Rollover Safety and Workable Boundary Suggestion of an Agricultural Platform with Different Attachments

Abstract

Overturning accidents in agriculture offroad vehicles have been reported annually around the world. Safety structures, such as rollover prevention, have been adopted to mitigate these accidents. Despite this, accidents remain persistent but less fatal. This study investigated the safe driving boundary for a multi-purpose platform with attachments (cargo, pepper harvester). Mathematical formulations of the roll and pitch motions for multi-purpose platforms were created. The critical overturning angle, at five load levels for each machine (0, 50, 100, 150, 200 kg), was determined using Recurdyn, a dynamic analysis software. Regression analysis of each coordinate of the center of gravity was conducted to verify the most critical coordinate against gaining loads. The critical overturning angle for the X and Y directions was found. The cargo and pepper harvester had 28.64° and 21.04° of critical overturning angles in the Y direction, respectively, with a full loading state while climbing a steep hill. The regression model of the X, Y, and Z coordinates of CoG suggested that the most critical coordinate of the CoG against the loads was in the Y coordinate for the pepper harvester and the X direction for cargo. This study could be applied to develop an awareness safety system that warns the operator of the risk of a fatal accident.

1. Introduction

Agricultural machines have eased operations and saved labor costs and manhours. However, components such as tractor attachments have caused accidents that resulted in severe injuries and death [1,2]. Efforts to reduce machine-related incidents are evident in applying the rollover protective structure (ROPS). Tractor overturning still frequently occurs, and the danger of using this agricultural machine is persistent [3,4,5]. Irrelevant to the severity of injuries, overturning occurs in two ways: longitudinally and laterally, and 25% of operators using ROPS-equipped tractors become injured in both ways [6]. The ROPS seems to only prevent severe injury, and accidents still occur.

In many countries, the majority of fatal tractor accidents involve overturning. In Spain, most fatal agricultural accidents were related to tractors, and 70% of deaths were due to overturning [7]. Tractor overturning is also the most frequent accident type in Portugal, and the majority are caused by sloping or irregular lands [8]. More than a hundred fatal agricultural tractor overturning accidents occur annually [9,10]. Similarly, 34.1% of the agricultural accidents in Turkey are caused by the sideway and backward motions of the tractor overturning on various road conditions [11,12]. In the United States, 32% of fatalities and 6% of non-fatal injury cases are tractor-related, with tractor overturning as the main reason. The fatality risk from tractor overturning has been reported as the most critical factor [13,14].

These fatal accidents necessitate recommendations for tractor safety under working boundary conditions. The overturning of a tractor is an instability or control-loss accident that is fatal to the operator [15]. Results of previous tractor stability experiments with various parameters (i.e., the center of gravity (CoG) position, wheelbase, and track width) showed that the slope angle significantly affected tractor overturning. Providing safety information to the operator could prevent these accidents [16,17]. In addition, working with the industry to promote research on new operational guidelines will help reduce overturning-related deaths and permanent disabilities [18].

A dynamic rollover model of a tractor under various conditions has been formulated to suggest safety guidelines and improvements for tractor stability. A mathematical model was also developed to examine a tractor’s lateral stability in different conditions [19,20]. However, a difference in the mathematical value was found, requiring additional experimental work [21]. Designing actual tests to determine overturning angles under various conditions, such as overturning in different directions and applying heavy loading, is complex [19]. Correspondingly, a computational simulation implemented an efficient method of analyzing a tractor’s safety dynamics. Baker and Guzzomi [22] found that the instability angle depends on the position of the posterior CoG. Gravalos et al. [23] also revealed that selecting a tractor’s rear parameters is essential in improving the stability of vehicles working on side slopes. Majdan et al. [24] measured the coordinates of the CoG with different rear wheel ballast weights and proposed regression models for the overturning angle and coordinates of the CoG. Sun et al. [25] simulated the sideways overturning characteristic of a tractor using a dynamic simulation and proposed safe working conditions. Similarly, Watanabe and Sakai [26] conducted a numerical analysis of tractor overturning under different conditions and verified the dynamic characteristics in fatal accident cases.

