Assessment of Safe and Sustainable Operation for Freight Transportation Company Based on Tire Set Configurations Used in Its Trucks’ Fleet

Abstract

This article investigates the safety potential of a freight transportation company, considering tire set selection as one of the most important aspects to ensure safe driving and a reliable transportation service. The revision of tire sets selection in large vehicle fleets is attributed to a new regulation from the United Nations to maintain non-deteriorating tire wet braking performance up to a minimum allowable wear limit, encouraging both safety and sector sustainability, as a significant part of tires are currently replaced before reaching a tread depth of 3 mm. In this research, an experimental test was conducted that involved four maneuvers with a truck using ten different sets of tires (including new and retreaded) to determine which set performs better in critical driving conditions. The results are then analyzed using the TOPSIS method where the most efficient set of tires and the best alternatives are selected. Finally, the safety of trucks on the road using the appropriate set of tires is evaluated by the estimated accident reduction potential. It should be mentioned that the optimal selection of the truck tire set is also important for sustainable transportation, as the pollution of worn tires remains a relevant environmental issue.

1. Introduction

Smooth and efficient freight transport on land roads depends mainly on traffic and weather conditions, and to consistently reduce the number of accidents, road infrastructure and vehicle safety is being improved, driver offenses monitored, and road safety training and education provided. Despite this, around 1.19 million people a year are still killed on the world’s roads and between 20 and 50 million suffer non-fatal injuries [1]. According to American accident statistics, fatalities involving heavy-duty vehicles (HDV) account for approximately 11% of the total, while in Europe this is up to 14% [2,3].

In addition to continually improving safety technology, it is important to consider the proper maintenance and operation of the vehicle fleet when analyzing the causes of vehicle accidents. These processes are particularly relevant for freight transport companies and focus mainly on road safety [4,5]. The use of defective or worn tires can be one aspect of poor fleet maintenance, and tire-related poor grip is directly related to both the cause and the consequences of an accident [6]. Tires are an essential part of the costs associated with truck operation; therefore, the rational and safety-optimized selection of the tire set is an important task in organizing the operation of a congested vehicle fleet.

The rational use of tires is important not only from a safety point of view, but also from an environmental perspective [7]. It is estimated that with the new regulation from the United Nations (UN R117-04) to maintain non-deteriorating tire wet braking performance up to a minimum allowable wear limit, there will be a reduction in tire production-based CO₂ emissions in Europe by 6.6 million metric tons. This step toward tidiness in the transport sector is driven by the trend that almost 50% of tires are discarded before they reach the 3 mm wear threshold.

The tires of commercial vehicles used for the regular commercial transport of goods are replaced in axles as required; that is, all tires are not replaced synchronously due to unequal wear, which would directly increase the costs for a logistic company. This often leads to the search for the most cost-effective tire replacement principles, while the complexity of predicting driving safety means that the matching of these combinations is usually not evaluated. This research aims to determine to what extent the handling characteristics of HGV vary when different tire sets are used and to what extent a different tire set can increase the safety potential of a company’s managed fleet.

The paper is organized as follows. The literature review is described in Section 2, and the materials and methods used for experimental research, including a description of the tire sets, are presented in Section 3. In Section 4, the experimental research with evaluation of the collected data is presented. Section 5 discusses the road safety assessment based on the estimated data for the truck fleet. In Section 6, the results and discussion regarding processed data are presented. Finally, Section 7 concludes the paper.

2. Literature Review

Theoretical and experimental research are aimed at assessing the most appropriate choice of vehicle structure, component parts, or materials to achieve the best possible dynamic and safety performance [8]. Here, variable speed conditions for a non-stationary system’s dynamics require specific studies which include experimental research with data-driven strategies [9,10]. In this case, tire interaction with the road surface is a key factor in the safe and reliable handling of the vehicle and the effective operation of the active safety systems [11,12]; therefore, experimental and theoretical tests are carried out to discover the relationship between road safety and tire properties or changes in tire state [13].

Experimental tests on the whole vehicle are the most widely used to assess braking efficiency and handling under different conditions. Stoklosa and Bartnik [14] conducted experimental brake tests on dry, wet, and snowy road surfaces using summer and winter tires, respectively. The tests examined the effect of tire inflation pressure on stopping distances. The braking tests from a speed of 50 km/h showed an increase in stopping distances of 0.49 m on dry asphalt pavement using summer tires with a pressure of 1.5 bar. Above the recommended tire pressure, stopping distances increased by 0.25 m on average. With winter tires, the braking distance increases in all cases, with the greatest increase of 0.33 m at 1 bar. Similar to this, Kreitzberg et al. [15] summarized a test carried out in winter conditions with full stop braking and stopping with obstacle avoidance. Measurements were carried out in 30 different passenger cars for a total of 200 tests: 69 straight-line braking tests from 50 km/h and 131 braking tests with an obstacle on the road. The test site was covered with a 3–5 cm thick layer of packed snow with ice cracks underneath and the temperature ranged from −7 °C to +3 °C. The hypothesis that heavier vehicles generally have a longer braking distance in slippery conditions was rejected. The main factor for the shorter stopping distances was the tread depth of the tires, with the year of manufacture of the tires remaining a secondary factor.

Malmivuo and Luoma [16] conducted a study to investigate the traction of studded and non-studded tires on snowy and icy surfaces. Nokian Hakkapeliitta 8 adapted tires were used and 150 runs were carried out on a track. The test also aimed to assess how the friction coefficient changes when different numbers of vehicles pass over the test section. For studded tires, the effect of the test lane loading on friction was found to be relatively small, but for non-studded tires, the lane loading resulted in a 24–31% reduction in grip under braking and a 13–19% reduction under acceleration.

