze the vibration response of an airdrop vehicle during landing. Key findings indicate that the suspension and wheel compressions are primary vibration sources with suspension impacts at maximum displacement causing significant acceleration spikes. Niu et al. [
35] investigated the cushioning characteristics of a double-chamber airbag system for heavy equipment recovery by developing an analytical model integrating dynamic equations, thermodynamic principles, and exhaust flow equations. Results indicate that increasing initial pressure reduces peak impact overload but raises landing velocity. Also, a higher secondary-to-primary chamber volume ratio lowers peak overload at the cost of increased landing speed.
Despite the above studies, most of the previous work only involves in the ground test or simulation, the uncertainty and complexity of the wind field in the actual airdrop process have not been taken into account. In particular, the key factors causing the landing rollover phenomenon are unclear. There is a lack of quantitative landing stability evaluation methods, which cannot effectively guide the design of anti-rollover devices for airdrop systems.
In view of the above situation, this article starts with dividing the landing impact process of the vehicle into two stages in
Section 2: the airbag cushioning stage and the rigid collision stage. In the airbag cushioning stage, a vertical impact test bench and a fluid–structure interaction (FSI) model is built up to obtain the terminal impact velocity in typical airdrop mission. In addition, an oblique impact test bench and a dynamic model is proposed in the following rigid collision stage. After the impact mechanism analysis, a terminal sideslip angle active control system is built up to help increase the vehicle’s landing attitude stability. In
Section 3, this work successfully achieves the terminal impact velocity through both the experiment and the fluid–structure iteration (FSI) model. Then, through the dynamic simulation model and the corresponding experiment, the effect of the terminal sideslip angle and the terminal impact velocity on the vehicle’s roll/pitch angle is uncovered. Afterwards, at the landing time, the angles between the vehicle’s horizonal velocity vector and its longitudinal axis are regulated to around 0°/180° under the assistance of the terminal sideslip angle active control system. Finally, several conclusions are drawn in
Section 4.
In general, this work studies the influencing factors of landing rollover in heavy vehicle airdrop systems (e.g., jeep or truck) and researches the active attitude adjustment anti-rollover control technology, laying a technical foundation for the engineering design and application of adaptive, highly reliable active attitude adjustment anti-rollover devices for heavy airdrop system during their airdrop process.
2. Materials and Methods
2.1. Experimental and Numerical Method in Airbag Cushioning Stage
To study the rollover phenomenon caused by the impact of the vehicle on the ground at the end of the airdrop, this paper tries to conduct experiments and simulation analyses in two steps: the airbag cushioning stage and the rigid collision stage.
In the first stage, the vehicle assembled with the airbag touches the ground and the gas inside the airbag is rapidly compressed. When the air pressure inside the airbag exceeds the predetermined pressure of the exhaust hole, the airbag begins to deflate until the internal gas is almost completely expelled. Upon the main airbag becoming nearly empty, the airbag loses its buffering effect and the descent speed of the vehicle that has not been buffered is defined as the terminal impact velocity ().
For obtaining the magnitude of the terminal impact velocity, a vertical impact test bench is plotted in
Figure 1. When fully deployed, the under-vehicle airbag reaches a height of 1.2 m. The weight of the vehicle is 7600 kg and the ground coordinate is defined in
Figure 1. In addition, the angle between the vehicle’s heading (longitudinal axis) and its horizontal velocity vector is defined as the sideslip angle (
). When the vertical landing impact test bench works, the vehicle is lifted by a cable to a height of 3.2 m and then conducted free fall afterwards to ensure a ground touching speed of about 8 m/s (typical touchdown speed in the airdrop mission).
Afterwards, an accelerometer (MPU9250) is used to record the vehicle’s overload (the acceleration of the vehicle divided by the acceleration of gravity) curve as shown in
Figure 2 wherein the accelerometer’s measurement range is ±16 g (g means the acceleration of gravity). The microcontroller unit (MCU) is powered by a 7.4 V battery and stores the acceleration data measured by the accelerometer. The velocity and displacement curves are then calculated by the acceleration curve (proportional to the overload curve) through integration. Through analyzing the velocity and displacement curves, the terminal impact velocity can be obtained.
where the
and
mean the exhaust flow coefficient and gas leakage coefficient,
and
stand for the exhaust port area and leakage area,
and
mean the pressure inside and outside of the airbag,
means the universal gas constant,
is the temperature,
is the specific heat ratio. Equations (1) and (2) come from the theoretical manual of LS-Dyna R13 Software.