The present study was conducted to simulate a multi-purpose platform’s lateral and longitudinal overturning characteristics with different attachments and loading conditions. This study aims: (1) to define the overturning motion of the vehicle in the X and Y directions, (2) to find out the overturning angle of the vehicle in each attachment with various loading conditions, and (3) to analyze the overturning characteristics in the X and Y directions and suggest the safety driving boundary of the vehicle with each attachment. This would be the initial research for an agricultural vehicle-overturning alert system, the cost benefits of which could be calculated as the ROPS system was [27,28] and would complement the driving safety of agricultural vehicles [29,30].

2. Material and Methods

2.1. Formulation of Motions

The multi-purpose platform was designed with a continuous track, which separated the wheels, not at the front and rear, but left and right. Typically, the dynamic motion of the vehicle needs to be considered when the force is acting on each of the four wheels. However, only the two forces acting on the multi-purpose platform used in this study were considered to explain the dynamic motion of the vehicle (left and right tracks).

The vehicle stays stable with the supporting force ($F_z$) and the reaction force ($R_z$) against the ground. When the force was acted on the vehicle in the X direction, the roll occurs with $\phi$ and the frictional force ($f$), and the reaction forces in each direction ($R_y$ and $R_z$) were affected newly. While these forces support the vehicle, the initial center of gravity ($Co{G}_i$) moved to the rolled center of gravity ($Co{G}_r$). Consequently, the roll motion of the multi-purpose platform could be expressed as the equilibrium state of the torque that could be found with the multiplication of the moment of inertia ($I_x$) with the lateral angular acceleration ($\ddot{\phi }$) and the sum of reaction torque (Figure 1a). The equation for the rolling motion is expressed as follows:

$$\begin{gathered} I_x\ddot{\phi }=F_{zr}\left[jcos\left({\phi }_i-\phi \right)-\left(e_y-y\right)\right] \\ -F_{zl}\left[jcos\left({\phi }_i+\phi \right)+\left(e_y-y\right)\right]cos\theta \\ +Hf+R_y\left(e_{y}sin\phi +e_{z}cos\phi \right)+R_z\left(e_y-y\right)cos\theta \end{gathered}$$

For analysis of the roll angle ($\phi$), the above equation could be expressed as follows:

$$\begin{gathered} \phi =\iint \frac{F_{zr}\left[jcos\left({\phi }_i-\phi \right)-\left(e_y-y\right)\right]}{I_x}dt \\ -\iint \frac{F_{zl}\left[jcos\left({\phi }_i+\phi \right)+\left(e_y-y\right)\right]cos\theta }{I_x}dt \\ +\iint \frac{Hf+R_y\left(e_{y}sin\phi +e_{z}cos\phi \right)+R_z\left(e_y-y\right)cos\theta }{I_x}dt \end{gathered}$$

Like the rolling motion, the pitching motion is shown in Figure 1b. The track vehicle design did not assort forward and backward wheels but only came up with the supporting force ($F_z$). When the force was acted on the vehicle in the Y direction, the pitch occurs with $\theta$ and the reaction force ($R_x$) was affected newly. While these forces support the vehicle, the initial center of gravity ($Co{G}_i$) moved to the $Co{G}_p$. Consequently, the pitch motion of the multi-purpose platform could be expressed as the equilibrium state of the torque that could be found with the multiplication of the moment of inertia ($I_y$) with the longitudinal acceleration ($\ddot{\theta }$) and the sum of reaction torque. The equation for the pitch motion is expressed as follows:

$$I_y\ddot{\theta }=\left[R_z{h}_f\cos \left({\theta }_1-\theta \right)-F_z{h}_{r}cos\left({\theta }_2+\theta \right)\right]cos\phi$$

For analysis of the pitch angle ($\theta$), the above equation could be expressed as follows:

$$\theta =\iint \frac{[R_z{h}_{f}cos\left({\theta }_1-\theta \right)-F_z{h}_{r}cos\left({\theta }_2+\theta \right)\left]cos\phi \right]}{I_y}dt$$

2.2. Modeling of the Multi-Purpose Platform with Attachments

The multi-purpose platform was designed with a continuous track used in various fields [31,32,33]. It can control each side of the track individually and is good at distributing the amount of weight with a large contact area to the ground.