Heavy vehicle tire testing is much rarer, as it is more costly and requires an expensive test vehicle, which can only be offered by new truck dealerships and logistics or haulage companies [17,18]. Tuononen [19] describes the characteristics of heavy-duty vehicle tires, which are estimated from optically measured tire carcass deformations. Accurate information on tire forces is important for the development of active safety systems and the prediction of vehicle stability under critical driving conditions. For research purposes, a dedicated sensor module has been developed to estimate tire deformation and forces that act on the tire in real time based on the displacement of the tire carcass. A Volvo FH12 truck is used for this test, with sensors inside the rim. The test was carried out on a snow-covered frozen lake, with a 150 m arc turning maneuver selected. The vehicle was accelerated to a threshold speed at which the towing vehicle starts to slip. The test showed that the chosen optical sensor technology is generally valid: the lateral and longitudinal force estimates are accurate in most cases, but the vertical force estimate is not reliable under heavy braking.

In addition to this, it is accepted that tire testing under variable road conditions is particularly relevant, but its accuracy is more difficult to ensure. Following this, tire contact patch deformation on wet surfaces was investigated by Niskanen and Tuononen [20]. An integrated optical sensor in the tire provides insight into the relationship and context of partial and full aquaplaning. Researchers have developed a real-time algorithm to detect partial aquaplaning at speeds up to 40 km/h. Optical, strain, or acceleration sensing inside the tire is relevant for all types of tires, while more promising applications have been tested on large or agriculture tires [21]; however, the mass integration of these sensors in tires is not yet ready, even though the installation of accelerometers is easier [22].

In contrast to the experimental assessment of overall braking performance or vehicle handling, tire grip properties are often analyzed separately. In this way, vehicle tire test benches offer many opportunities to investigate a wide range of tire traction, wear, noise, and other parameters; however, they are cost and time consuming and, compared to a detailed mathematical model (computerized calculations), do not always provide an accurate assessment of all the desired parameters. In these conditions, it is difficult to assess the load on the tire and choose the right road surface [23]. Furthermore, on test benches, the tire is usually not in natural contact with the road surface; that is, a large diameter roller that results in an incomplete contact patch [24] or a conveyor belt [25].

The comparison between experimental data and simulation results is a common approach in current research. Lugaro et al. [26] have found that both the measurement and simulation results show similar slip performance results; however, the measured stopping distances between the different tests differ by up to 2 m in individual cases. This is mainly due to the different properties of the pavement in the calculation model and in the test area; however, surface roughness causes high nonlinearities for tire grip and wear properties; therefore, these contact features must be taken into account during tire performance testing [27]. In general, stopping distances are an important aspect in assessing the technical capability of trucks to improve safety, as they determine the probability of avoiding an accident or minimizing its consequences. For heavy-duty vehicles, this safety criterion is very important because the stopping distance is longer than for other road users due to the high mass. In addition, the high weight of trucks often leads to multi-crashes, especially on highways or mountain roads [28]. Similarly, stopping sight distance (SSD) is often evaluated in road safety projects and it consists of stopping distance, driver reaction time, and brake system deployment time. For a truck with a trailer, the safe stopping sight distance at 90 km/h is 175 m, while for an average passenger car, this distance is 145 m [29].

A potential collision situation with a slower vehicle in front was analyzed by Christ [8], who found that for each of the parameters tested (vehicle speeds, distance between vehicles, driver reaction time, brake response time, and tire-pavement grip), the relative risk level was reduced; however, it was highlighted that a 5% increase in grip would result in a 4% reduction in accident rates. The author also points out that the increased use of advanced safety systems will reduce the influence of the driver and increase the impact of the grip on the risk of a crash [30,31]. Despite the fact that the preventive measures are always a positive step towards road safety, when an accident occurs, the driver’s actions and vehicle condition remain the only factors. When a vehicle is braking, the tire contact patch with specific tread design and rubber compound is the only place the vehicle comes into contact with the road surface [32].

In this context, individual efforts or trends to significantly reduce accident rates are not sufficiently effective. A freight transport fleet is an appropriate ecosystem for the assessment of a wide range of different safety measures, such as the following: driver selection and training, improving driving culture, route planning and management systems, driver workplace improvements, safety technology, and fleet maintenance [33,34]. There is no single best strategy to improve road safety, and the most effective is usually a combination of road measures that affect individual fleets differently. Automated fleet management and routing is another important tool to improve safety in freight transport companies [35]. Route planning allows you to find the shortest or fastest route, avoid congestion, and reduce driver anxiety and fatigue. Such smart systems also allow one to monitor driver’s well-being and drowsiness, and avoid the urge to engage in extraneous and irritating activities. Smart vehicle monitoring systems include systems that allow the monitoring of technical conditions, brake wear, maintenance intervals, tire inflation, or even general wear.

The global automation of the freight fleet can improve road safety if all vehicles meet at least the minimum requirements of the model. In line with this, Wang et al. [36] estimated in their study that, by achieving certain vehicle and road surface parameters, using a safe distance model and vehicle communication, rear-end collisions can be completely avoided. In general, vehicle safety assessment is a broad and widely researched area, including tire testing; however, the research on trucks is more specific, especially when it comes to assessing the effects of different tires and their features. The specifics of freight transport companies lead to a compromise between rational tire selection and maximum possible driving safety; therefore, in this paper, the option of individual tire set selection as a sustainable alternative is planned after experimental tests and a truck fleet safety assessment. From the sustainability point of view, retreaded tires are included in this research as a valuable option to reduce the amount of waste [37].