Apart from the above experiment measurement devices, this study also proposes a fluid–structure interaction (FSI) model to simulate the vehicle’s impact process. The buffer airbag of the airdrop system is fixed to the vehicle through the front and rear beams and modeled based on the control volume (CV) method. Since the complex contact model between the buffer system and the vehicle has little impact on the finite element analysis, the front and rear beams of the buffer system are simplified as part of the vehicle model. Also, the impact of the vehicle’s gun barrel on the calculation of airbag buffering process is minimal. Therefore, the gun barrel is deleted for simplicity. At the beginning of the simulation, the height of the base plate from the ground is set to 0.751 m and the vehicle falls freely at an initial speed of 7.02 m/s so that the speed at which the base plate contacts the ground is around 8 m/s. The meshes of the vehicle as well as the airbags are plotted with the Hypermesh 13.0 software as shown in
Figure 3. The total cell numbers for the vehicle, airbag, and ground are 500,000.
In addition, in this simulation model, the airbag exhaust port is usually bonded with nylon buckles and the deflation pressure is very low. Thus, the default deflation pressure of the exhaust port is assumed to be the ambient pressure . When the pressure inside the airbag () is higher than , the mass flow through the vents and leakage can be expressed using Equations (1)–(3). Wherein, means vent orifice coefficient, is the vent orifice area, means the orifice coefficient for leakage, is the area for leakage, is the ideal gas constant, stands for the specific heat of gas (equals to 1.4 for air), denotes the internal temperature of the airbag gas. The main bag’s exhaust hole area is set as 0.018 m2 while the auxiliary bag’s exhaust hole area is set as 0.028 m2. Both the main bag’s vent hole area and the auxiliary bag’s vent hole area are given as 0.030 m2.
2.2. Experimental and Numerical Method in Rigid Collision Stage
Right after the airbag cushioning stage, the second stage starts when the bottom surface of the vehicle touches the ground. In that case, the vehicle collides with the ground directly and may cause the rollover or forward-backward flipping along the pitch axis as shown in
Figure 4. The analysis in this step can also be divided into experiment study and numerical study.
In the actual airdrop process, due to the influence of wind, the vehicle has a certain horizontal velocity at the start of the rigid collision stage. The initial vertical descent velocity is equal to the terminal impact velocity at the end of the airbag cushioning stage. In this work, the vertical and horizontal movements of the airdrop system are treated as two independent parameters, while the initial pitch angle is assumed to be zero during the rigid collision stage. The differences in the vertical deformation of each airbag are neglected.
Thus, on one hand, the structure of the oblique impact test bench is plotted in
Figure 4. Unlike the experiment instruments in
Figure 1, the oblique impact test bench gives the vehicle a horizontal velocity (
). The angle between the longitudinal direction and the horizontal velocity vector when the bottom surface of the vehicle touches the ground is defined as the terminal sideslip angle (
).
On the other hand, using the terminal impact velocity at the end of the airbag cushioning stage as the initial descent velocity for the collision process, the collision process between the vehicle and the ground under different terminal sideslip angle (
) was calculated based on an ADAMS model. The critical horizontal velocities leading to the rollover (or backward/forward flipping) phenomenon under different terminal sideslip angles are then obtained to be an evaluation index for the anti-rollover performance of the vehicle. In the ADAMS model, the parameters setting for the calculation of normal force and friction force are listed in
Table 1 where the moments of inertia are defined in body axis.
