The Impact of Incorporated Nonwovens on the Surface Roughness of Sport Pitches

Abstract

Sport pitches have to meet high requirements and are highly maintained. To reduce the costs, vertically incorporated biodegradable nonwoven should raise water from the drainage layer to the root zone. To incorporate the nonwoven in existing sport pitches, a device was developed. In the present study, the variation of the roughness after the incorporation of the nonwoven was evaluated. Different measurement methods for the roughness of sport pitches were evaluated. A low roughness is important for the players to avoid injuries to the athletes. A point laser, an ultrasonic sensor and a feeler wheel were tested for their suitability for measuring the surface roughness of turf sport pitches. The sensors were fixed to a measurement frame and pulled over the test plots before and after the incorporation. The feeler wheel showed the lowest standard deviation and a significant difference after incorporating the nonwoven. These results indicate that the feeler wheel is best suited to determine the effects of the incorporation device in relation to the surface. Comparing the results of the feeler wheel indicates that the developed device increases the roughness of the surface, but the effects are minor.

1. Introduction

There are 35,993 sport pitches for soccer in Germany [1]. Sport pitches are intensively used and highly maintained. Normally, they are constructed by a layer-by-layer structure, according to the German standard DIN 18035-4 [2]. This structure consists of three layers as described from the top: rootzone, drainage and subgrade. The lowest layer is the subgrade, which consists of the existing subsoil. A drainage layer is installed when the subgrade is water impermeable. The turf grows on the root zone layer. The described construction method is comparable to that of the United States Golf Association (USGA), which is used on golf courses in the United States of America. The USGA drainage layer is made of gravel with a particle size distribution of 65% between 6.4 and 9.5 mm. The high water permeability is necessary to drain excess water from sports surfaces. At the same time, supplying turf grasses with sufficient water is a major challenge for the planning and construction of sports pitches [3]. Therefore, artificial irrigation is necessary.

To reduce the irrigation costs of a sport pitch built according to DIN 18035-4 [2], an improvement of the standard construction method was developed [4]. A vertically incorporated biodegradable nonwoven should raise water from the drainage layer to the root zone. To vertically incorporate the nonwoven in existing sport pitches, a device was developed that precuts the turf vertically, opens a slit, places the nonwoven vertically into the slit, closes it up and finally compacts the root area again [5]. This device literally causes a change in the surface due to the high interference with the turf.

A low roughness of the surface is important for the playability of the sport pitch. If the surface profile is not level, the players can be injured and the desired mowing height cannot be fulfilled. In addition, an increased roughness is associated with a decreasing ball rolling length. Roughness is defined by the deviation from the ideal surface of the object. On sport pitches, the roughness is the deviation of the profile from an ideally leveled sport surface. However, reviewing literature showed no significant research on the surface roughness for sport pitches with natural turf. This does not apply to artificial turf. Here, the roughness of the surface is of high importance for the playability and the injury rate in sports [6,7].

Much research exists on the measurement of road or field profiles. Description and classification of road surface profiles are standardized by ISO 8608 [8]. According to this standard, the fast Fourier analysis is used to calculate the power spectral density. This power spectral density is used to classify the roughness of a road in eight classes, but the standard does not specify an explicit measurement method [9]. The power spectral density evaluation is strongly dependent on the measurement method and the measurement frequency [10]. The international roughness index was established by the World Bank and is calculated by using a measured profile. It is smoothed by a quarter-car model of a so-called golden car with a speed of 80 km h⁻¹ [11]. Both indexes are used to assess the riding quality on road pavements. The presented study deals with the roughness of sport pitches that can be better compared to agricultural areas rather than to roads. The power spectral density according to ISO 8606 is even used to describe country roads or grass fields to evaluate the fatigue life of agricultural machinery [12]. However, according to the German standard DIN 18035-4, the evenness of a sport pitch is measured with a 4 m long leveling bar and a measuring wedge with a measurement resolution of 1 mm [2]. The unevenness of the sports pitch surface must be less than 20 mm over a length of 4 m [2]. This method is normally used for the acceptance of the work performed on a sport pitch, but not in research.

In laboratory studies, the roughness of soil surfaces is often measured with a pin meter or a roller chain, when using contact devices [13,14]. Noncontact measurements of soil surfaces are performed by photometers, infrared sensors, ultrasonic sensors, laser technics and satellite radar [13]. The major disadvantage of the first-mentioned contacting instruments is the deformation of the soil surface during the measurement. Noncontact measurements are widely used, but they require more technical and expensive equipment. Furthermore, the resolution accuracy of the height in airborne measurement methods is not sufficient, or measurement errors occur due to the turf [15]. In the present study, the surface of an experimental sport pitch needs to be evaluated. Because a standard soccer field has a size of 4000 to 10,000 m², the measurement would be time-consuming if using a pin meter. Therefore, in the present study, the measurements were performed by a measurement frame. It carries contact and noncontact sensors. A feeler wheel, an ultrasonic sensor and a laser pointer were attached to the measurement frame. The use of laser measurement methods with static apparatus for determining soil surfaces is widely known, but the use on turf-covered areas by using continuous measurement methods is a new approach [16].