3. Research Conditions and Used Tire Sets

Test sessions on braking performance, hill climb, and cornering were conducted on an enclosed proving ground to determine which tire set gives the best handling for the truck (Figure 1). Three selected maneuvers were all carried out on the same MAN TGX 18.470 truck with a tri-axle trailer and measuring equipment mounted on a cabin (Figure 2). The inertia measurement unit (IMU) together with the GPS sensor from “Race Technology” (Nottingham, UK) were used to log vehicle acceleration, yaw rate, and GPS data. The truck was driven by the same instructor, experienced in driving such vehicles.

The brake test, as one of the most important in a potential emergency situation, was carried out in two phases: on a wet asphalt surface with an initial speed of 60 km/h and on an artificially slippery surface with a water film from an initial speed of 40 km/h (Figure 1). In both braking maneuvers, the driver applies the brake pedal fully to achieve the shortest braking distance. All the assisted driving and safety systems of the truck were active during braking.

The downhill slippery cornering maneuver was carried out with a steady steering input at 20 km/h (Figure 1). For safety reasons, this maneuver was carried out without a trailer. Finally, the slippery hill climb was carried out from the same starting position each time, with constant accelerator pedal depression, and with an active traction control system.

The truck was tested with seven different commonly used tire sets consisting of new and used tires from different manufacturers and with different seasonality. To examine whether cheaper retreaded tires are as safe on the road as new tires, new and retreaded tires were compared on different tire sets. The trailer braking test used retreaded universal tires (three sets) of different patterns, retreaded by different tire retreaders. This range of selected tires is usually used by freight companies in Central and North Europe where the road surface is usually wet or snowy in autumn and winter seasons.

Table 1. Tires with their short description and assigned coding are used for experimental research.

Tire Tread Design	Tire Properties	Tire Tread Design	Tire Properties
	UGMAXS		ENRUGMAXD
	New, winter		Retreaded from 3-year-old carcass
	‘Goodyear’		‘Goodyear KMAX D’
	(tread hardness 66.7)		(tread hardness 63.5)
	(Colmas-Berg, Luxembourg)		(Colmas-Berg, Luxembourg)
	NF501		GUMRZT220
	New, all-season		Retreaded from a 3-year-old carcass
	‘Kama NF501’		‘Vipal VT220’
	(tread hardness 68.5)		(tread hardness 68.6)
	(Kruševac, Serbia)		(Nova Prata, Brasil)
	KMAXSG2		GUMM788
	New, all-season		Retreaded from 3-year-old carcass, all-season
	‘Goodyear KMAX S’		‘Bandag’
	(tread hardness 71.9)		(tread hardness 57.6)
	(Colmas-Berg, Luxembourg)		(Muscatine, IA, United States)
	MOTHKPLD		MOTHKPLF2
	Retreaded from a 3-year-old carcass		Retreaded from 3-year-old carcass, all-season
	‘Noktop’		‘Kraiburg’
	(tread hardness 63.4)		(tread hardness 65.6)
	(Nokia, Finland)		(Geretsberg, Austria)
	UGMAXD		ENRWSS
	Retreaded from 3-year-old carcass, all-season		Retreated from 3-year-old carcass, all-season
	‘Nokian Hakkapeliitta Truck E2’		‘Ringtread Blackline WSS’
	(tread hardness 65.4)		(tread hardness 67)
	(Nokia, Finland)		(Henstedt-Ulzburg, Germany)
	MOTHKPLE
	New, winter
	‘Goodyear UltraGrip Max D’
	(tread hardness 66.5)
	(Colmas-Berg, Luxembourg)

shows a list of tires used and a short description. Each tire description also includes the tread hardness measured during the test according to Shore A class [38].

Although seven different tire configurations are used for the tractor and three for the trailer, most of them use the same tires on a single axle (

Table 2. Tire sets used in experimental research.

No of Tire Set	Vehicle Axis
	Front (Steered)	Rear (Drive)	Trailer
	Code for Tire Set
1	UGMAXS	MOTHKPLD	–
2	NF501		–
3	KMAXSG2		–
4		UGMAXD	–
5		MOTHKPLE	–
6		ENRUGMAXD	–
7		GUMRZT220	–
8			GUMM788
9			MOTHKPLF2
10			ENRWSS

). In this way, the first three sets of tires consisted of identical tires for the rear axle of the truck and different tires for the front axle of the truck, so that for the first three sets of tests, only the front tires were compared. By evaluating tires in this way, it is theoretically possible to select a tire set that will be even more effective in braking tests.

4. Experimental Research

4.1. Evaluation of Brake Efficiency

The mean fully developed deceleration (MFDD) was calculated for the braking efficiency analysis that does not consider the initial and final braking stages and is calculated from raw longitudinal acceleration data [39]:

$$\mathrm{M}\mathrm{F}\mathrm{D}\mathrm{D}={\displaystyle \frac{v_a^2-v_b^2}{2\left(S_b-S_a\right)}},$$

(1)

here, v_a—vehicle speed corresponding to 80% of the initial speed (0.8v_o); v_b—vehicle speed corresponding to 80% of the initial speed (0.1v_o); S_b—distance traveled in braking mode between v_o and v_b; S_a—distance traveled in braking mode between v_o and v_a; and v_o—initial vehicle speed before braking.

The brake data on wet and artificially slippery surfaces are summarized in

Table 3. Data evaluated from two braking tests.