2.3. Setup of the Terminal Sideslip Angle Active Control System
Since the terminal sideslip angles affect the landing attitude stability, it is fundamental to design certain control systems to achieve certain terminal sideslip angles and help increase the landing attitude stability. As shown in
Figure 5, the micro controller unit (MCU) used in the control system is the STM32F407 chip manufactured by O1 studio. The wind sensor powered by an independent 12 V battery (manufactured by GT electronics technology) is connected to the MCU through transistor-transistor logic (TTL) serial communication, while the UB 482 module used to obtain the high precision orientation of the vehicle is connected to the MCU through UART port. The GPS receiver embedded in UB 482 module can receive satellite signals and calculate the current position and velocity of the vehicle in real time to provide initial position information and gyro drift correction; magnetometers embedded in UB 482 module can be used for initial heading alignment and gyro heading drift correction.
To guarantee enough steering torque, two 150 mm ducted fans are connected in parallel to form a thrust generation group. Also, since the forward and reverse rotation speeds of ducted fans are different under the same voltage, two thrust generation groups aligned in opposite direction are needed in total. For each ducted fan, two 25.9 V batteries and one voltage controller are needed. The 25.9 V battery’s capacity is 6000 mAh and its maximum current is 189 A. With a single operation time of 2–3 min, it can be repeatedly charged and used, meeting the requirements of small size and light weight.
4. Conclusions
In this paper, the experiments and simulation analyses are mainly focused on the vehicle’s roll and pitch stability during the airbag cushioning stage and the rigid collision stage in the airdrop process. In the airbag cushioning stage, a vertical impact test bench and a fluid–structure interaction (FSI) model are built up to study the vehicle’s impact process. The buffer airbag of the airdrop system is modeled based on the control volume (CV) method. In both the numerical model and experiment, the speed at which the bottom of the airbag contacts the ground is around 8 m/s. In the following rigid collision stage, an oblique impact test bench and a dynamic model are proposed. Wherein, the terminal impact velocity obtained in the airbag cushioning stage is given as the initial descent velocity in the rigid collision stage. It is noted that the vertical () and horizontal () movements of the vehicle in body axis are treated as two independent parameters in this study. The initial pitch angle is assumed to be zero during the rigid collision stage and the differences in the vertical deformation of each airbag are neglected. After analyzing the impact process, a terminal sideslip angle active control system is built up to help increase the landing attitude stability. The O1 studio STM32F405 chip is used in the control system as the micro controller unit (MCU). Simultaneously, a wind sensor and a UB 482 module are used to obtain the horizontal velocity vector as well as the longitudinal axis of the vehicle. Four ducted fans and the corresponding PD control algorithm are also applied to adjust the azimuth angle of the vehicle. From the numerical and experimental studies in this paper, the following key conclusions can be drawn:
In the airbag cushioning stage, the peak overload in simulation and two experiment trials are 11.2 g, 12.9 g, and 10.8 g, respectively. The terminal impact velocity is obtained as around 2 m/s.
In the rigid collision stage, the increase in the initial descent velocity can increase the peak roll angles increase and reduce the roll stability of the vehicle. The peak roll angles increase dramatically with the increase in the initial horizontal velocity. Although the vehicle shares very good roll stability when the terminal sideslip angle is 0° or 180°, it does not mean the vehicle can sustain its pitch stability with huge initial horizontal speed. The critical horizontal velocity almost tripled from about 5.3 m/s to 14.7 m/s if turning the terminal sideslip angle from 90° into 0°/180°.
With the help of the sideslip angle active control system activated when the vehicle reaches 250 m height, the angular velocity can be kept within a relatively small value. Also, at the landing time, the angles between the vehicle’s horizonal velocity vector and its longitudinal axis get quite close to 0°/180° whatever the initial airdrop height is. It means that the vehicle can maximize its critical horizonal velocity.
To summarize, this work analyses the roll/pitch stability of the vehicle during the airbag cushioning stage and the rigid collision stage and draws several theoretical conclusions. Then, a sideslip angle active control system is built up to improve landing stability according to the theoretical conclusions. This study establishes a technical foundation for the design of a highly reliable anti-rollover device capable of active attitude adjustment for the airdrop vehicle. In fact, the current airdrop test data is still insufficient, and the thrust of the ducted fan may have caused a lag in vehicle attitude adjustment. Thus, it is fundamental to further improve the thrust generated by the active attitude control system in the next work.