Table 1. Specifications of the vehicles.

Parameters (Unit)	Specifications
	Pepper Harvester			Cargo
Length × width × height (m)	4.0 × 1.6 × 1.9			2.1 × 1.6 × 1.5
Ground clearance (m)	0.6			0.6
Maximum loading weight (kg)	200			200
Original coordinate	X	Y	Z	X	Y	Z
	−51.97	−1288.03	841.61	−28.75	1222.52	657.34

presents the specifications of each attachment, namely the pepper harvester (CH301, TYM Co., Ltd., Daegu, Korea) and the cargo (DC30, TYM Co., Ltd., Daegu, Korea), used in this study. Kim and Park [34] analyzed the dynamic properties of a semi-crawler-type mini-forwarder with a simplified model that can be applied to computational simulations. The continuous track and the attachments were also simplified in the simulation method to reduce the computations and facilitate the simulations.

2.3. Experimental Set-Up

To suggest the safe driving boundary of the vehicle for each attachment, five loading conditions were applied (i.e., 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg). The overturning simulation was conducted in the X and Y directions. The overturning angle was defined when the contact between the ground and the platform became zero.

Table 2. Weight of each vehicle with the loading conditions.

Load (kg)	Vehicles
	Pepper Harvester	Cargo
Empty	2270	1550
50	2320	1600
100	2370	1650
150	2420	1700
200	2470	1750

presents the vehicle weight in each condition. The overturning angles in the X and Y directions were analyzed with each loading condition, while the pure overturning angle was analyzed in an empty loading condition. The CoG prediction model was analyzed with SPSS (IBM SPSS Statistics 26, IBM, Armonk, NY, USA) linear regression for dynamic stability of the vehicle for each attachment.

2.4. Simulation Environment Set-Up

Figure 2 shows the general model of the multi-purpose platform. Each trackside was supported by the ground force (F_G), friction force (f), and reaction force y (F_y) generated as the plane turns in the X direction. Overturning occurs when the contact between the ground and crawler reaches zero as the plane turns in the X direction. In the same way, overturning will occur in the Y direction if the plane turns in the Y direction. Two coordinate systems were utilized. One local coordinate system is located at the center of gravity (CoG_m) and the global coordinate system (CoG_O). This study used the program Recurdyn (V8R5, Functionbay, Korea) for the dynamic analysis.

The overturning simulations were conducted for analysis of lateral and longitudinal stability. However, these computational works were performed with some assumptions that could not be applied to the system. The general model was based on the following assumptions:

• There is no yaw motion.
• The crawler comes into contact with the ground in areas.
• The normal friction coefficient is generated at the crawler–ground interface.
• CoG_m will shift in all directions with the loading conditions.
• The ground is a rigid body, and there is no deformation.
• The external conditions such as temperature and wind are ignored.

3. Results

3.1. Overturning Angle of the Vehicles with the Loading Conditions

The overturning angle of the pepper harvester was determined under different loading conditions (

Table 3. Overturning angle of the pepper harvester.