No of Tire Set	Average Deceleration, m/s2	Maximum Deceleration Value, m/s2	MFDD, m/s2	Braking Distance, m	Braking Time, s
Braking on wet asphalt pavement
Front axle under consideration
1	6.05	7.46	6.67	28.32	3.15
2	5.66	6.79	6.05	28.59	3.37
3	6.25	7.83	6.78	27.73	3.11
Rear axle under consideration
4	6.59	7.90	7.17	26.06	2.95
5	6.36	7.96	6.79	24.74	2.92
6	6.49	8.07	7.11	24.90	2.89
7	6.51	8.14	7.00	24.30	2.91
Trailer
8	6.31	7.97	6.90	26.70	2.99
9	5.82	7.50	6.53	27.37	3.18
10	6.31	7.76	7.02	25.35	2.92
Braking on artificially slippery surface
Front axle under consideration
1	2.11	4.74	2.48	46.22	6.25
2	2.55	4.65	2.89	37.79	5.39
3	2.24	5.25	2.44	39.04	5.76
Rear axle under consideration
4	2.18	4.89	2.37	39.67	5.89
5	1.90	4.67	2.01	39.91	6.21
6	1.82	4.67	2.05	48.12	6.94
7	1.94	4.86	2.11	42.48	6.35
Trailer
8	1.78	4.59	2.17	48.58	6.89
9	1.80	4.39	2.09	51.07	7.08
10	1.77	4.64	2.01	46.02	6.73

.

The braking performance results show certain trends between tire sets, where the most efficient tire sets also change with different pavement conditions, making it difficult to clearly identify which tire set is the most efficient and the best possible alternatives for safe driving. Here, the technique for order of preference by similarity to ideal solution (TOPSIS) methodology is used for a more accurate assessment of results. This approach is used in a wide range of areas where different possible alternatives are analyzed [40,41].

In this multi-criteria decision making (MCDM) method, it is assumed that the tire sets to be tested are defined as alternatives denoted by A₁, A₂, …, A_m, and the criteria (MFDD, stopping distances, and other indicators) against which the alternatives are evaluated are denoted by C₁, C₂, …, C_n. Also, x_ij (i = 1, 2, …, m; j = 1, 2, …, n) denotes the value assigned to criterion i of alternative j and $X={\left[x_{ij}\right]}_{mn}$ is the decision matrix. The associated weight value of each criterion is denoted by $X=\left[w_1,w_2,\dots ,w_n\right]$, when ${\sum }_{j=1}^n{w}_j=1$ $\sum w_j=1$ [42].

The decision matrix is then normalized according to the following equation:

$$\overline{r_{ij}}={\displaystyle \frac{x_{ij}}{\sqrt{{\sum }_{k=1}^m{x}_{kj}^2}}}; i=1,\dots ,m; j=1,\dots ,n,$$

(2)

here, r_ij—the normalized value i of criterion j for alternative A_i.

A weighted normalized decision matrix is then calculated:

$$v_{ij}=w_j·r_{ij}; i=1,\dots ,m; j=1,\dots ,n,$$

(3)

here, w_j—the weight of criterion or attribute j.

The ideal positive and ideal negative solutions for each criterion are then determined:

$$A^{+}=\left\{v_1^{+},\dots ,v_n^{+}\right\},$$

(4)

$$A^{-}=\left\{v_1^{-},\dots ,v_n^{-}\right\},$$

(5)

here, A⁺—specifies the ideal positive solution; and A⁻—ideal negative solution. If criterion j is useful, then it is assumed that $v_j^{+}=max\left\{v_{ij},i=1,\dots ,m\right\}$ and $v_j^{-}=min\left\{v_{ij},i=1,\dots ,m\right\}$. If the criterion is useless, then $v_j^{+}=min\left\{v_{ij},i=1,\dots ,m\right\}$ and $v_j^{-}=max\left\{v_{ij},i=1,\dots ,m\right\}$.

The distances from each of the alternatives i to the best positive and negative solution are then calculated:

$$D_i^{+}=\sqrt{{\sum }_{j=1}^n{\left(v_{ij}·v_j^{+}\right)}^2},i=1,\dots ,m,$$

(6)

$$D_i^{-}=\sqrt{{\sum }_{j=1}^n{\left(v_{ij}·v_j^{-}\right)}^2},i=1,\dots ,m.$$

(7)

Finally, each alternative is evaluated in terms of its relative proximity to the ideal solution and distance from the negative solution:

$$y_i={\displaystyle \frac{D_i^{-}}{D_i^{+}{·D}_i^{-}}}$$

(8)

The performance indicator shows which of the alternatives, in this case a set of tires, is the best in terms of the selected evaluation criteria.

This analysis requires the selection of appropriate values or weights for the criteria. This is already a human factor, but for a more accurate selection of the weights, a correlation analysis of the different braking criteria is carried out using Microsoft Excel Solver software (v16.0) to discover the relationship between the different measured braking test criteria and the extent to which they are dependent on each other. The strategy chosen to weight the criterion most heavily was to use the less correlated measurements, as, because they are less correlated, they are less likely to influence the results of the analysis. The weights of the other criteria were chosen based on the practical use of the measurement results. In studies where the selection of weights is important, the TOPSIS method often uses expert interviews [43,44], but in this case, it was decided that were would be a focus on the opinion of the researchers who conducted the test. The weights for the “Race Technology” accelerometer performance criteria are shown in

Table 4. Values of alternative criteria.

