Journal of Terramechanics最新文献

With the advent of artificial intelligence, the analysis of systems related to complex processes has become possible or easier. The interaction of the traction factor of off-road vehicles with soil or other uncommon surfaces is one of the complex mechanical problems, which has been very difficult to model and analyze in conventional and previous methods due to numerous and variable parameters. This review article delves into the imperative and progression of integrating AI algorithms within the realm of modeling and predicting target parameters in Terramechanics engineering. Such endeavors are especially pertinent to predicting various soil properties, including soil compaction, traction, energy consumption, deformation, and associated factors. The application of AI encompasses various facets, including modeling and predicting traction, soil sinkage, rut depth, contact area, soil stress, density, and energy wasted on the traction device’s movement on the soil. The present study evaluates the solutions and benefits offered by AI-based methodologies in addressing soil-machine interaction challenges. Furthermore, the study investigates the constraints inherent in utilizing these methodologies.

On soft terrain, the rover wheels are easy to slip, sink, or even fail to move. This paper designs a soil-plowing wheel which is two-sided closed and without tread. The discrete element simulation shows that the wheel could grasp soil through both sides and plowing soil and that the ability to gain drawbar pull is not significantly reduced. The wheel is fabricated and tested to measure its sinkage, slip rate and drawbar pull. The wheel has high sinking, high slip and high drawbar pull. And the wheel is tested to verify the passability on five terrains of flat ground, climbing, out of sinkage, obstacle crossing and hard ground. The wheel exhibits good passability in all terrains. The soil-plowing wheel is tested verify the passability on three terrains of obstacle crossing, out of sinkage and climbing and using a three-rockers six-wheels rover. The wheel can pass through all terrain. More importantly, the wheel has an excellent ability to get out of sinkage. And it takes only 25.43 s for all six wheels to get out of sinkage. It is believed that the structure and test results of this wheel are valuable for the subsequent development of unmanned rover wheel.

Bulldozers are one of the major off-road machine systems for cutting and transporting granular materials during earthmoving operations. With the growing demand for energy-efficient and accelerated optimization design cycles and automated earthmoving processes, researchers and engineers are exploring methods to model the soil-to-bulldozer interaction. This study proposes a similitude scaling law for the soil and scaled blade systems, providing an alternative approach to costly and time-consuming field-based design verification and validation for product engineering of Ground-Engaging Tools (GETs). In this soil bin study, we examined the interaction between scaled bulldozer blades and cohesive-frictional artificial soil, aiming to establish scaling laws of geometrically scaled blade ratio to two blade performance responses, soil reaction forces and soil mass. A randomized complete block design with five replications was conducted in a soil bin using five 3D printed geometric scales of the blade, λ = 1/8, λ = 1/9, λ = 1/11, λ = 1/13, and λ = 1/15, with λ = 1 representing the full-scale geometry of a Caterpillar D3K2 XL bulldozer blade. Blade soil cutting forces were measured using a load cell instrumented blade dynamometer carriage on a cohesive-frictional artificial soil in the bin. Each scaled blade traveled at a constant speed of 213 mm/s and the tool depth was set to 30 % of the blade height. After reaching full load, the cut soil mass and pile dimensions (height, width, and rupture length) above the undisturbed soil were also measured. A scaling law model was established between soil horizontal reaction forces and the five geometric blade scale ratios with a high coefficient of determination, R², of 0.9898. Similarly, the scaling law (R² = 0.9951) was established between the five geometric scales and soil mass. The findings demonstrate that a scaling law model can be used for predicting the soil horizontal reaction force and soil load. The scaling law can be utilized for optimizing energy and productivity, enhancing GET product design optimization, and developing algorithms for energy-efficient automation of earthmoving processes.

Difference thresholds of whole-body vibration is important to determine perceptibility of changes in vehicle vibration and can be used to guide ride comfort improvements. It is postulated that estimated difference thresholds in a laboratory setting should be applicable to real-world driving conditions given that the stimuli are similar. This study considers the aspect of vehicle vibration associated with the stimulus. A validated non-linear SUV vehicle model is simulated on a 4-poster test rig and driven in a straight-line over a rough road. This allows for the vehicle vibration to be compared between vertical excitation only (4-poster) and complete excitation (straight-line driving) by the road profile at each of the four wheels. Results show that differences in the seat vibration exists between the 4-poster test rig and straight-line driving simulations. These differences are larger than difference thresholds implying that they would most probably be perceivable. Further investigations are needed to determine the influence of differences in vibration stimuli on difference thresholds.

The deep sea polymetallic nodule mining vehicle maneuverability depends on the vehicle track parameters and track configuration. The traction force offered by the deep sea soil is very limited for the mining vehicle during dynamic operating conditions on the seabed and it is very critical to maneuver against the external resistances. The present study strives to arrive at optimum track parameters for enhancing the traction of the vehicle for the pre-determined seabed conditions. The efficacy of the four tracks in Inline and Offset track configurations on the soft soil has been compared. To improve the traction force estimation, the existing mathematical model was modified with the inclusion of dynamic variation of shear stress-shear displacement characteristics and variation in shear residual displacement concerning the track parameters. The modified mathematical model was solved in a well-established mathematical tool and found that there are 30 percent improvements in the traction force generation for the offset configuration over inline track configuration. The optimum track length to width ratio ($L/b$) was also estimated for the given contact area to configure the vehicle track for improvement of the traction. Further, a Multi-Body Dynamic (MBD) analysis has been carried out in commercially available soil-machine interaction tool for the inline and offset track configurations with actual measured seabed soil parameters. The MBD analysis proved that the sinkage and vehicle gradient is significantly increased in the inline track configuration due to disturbance created by the front tracks. The simulation results confirm that the offset track configuration is suitable for the deep sea soil conditions for handling the higher payload of a deep sea mining vehicle.