Load (kg)	Direction
	X		Y
	Right (°)	Left (°)	Front (°)	Rear (°)
Empty	39.39	31.19	35.01	28.17
50	39.22	31.46	36.20	26.67
100	39.14	31.60	37.50	25.51
150	38.80	31.87	38.91	23.37
200	38.29	32.13	39.61	21.04

). The overturning angles of the pepper harvester in the X direction were 39.39°, 39.22°, 39.14°, 38.80°, and 38.29° on the right side and 31.19°, 31.46°, 31.60°, 31.87°, and 32.13° on the left side in each loading condition, respectively (i.e., 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg). Comparing the overturning angle on the right and left sides, the critical overturning angle was found on the left side (31.19°, empty loading). This incident was because of the leaning to the left-side driver’s seat. In the case of empty loading, which means the machine is stable without the load except for the driver’s weight, the critical load for the dynamic accident is only the load on the seat, which is changeable based on the driver. For dynamic stability in the case of empty loading conditions, weight balance of the front section using extra weight is necessary. There is no disadvantage of gaining extra weight on the vehicle for the weight balance. Varani et al. [35] reported that the changes in tractor mass distribution did not produce practical effects on machine performance and fuel consumption. Similar results were observed in other works [36,37,38,39].

The overturning angles of the pepper harvester in the Y direction were 35.01°, 36.20°, 37.50°, 38.91°, and 39.61° at the front and 28.17°, 26.67°, 25.51°, 23.37°, and 21.04° at the rear in each loading conditions, respectively (i.e., 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg). The minimum overturning angle in the Y direction was found at the rear (21.04°, full loading). In the full loading state of storage with 200 kg of hot peppers, the overturning danger was increased by 24% compared with the empty loading state at the rear. To prevent a fatal accident in the Y direction at the rear, yield estimations of hot peppers should be considered, with a notification to the driver about a possibly fatal accident.

In addition, regression analysis results revealed the weakness of stability in each direction (Figure 3). The slope of each line in Figure 3 could be explained by the number of differences with the load, and the steeper the slope, the more unstable the vehicle is. The slope in the Y direction (Figure 3b) was seven times steeper than in the X direction, which indicates that the instability of the pepper harvester in the Y direction is seven times more dangerous than in the X direction.

The overturning angle of the cargo was also determined under different loading conditions (

Table 4. Overturning angle of the cargo.

Load (kg)	Direction
	X		Y
	Right (°)	Left (°)	Front (°)	Rear (°)
Empty	44.85	37.21	45.44	31.34
50	43.82	37.46	45.11	30.76
100	43.23	37.71	44.94	29.85
150	42.30	38.03	44.96	29.36
200	41.87	38.18	44.57	28.64

). The overturning angles of the cargo in the X direction were 44.85°, 43.82°, 43.23°, 42.30°, and 41.87° on the right side and 37.21°, 37.46°, 37.71°, 38.03°, and 38.18° on the left side in each loading condition, respectively (i.e., 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg). Similar incidents were found with the hot pepper harvester. The impact of the driver’s seat was indicated, as it determined the overturning angle with the pepper harvester and the cargo.

The overturning angles of the cargo in the Y direction were 45.44°, 45.11°, 44.94°, 44.76°, and 44.57° at the front and 31.34°, 30.76°, 29.85°, 29.36°, and 28.64° at the rear in each loading condition, respectively (i.e., 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg).

Likewise, the regression analysis results of the cargo also revealed the weakness of stability in each direction (Figure 4). The slope in the X direction (Figure 4a) was found to be steeper than in the Y direction, which indicates that the instability of the cargo in the X direction is more dangerous than in the Y direction. However, the overturning accident still occurs in the Y direction at 28.64°, even when the stability weakness was found in the X direction.

Previous studies found a reduction in fatal accidents in agriculture and the cost-effectiveness of preventing disablement of an operator through the hardware structure, ROPS [27,28]. However, fatal accidents continue to occur and need to be reduced in other ways. The software structure, a safety alert system, could be the solution [29,30]. Autunes et al. [8] reported that the information about all the accident circumstances, and the use of ROPS in tractors, is often removed to facilitate maneuvering in farmland [7,29]. Similarly, Di Nocera et al. [1] also mentioned that the industry still needs to consider this, and a system for operationalizing an operator’s attention is required. Ahmadi [20] suggested the requirement to design an alarm system to warn the tractor operator of possible overturning or skidding. Consequently, the analysis defining the overturning characteristics of the agricultural tractor is generally required, and the advantages of tractor stability should be reinforced by catching the operator’s attention.