Criteria	Criteria Value, wj
Average deceleration, m/s2	0.1
Maximum deceleration, m/s2	0.1
MFDD, m/s2	0.35
Braking distance, m	0.35
Braking time, s	0.1

.

The analysis of the numerical results shows that the relationship between stopping distance, stopping time, and the MFDD value remains, despite the fact that the MFDD value does not take into account the initial braking pulse and the end of the braking (braking repulsion). Some of the values indicate that the full deceleration may not have been developed from the start of braking (because of tire slip, delay in ABS activation, and delay in brake force control).

After analyzing the braking maneuver test measurements using TOPSIS analysis, the specific rankings of the tire sets obtained according to the braking criteria evaluated are prepared in

Table 5. Results of the brake test according to the TOPSIS performance index y_i.

No of Tire Set	Braking on Wet Asphalt Pavement	Braking on Artificially Slippery Surface	Evaluation Ranking Position
Front axle under consideration
1	0.74	0.05	3
2	0.21	0.89	2
3	1.00	0.53	1
Rear axle under consideration
4	0.56	0.86	1
5	0.68	0.51	3
6	0.83	0.16	4
7	0.86	0.46	2
Trailer
8	0.56	0.62	2
9	0.16	0.24	3
10	0.93	0.61	1

.

The most effective braking on wet asphalt was achieved with tire set 3, consisting of a new Goodyear KMAX S all-season tire in the front and a NOKTOP retreaded tire in the rear (Table 1), but under different braking conditions, such as on artificially slippery surfaces (imitation of ice), tire set 2, consisting of a new Kama NF501 all-season tire in the front axle of the towing vehicle and a retreaded winter tire with a Noktop tread in the rear axle, was more effective. However, the difference in test results was not sufficient to allow the overall performance of set 2 to exceed that of set 3. The most efficient tires were those with many small tread blocks with transverse grooves. Since the artificially slippery road surface corresponds to ice-covered asphalt and the geometry of the described tires is designed to cling in sipes to a low-grip road surface, the results seem to justify the predicted tire performance.

When comparing tire sets for the rear axle of the truck, tire set 4, consisting of Goodyear KMAX S all-season tires for the front axle of the truck and retreaded NOKIAN all-season tires, was the most effective in the braking test. However, this is only an optimal option, but not the best one for braking on different surfaces. In the tests, tire set 6 was the most effective on wet asphalt and tire set 4 on artificially slippery surfaces, but these tire sets were less effective on other surfaces.

Of the three different tire configurations for the trailer, the most efficient is tire set 10, with the finest tread and fine grooves angled at 45 degrees. The truck with this set of tires stopped 1.95 m shorter and had an MFDD of 0.36 m/s² higher than the worst performing trailer tire.

The results partially supported the hypothesis that a single set of tires would not be the most efficient for braking on both surfaces, due to differences in tread pattern properties, seasonality, and tread compound hardness. On wet asphalt, tires with a harder compound and fewer transverse grooves were more effective, but large tread blocks and fewer transverse grooves did not grip as well on artificially slippery surfaces, which had a higher water film, and tires with these geometries are more likely to suffer aquaplaning.

4.2. Hill Climb Evaluation

The hill climb allows the traction of the drive axle tires to be assessed when starting from a standstill on slippery road surfaces. Two parameters were chosen for the qualitative evaluation of the test: the average longitudinal acceleration over a distance of 35 m and the time taken to reach the distance of 35 m. The test used the same truck as in the braking test, without a trailer, with tire changes on the drive axle. The same maneuver start location was chosen for the runs, and the end location was determined after 35 m from the start of the maneuver based on GPS sensor data. The accelerator pedal was released after the end of the maneuver mark, thus avoiding braking or accelerator release before the maneuver was completed. The test results were analyzed using the TOPSIS analysis method described in the previous section and the results are presented in

Table 6. The results of a hill climb for a truck on an artificially slippery slope.

No of Tire Set	Average Value of Longitudinal Acceleration After 35 m, m/s2	Time Taken to Drive 35 m Climbing, s	TOPSIS Performance Indicator yi	Evaluation Ranking Position
1	1.20	15.21	0.34	4
4	1.34	13.89	0.81	2
5	1.26	15.19	0.42	3
6	1.19	16.72	0.12	5
7	1.38	13.25	1	1

.

According to the data, the most effective tire set for the artificially slippery climb was tire set 7, which consisted of a new Goodyear KMAX S all-season tire, which did not affect grip in this maneuver, and a retreaded rear tire with Vipal VT220 winter tread. These tires had the hardest tread on the Shore A scale. Compared to braking tests, where the softest tires were the most effective on wet asphalt, the results are reversed in this test: the second best-performing tire was the NOKIAN HAKKAPELIITTA TRUCK E2 retreaded winter tire, while the tire set 5 (the new Goodyear UltraGrip Max D tires) was better than the Goodyear KMAX S and KMAX D retreaded tires with similar tread, but this may also be due to the tire’s sipes and the hardness of the rubber compounds.

4.3. Driving Downhill with a Corner Evaluation

During the driving downhill with a corner experiment, the average lateral acceleration and yaw rate acting on the truck from its initial onset (the start of the track curve) to the full stop were analyzed. This test checked whether the truck was turning enough to maintain its trajectory and whether the unladen drive axle (without trailer) was slipping.

The measuring equipment has a GPS sensor, so it was possible to identify a single point during each test run at which cornering starts, and to sample the exact point at which cornering intensity starts to increase. From the results of the measurements (

Table 7. Results of a driving downhill with a corner evaluation.