The soil in southwest China is a cohesive soil in which discrete cohesive particles and aggregates coexist. In view of the problem that there are many studies on discrete cohesive particles and a lack of research on aggregates, discrete element software DEM is used to conduct a study on cohesive particles and agglomerates parameter calibration. The angle of repose is selected as the target value to calibrate the simulation parameters of sticky particles. Then, the simulation parameters of the viscous particles are used as the basis for the calibration of the contact parameters of the agglomerates, and shear experiments are conducted on the agglomerates, with the ultimate shear depth and ultimate shear force as target values. The results show that the parameters of the agglomerate are: Normal Stiffness per unit area is 5.627 × 10⁸ N/m³, Shear Stiffness per unit area is 4.359 × 10⁸ N/m³, Critical Normal Stress is 3.5 × 10⁶ Pa, Critical Shear Stress is 4.5 × 10⁶ Pa and Bonded Disk Radius is 5.43 mm. Through the particle sliding friction angle test and the agglomerate compression test, it was verified that the errors of sticky particles were 0.30 % and 0.37 % respectively, and the error of agglomerates was 1.69 %.

Vegetation override is an important aspect of off-road ground vehicle mobility. For autonomous ground vehicles (AGV), path-planning in off-road environments may be informed by the predicted resistance of vegetation in the navigation environment. However, there are no prior measurements on the override resistance of small stems ($<$2.5 cm) and groups of small stems on medium-sized ($\approx$1000 kg) vehicles. In this work, a series of override measurements for clumps of small vegetation that are relevant for off-road navigation by intermediate-sized AGV is presented. The development and calibration of a custom-made pushbar system with integrated load cells for directly measuring override forces is also presented, and a comparison of the results of the experiments to models developed for override of larger single stems is made. It is found that for clumps of small vegetation, the total override force is best predicted by the diameter of the largest stem in the clump. Additionally, it is found that the equations developed for larger stems under-predict the override forces exerted by smaller stems by about a factor of two.

The interaction between deformable terrain and wheels significantly affects wheel mobility. To accurately predict vehicle mobility or optimize wheel design, an analysis of this interaction is essential. This study develops a hybrid terramechanics model (HTM) that integrates the semi-empirical model (SEM) and the discrete element method (DEM) using artificial neural networks (ANNs). The model overcomes the limitations inherent in SEM and DEM approaches. We used DEM simulations to analyze the impact of wheel design parameters and slip ratio on terrain behavior. ANNs were subsequently developed to predict dynamic sinkage in real time based on these results. A new concept, termed bulldozing angle, was introduced to define additional terrain–wheel contact caused by dynamic sinkage. Based on this concept, we predicted the bulldozing resistance exerted on the wheel. By combining SEM, ANNs, and DEM, we developed an HTM capable of terrain behavior analysis. Lastly, we conducted a comparative analysis between the SEM, HTM, and actual test data. The results confirmed that the predictive accuracy of the HTM surpassed that of the SEM across all slip ratios.

Modeling the wheel-soil interaction of small-wheeled robots in granular media is important for robot design and control. A wide range of applications, from earthmoving for construction and farming vehicles to navigating rough terrain for Mars rovers, motivate the need for a model that can predict the force response of a wheel and the terrain shape afterward. More importantly, the speed, accuracy, and generality of the model should be considered for real-world applications. In this paper, we offer a straightforward sand deformation simulator to simulate the soil surface and integrate it with 3D-RFT in order to capture the soil motion caused by the wheel. The proposed method is able to: (1) estimate three-dimensional interaction forces of arbitrarily shaped wheels traveling in granular media; (2) simulate the soil displacement from the trajectory; and (3) perform the force calculation in real-time at 60 Hz.

Accurately estimating surficial soil moisture and strength is integral to determining vehicle mobility and is especially important over large spatial extents at fine resolutions. While methods exist to estimate soil strength across landscapes, they are empirical and rely on class average soil behavior. The Strength of Surficial Soils (STRESS) model was developed to estimate moisture-variable soil strength with a physics-based approach. However, there is a lack of field data from a diverse landscape to evaluate the STRESS model. To test the STRESS model, a field study was conducted in northern Colorado. Soil moisture and strength were measured on 10 dates. Data from the surficial layer (0–5 cm) were used to test the STRESS model and determine if soil strength trends could be estimated from topographical and soil textural differences. High variability was observed in soil strength measurements stemming from fine-scale terrain features and user variability. Observations show lower soil strengths for greater soil moistures, steeper slopes, higher vegetation, and lower soil fines content. STRESS estimated rating cone index values with a mean relative error of 37.5 %, while a pre-existing model had a mean relative error of 47.4 %. The STRESS model outperforms the current RCI prediction method, but uncertainty remains large.