For the static state of the empty loading condition, the driver’s seat on the left side caused the CoG to lean to one side. The left side of the vehicle had an advantage of 7° compared with the right side. Gaining weight increased the roll angle of turning on the right side. The same results were obtained in both the pepper harvester and the cargo.

3.2. X, Y, and Z Coordinate Change with the Loading Conditions

Figure 5 shows the results of the CoG coordinate displacement of the pepper harvester under each loading condition. The X coordinates of the CoG were −51.97, −49.33, −46.81, −44.39, and −42.08 at the loading conditions of 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg, respectively. The Y coordinates of the CoG were −1288.03, −1256.37, −1226.08, −1197.08, and −1169.29 at the loading conditions of 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg, respectively. The Z coordinates of the CoG were 841.61, 842.10, 842.57, 843.02, and 843.44 at the loading conditions of 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg, respectively. In addition, the most influenced coordinate as the load became heavy was investigated through regression analysis. The slope of the line in Figure 5 explains the severity of the load’s influence on the displacement of coordination. The steeper the line, the more displacement occurs. It was found that the Y coordinate was the most influenced coordination against the load (i.e., 0.594; R² = 0.99) because the pepper harvester has a minimum overturning angle of 21.04°. It is important to find the most variable direction by comparing each regression model [40] (Unver-Okan et al. 2020). The instability in the Y direction is seven times more dangerous than in the X direction.

Figure 6 shows the results of the CoG coordinate displacement of the cargo with each loading condition. The X coordinates of the CoG were −28.75, −15.68, −3.46, −7.99, and −18.77 at the loading conditions of 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg, respectively. The Y coordinates of the CoG were −1222.52, −1216.13, −1210.16, −1204.55, and −1199.25 at the loading conditions of 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg, respectively. The Z coordinates of the CoG were 657.34, 670.18, 682.20, 693.48, and 704.08 at the loading conditions of 0 kg, 50 kg, 100 kg, 150 kg, and 200 kg, respectively. Like the pepper harvester, the most influenced coordinate of the cargo was revealed through the regression analysis of the Y coordinate (i.e., 0.234; R² = 0.99). Considering the most influenced coordination, the cargo’s stability characteristic may overturn at the minimum angle of 28.64° in the Y direction, even though the line slope (overturning angle vs. loads) was steeper in the X direction than could be explained.

To prevent a fatal accident, balancing the vehicle’s CoG is required with additional weight, such as adding ballast. Similarly, the importance of adjusting the CoG for stability was emphasized in previous studies. Adjusting the CoG, lowering the height of the CoG, and balancing the leaning CoG by shifting or adding weight are effective and necessary [20,23]. Adjusting CoG reinforces the multi-purpose platform driving performance, such as climbing steeper slopes and driving on the side slopes [41,42].

4. Conclusions

This study analyzed the stability of a multi-purpose platform attached to a pepper harvester and cargo under different loading conditions. For the pepper harvester, the critical overturning angles were 31.19° and 21.04° in the X and Y directions each. Additionally, the Y direction was the most influenced coordinate against the load (i.e., 0.594; R² = 0.99). For cargo, the critical overturning angles were 37.21° and 28.64° in the X and Y directions. The X direction was the most influenced coordinate against the load (i.e., 0.234; R² = 0.99).

The results showed that the platform was at risk of overturning in the Y direction when it was attached with a pepper harvester and in the X direction when it was attached with the cargo. However, verification is needed for future work to test the developed model.

The results of this study can be used for other safety systems, such as a safety-alert system that warns the operator and prevents fatal accidents, overturning, and skidding. The stability characteristics analysis for the general agricultural tractor is required to reduce the number of fatalities.