No of Tire Set	Mean Lateral Acceleration, m/s2	Mean Yaw Rate, deg/s
1	3.637	19.34
2	2.424	18.66
3	2.947	20.11
4	2.703	17.72
5	2.800	18.22
6	2.912	20.39
7	2.034	16.12

), it can be tentatively stated that the truck handles best under these conditions with tire sets 1, 3, and 6, but the values given depend on the slope and angle of the track, and the maneuver path and the driver’s action (maintaining a constant speed). Therefore, the yaw rate and driving speed data in the time domain were used to evaluate this maneuver.

These characteristics show whether the steering or drive axles were slipping, as well as the speed at which the maneuver was carried out at the apex of the turn, when the maximum yaw rate is developed. All maneuvers were started at an average speed of 26 km/h and the downhill driving was at idle speed in the engaged gear, without the accelerator or the service brake pedal, maintaining the same trajectory (driver factor). Figure 3 shows three runs of the truck with tire set 6.

The graph shows that between the sixth and seventh second, all three tests showed a significant variation in the yaw rate. This is because the rear axle of the truck was slipping when the truck was in the turning circle, resulting in a sharper rotation around the vertical axis, while the reduced intensity indicates the slipping of the steering axle (vehicle understeering). The front axle of the tractor slipped in the first and second tests and the rear (drive) axle lost traction in the third maneuver, so this set of tires did not give stable results in this test. For all tire sets, tire set 5 was the most stable one (Figure 4).

It was observed that a stable yaw rate was maintained with tire set 5. The graph does not show any significant spikes in the yaw motion, so it can be concluded that there was no significant tire slippage and the handling of the truck remained smooth and predictable. The average initial speed of the three tests was 26 km/h, which was not lower than that for the other set of tires.

5. Safety Assessment of the Truck Fleet

One of the most important issues discussed in the field of road transport is the maximum permissible speed limits and their link with the frequency of accidents. Most drivers believe that they are capable of choosing a safe speed, but there is not a country that has not introduced speed limits. However, despite driver training and education, 30–50% of all drivers exceed the maximum speed limit. Speeding significantly increases the risk of accidents and injury, or even death [45]. Elvik et al. describe the emergence of the Power model and its application to road safety assessment, where the aim is to assess how a change in speed determines the risk of certain factors related to road safety. This is carried out by reviewing and summarizing a large number of studies and accident statistics that have assessed the effect of speed changes on the number and severity of accidents. It also compares whether the Power model can be accredited as suitable for use in practice, and what factors may lead to a discrepancy between the theoretical calculations.

The model is suitable for describing the severity of an accident; however, it does not specify the probability of an accident. The model has implications for the consequences of accidents involving injuries and fatalities, as it derives a degree indicator. According to the model, the number of fatal accidents, accidents with serious injuries, and all accidents with injuries recorded by the police (including fatal and serious injury accidents) varies in proportion to the change in the speed of driving, when raised by a certain degree, which is dependent on the severity of the accident or the injury under investigation. In general, the model predicts the change in the number of accidents or the number of injured or killed road users as the speed changes.

When analyzing specific aspects of road safety, the degree to which the speed differential is raised varies depending on the aspect being analyzed, as all aspects vary differently. A specific dataset is defined for the degrees and their thresholds used to assess the different severities of an accident or injury [45].

Based on this, the report of Institute of Transportation Engineers [46] used the accident rate of change method to assess the road safety risks of the experimental test. The crash modification factor (CMF) is calculated by raising the ratio of the original and modified speeds by a certain grade:

$$\mathrm{C}\mathrm{M}\mathrm{F}={\left({\displaystyle \frac{v_a}{v_b}}\right)}^x$$

(9)

here, v_a—initial speed or speed before traffic regulation changes; v_b—actual speed or speed after traffic regulation changes; and x—grade indicator by type of accident or injury.

This approach is used to select or modify the maximum speed limit on a given section of traffic, but in this case, it is applied to an experimental study on tires and how different tire performance results can determine the road safety potential.

On the basis of the braking tests carried out in this research, it was chosen to assess the difference in speeds during full braking on different surfaces. The CMF is calculated by selecting the initial speed of the most efficient tire set in terms of braking distance and the speeds of the other tire sets in the same test series in the denominator of Formula (9). As the speed is selected at the time of braking, a situation is modeled in which there is already a probability of an accident, i.e., an obstacle in the road, and therefore no rate of change in accident occurrence is calculated, only the relative rate of change in the different injuries. The speed data for the conducted braking tests are available at 100 Hz; therefore, it was chosen to assess the CMF over the entire braking maneuver range. This allows the performance of different tire configurations to be seen and the variation throughout the maneuver to be seen, as an obstacle collision can occur at any time during braking or at any speed in the braking interval. Although a CMF of 1 as a reference value is derived from the most efficient speed parameters of the tire set, it can be seen from this analysis whether there are times during braking when a more useful CMF value for road safety is calculated for a vehicle with an experimentally captured longer braking distance.

The CMF analysis of the experimental tests compared the front, rear, and semi-trailer tire configurations on wet asphalt and artificially slippery surfaces. For the CMF assessment, the grade x = 4.1 is selected for fatalities and x = 2.2 for severe injury change. Among the rear tire configurations of the truck, the shortest braking distance was found in the initial analysis with tire set 7 on wet asphalt, resulting in a CMF value of 1 for this set of tires.

Figure 5 shows that the instantaneous (residual) speed during braking varies differently for each set of tires, giving a different CMF value. Tire set 3 has a lower relative CMF than the others in the entire speed range, indicating that it is the most unsafe tire in an accident. The vehicle with the tire set 6 maintains a higher speed than the reference tire set 7 (CMF < 1) during half of the braking maneuver, but after slowing to a speed of v_a = 24 km/h, the CMF has a lower speed value than set 7 (starts to generate more effective braking) and thus a reduced change in fatalities in the accident. At the end of braking, as the speed drops to 11–12 km/h, there is a sharp drop in CMF of sets 4 and 6, indicating better wet grip of set 7 at low towing speeds.

In order to visually evaluate the most efficient tire configuration and its alternatives, a histogram of the distribution of values is prepared and presented next (vertically) to the CMF values. The width spread of the histogram indicates the mean CMF value and the height of the histogram indicates the standard deviation.

The CMF graphical comparison curves for the risk of injury to road users do not differ much from the relative value of the risk of fatality, because CMF calculation uses the same speeds for the same experimental analysis, but the grade is selected x = 2.2. Figure 6 shows a CMF plot of the front tire test on an artificially slippery surface, evaluating the change in road traffic injuries.

The most effective in terms of braking distance in the initial analysis is tire set 2, and its CMF is used as a reference. From the graph it can be seen that the initial speed at the start of braking was not completely uniform, since the CMF curve for set 3 starts higher than for set 2. This shows that the denominator of the CMF equation had a lower speed than the numerator, so the vehicle with set 3 had a lower speed at the start of the brake. If the curve starts below the reference CMF value, then the speed at the start of braking is higher. The CMF curves for tire sets 1 and 3 decrease uniformly, which means that the value of the speed difference increases almost equally throughout the braking period. At a reference braking speed of 20 km/h, an increase in injury change during the accident of 29% with tire set 3 and 42% with tire set 1 is calculated.

A situation can be modeled where X collisions occurred at 20 km/h when braking from 60 km/h on a snow surface, causing injuries to Y road users, with a vehicle with tire set 2 in use. The calculations presented here would then predict that the injury rate would be 29% higher under these conditions with tire set 3.

It is difficult to estimate from the CMF graph exactly how much the risk of fatality increases with the most ineffective tire set; therefore, the mean CMF value and standard deviation for each alternative tire set were calculated. A CMF value less than one indicates an increase in the change or number of injuries in the accident under investigation, based on statistical data, while values greater than one indicate a decrease in the change in the accident under investigation when compared to the reference speed. Comparisons of CMF values converted into percentages for the front tires are shown in

Table 8. CMF comparison of front axle tire sets under braking on different surfaces.

Road Surface	No of Tire Set			No of Tire Set
	1	2	3	1	2	3
	CMF for Road Fatalities			CMF for Road Traffic Injuries
Wet asphalt	10% (0.27)	10% (0.38)	–	7% (0.20)	14% (0.29)	–
Artificially slippery	50% (0.33)	–	29% (0.41)	35% (0.29)	–	21% (0.30)

, for the rear tires in

Table 9. CMF comparison of rear axle tire sets under braking on different surfaces.

Road Surface	No of Tire Set					No of Tire Set
	3	4	5	6	7	3	4	5	6	7
	CMF for Road Fatalities					CMF for Road Traffic Injuries
Wet asphalt	49% (0.31)	10% (0.11)	21% (0.27)	13% (0.25)	–	34% (0.28)	5% (0.06)	14% (0.21)	9% (0.20)	–
Artificially slippery	–	−5% (0.40)	31% (0.31)	47% (0.28)	35% (0.27)	–	−1% (0.16)	20% (0.22)	30% (0.23)	21% (0.20)

, and for the trailer in

Table 10. CMF comparison of trailer tire sets under braking on different surfaces.

Road Surface	No of Tire Set			No of Tire Set
	8	9	10	8	9	10
	CMF for Road Fatalities			CMF for Road Traffic Injuries
Wet asphalt	28% (0.32)	9% (0.44)	–	19% (0.25)	10% (0.31)	–
Artificially slippery	11% (0.14)	16% (0.35)	–	6% (0.08)	13% (0.28)	–

.

According to the safety potential evaluation, the mean value of the coefficient of variation of tire set 4 for road traffic injuries when braking on wet asphalt pavement is $\overline{x}$ = 0.949 and the standard deviation is σ = 0.059. Therefore, the vehicle with tire set 4 results in a 5.1% increase in the change in road fatalities (Table 9) compared to the most efficient tire set 7. The CMF values are reported in the same way in Table 8, Table 9 and Table 10. Negative values indicate that the tire set has a higher safety potential than the reference for the entire braking interval.

6. Results and Discussion

An analysis of the experimental test results identified the tire sets that were the most effective in the selected maneuvers. In the braking tests with the trailer, tire set 10 was the most effective one, with a braking distance of 1.98 m shorter than the other alternatives on wet asphalt surfaces. The analysis also identified the best tires for the maneuver on different surfaces, as well as the alternatives in order of performance, so that other tire combinations are possible if necessary to adapt to road conditions or financial resources.

According to the Shore A measurement of the hardness of the tread rubber compound, it can be seen that harder tires were more effective on wet asphalt and softer tires were more effective on artificially slippery surfaces. Tire set 3 has the softest tread (63.4 Shore A) and is the most effective at braking on artificially slippery surfaces but the least effective on wet asphalt, while tire set 7 has the hardest tread (68.6 Shore A) and is therefore the most effective on wet asphalt and the least effective on slippery surface. The same correlation was observed between the tread hardness and braking performance of the trailer tire sets, while no correlation was observed between the tire sets of the steering axle of the truck.

During the uphill maneuver, the test with the tire set 7 was the most efficient, reaching the target distance of 35 m in 13.25 s. It was 0.7 s faster than the second-best tire set (3.5 s faster than the worse tire set). The tires from set 7 had the highest tread hardness of all the tire sets compared, while the least effective tire set was the one with the lowest measured hardness.

The graphs of yaw rate and driving speed were used to assess the front and rear axle slip and the stability of its intensity during downhill driving with the cornering maneuver. The most efficient and constant tests were carried out with the tire set 5 (medium tread hardness), as no significant front or rear axle slip was observed (as it was for other tire seats) and an average lateral acceleration of 2.8 m/s² was achieved.

The TOPSIS analysis evaluates each alternative against a set of criteria to identify the alternative that best meets these criteria. The method is based on the concept of “closeness to the ideal solution”, where the ideal solution represents the best possible result for each criterion and the worst solution represents the worst possible result. While TOPSIS offers a systematic approach to ranking alternatives according to their proximity to the ideal solution, it is important to recognize that real solutions often involve uncertainty and subjective judgment, as the weights of the criteria are chosen by the expert carrying out the analysis. In these types of analyses, it is important to ensure the robustness and reliability of the results, so a sensitivity test is carried out to see whether changing the weight of a criterion changes the position of the objects in the final ranking.

In the TOPSIS sensitivity analysis for braking tests, the weights of the criteria were changed by dividing the values equally or by giving one or the other of the criteria the majority of the weight. In the analysis of deceleration performance, giving more weight to the mean and maximum accelerations changes the ranking of the tire sets on the trailer: on wet asphalt, the tenth and eighth tire sets are ranked first and second, respectively, and on slippery surfaces, vice versa.

The best tire sets for each axle of the truck have remained the same, regardless of which feature is given more weight. This may be due to the fact that most of the criteria are quite highly correlated and interdependent. The best sets do not change with changing weights, but if alternatives are considered, such as the next best set of tires (if the first set is not available), the sensitivity test may lead to alternatives being ranked. It is therefore important to consider the importance of each of the criteria more explicitly, or to use a different method of analysis.

7. Conclusions

This paper describes the safety assessment of trucks based on experimentally tested different tire sets in four maneuvers on a vehicle test track with two types of low friction surfaces. A total of 11 tire sets were divided into three groups by axle: front axle, rear axle, and trailer. From the experimental test results, the most significant indicators of the truck’s driving dynamics for initial comparison were determined: mean fully developed deceleration (MFDD), braking distance, braking time, and mean and maximum longitudinal acceleration. On the basis of that the processing of the test data was carried out by applying a multi-criteria decision analysis, in which the output criteria were evaluated by selecting different weights, and a sensitivity analysis was carried out to test whether changing the weights of the criteria changes the ranking of the tire sets.

In the braking tests on the trailer truck, the most effective tire set consisted of new all-season front tires and a retreaded all-season rear tires. The main feature that led to shorter brake distances on both surfaces is the size and layout of the tread blocks. Larger tread blocks and grooves angled at 45 degrees resulted in better braking on wet asphalt, while smaller tread blocks resulted in shorter braking distances on artificially slippery surfaces that replicate ice traction.

A slippery hill climbing maneuver compared tire sets on the drive axle of a truck and the results showed that that the tire set with retreaded rear tires with a small asymmetric pattern had the best grip. In this course, the downhill cornering maneuver was conducted to test the ability of the truck to maintain stability and handling on slippery road surfaces. The most effective tire in this maneuver was a set of new all-purpose front and rear tires designed for the winter season.

In the case of the rear axle tire assemblies, the retreaded tires were more effective in braking and hill climbing maneuvers, while the new tires performed better in the downhill cornering maneuver. The new rear tire configuration was ranked third out of five in the final braking maneuver ranking. The three sets of tires tested in the maneuvers on the front axle of the truck were new and the most effective was an all-season tire with the largest tread blocks and 45-degree grooves for water ejection. All sets of tires in the trailer maneuver had the same tread block arrangement, but the size of the blocks varied, with the most efficient tires having the smallest tread blocks.

The analysis of the accident‘s crash modification factor (CMF) from braking tests allowed the relative change in the safety of the vehicle fleet to be determined when the tire set with the shortest braking distance was taken as a reference. It shows that the use of less efficient tire sets than the reference tire set increases the change of the fatality rate in an accident by 19.7% on wet asphalt and by 26.6% on an ice-imitating pavement. The rate of change in injuries under the same conditions increases by 13.9% in an accident on wet asphalt and by 16.1% on artificially slippery surfaces.

The limitations of this study are due to the relatively small number of different tires (models) used, but the aim of the study was to identify the tire combinations used by one of the largest road haulage companies in Europe and to suggest best tire combinations. External factors such as wind or air temperature were not measured separately, but on the testing day, these factors did not differ and their effect was considered negligible. In addition, selected tire sets and truck type (including load and trailer configurations) do not completely represent global fleet compositions; however, the context of the research is best suited to the central and northern European regions.

The methodology described and agreed between the experiment and the CMF model can be applied in other cases, when it is desired to predict the positive or negative effect of any other factor on one of the accident rates. For example, in the case of a vehicle in bad technical condition (reduced braking performance), in the case of eco-driving mode selection (speed tolerance on downhill gradients with cruise control), or in the case of an independent study of the increase in reaction time of a fatigued driver. In our future research, additional tire models and maneuvers are planned at higher initial speeds for even more real traffic conditions typical for rural roads and highways where freight transportation is organized.