There are many indexes or methods to compare the roughness of a soil surface. Following Gadelmawla et al. [17], the indices are divided into amplitude, spacing and hybrid parameters. The amplitude parameter describes the vertical characteristics and the spacing parameters measure the horizontal characteristics of the surface deviations. The hybrid parameters are a combination of both [17]. Amplitude parameters are the most common. The surface roughness of turf sports fields is described by the parameter of maximum roughness. This value is based on the depth gauge. It is the maximum deviation at one point from a straight line formed by two maximum points of the measured profile. If several measurements are taken on one sports field, the arithmetic mean of this maximum roughness is used to describe the evenness [2]. Therefore, the root mean square was used in the presented study to describe the changes of roughness on the experimental sport pitch. The main purpose of the present study was to evaluate the impact of incorporating a nonwoven into a sport pitch in relation to the surface roughness.

2. Materials and Methods

In the present study, the variation of the roughness after the incorporation of the nonwoven was evaluated. Additionally, this publication evaluates different measurement methods for the roughness of sport pitches. To compare the different measurement methods, a rectangular measurement frame was used (Figure 1). Its dimensions were 1600 mm in width and 1800 mm in length. It was made of square steel tubes with dimensions of 40 × 40 mm and consists mainly of four crossbars connected by longitudinal beams. The measurement frame had four free rotating wheels for driving. The wheels were continuously height-adjustable in order to adjust the measurement frame to the working distance of the sensors. The internal orientation of the measurement frame was measured by a VN-100 inertial measurement unit (manufactured by VectorNav Technologies, Dallas, TX, USA). The measurement resolution of the sensor is 0.05°, with repeatability equal to 0.2°. For a dynamic determination of pitch and roll, the deviation from the root mean square is 1°. The inertial measurement unit was attached to the rear crossbar of the measurement frame.

The external orientation of the measurement frame was determined with an M900 retroreflecting prism and a highly accurate robotic landscape surveying instrument (SPS930) (manufactured by Trimble, Sunnyvale, CA, USA). The retroreflecting triple prism was fixed to the third crossbar at a height of 430 mm above the measurement frame to plot the coordinates in a Cartesian coordinate system. By fusing the data of the inertial measurement unit and the Cartesian coordinates, the position of the measurement frame was calculated. Because the distances between the measurement frame and the sensors were known, the measured points could be integrated into the same coordinate system.

Three sensors were fixed on the measurement frame. The first one, pointing in the direction of travel, was a UM30-212113 ultrasonic sensor (manufactured by SICK AG, Waldkirch, Germany), with a frequency of 400 kHz. It can measure natural objects within an operating range of 65 to 350 mm. The accuracy is ±1%, with a resolution of more than 0.18 mm. The measured distance is exported via an analog output of 0 to 10 V with a 12-bit resolution. The measured distance was converted by an analog–digital transmitter with 16-bit resolution (RedLab 1608FS-PLUS) (manufactured by Meilhaus Electronic, Alling, Germany). The second sensor in the direction of travel was a long-range DT-500 distance sensor (manufactured by SICK AG, Waldkirch, Germany) with an operating range of 80 to 15,000 mm. It has a 12-bit resolution and an accuracy of ±3 mm. To reduce measurement errors, the laser beam was protected against sunlight with a tube.

Afterward, a feeler wheel was installed to measure the contact surface. The feeler wheel had a weight of 6100 g and an inflated rubber tire with a diameter of 280 mm. The tire had a width of 65 mm. Therefore, no deformation of the surface was expected and observed. Due to its weight, the feeler wheel follows the contour of the surface and pushes down the grass. The feeler wheel was mounted on a quadratic steel tube. This tube was guided by a larger quadratic square tube in height and could move without friction. The movements of the feeler wheel were measured from the frame by a midrange distance sensor. Because a reflector was installed on the feeler wheel, the distance sensor reached a resolution of 1 mm and an accuracy of ±10 mm. For data acquisition, a software program based on Microsoft C++ was used [12]. In postprocessing, the cubic spline interpolation routine in MATLAB (R2018b Update 2 Version 9.5.0.1033004) was used to synchronize the sensor data with the timestamps of the total station. By fusing the Cartesian coordinates of the measurement frame with the sensor data, the coordinates of the surface can be calculated. For the validation of the measuring setup, a wooden trapezoidal bump fixed to a level ground was measured. It had a height of 40 mm and a slope of 45 degrees on both sides. Then, the calculated height profile was compared with the height profile of the trapezoidal bump.

Figure 1. The measurement frame with the retroreflecting prism and the ultrasonic sensor, the feeler wheel and the long-range sensor. The retroreflecting prism has a distance measurement accuracy of ±(4 mm + 2 ppm). The ultrasonic sensor has an accuracy of ±1%, with a resolution of more than 0.18 mm. The long-range sensor has an accuracy of ±3 mm. The displacement of the feeler wheel according to the measurement frame was measured with an accuracy of ±10 mm by a midrange distance sensor.

Figure 1. The measurement frame with the retroreflecting prism and the ultrasonic sensor, the feeler wheel and the long-range sensor. The retroreflecting prism has a distance measurement accuracy of ±(4 mm + 2 ppm). The ultrasonic sensor has an accuracy of ±1%, with a resolution of more than 0.18 mm. The long-range sensor has an accuracy of ±3 mm. The displacement of the feeler wheel according to the measurement frame was measured with an accuracy of ±10 mm by a midrange distance sensor.

2.1. Description of the Trial Area

To measure the changes in the roughness of a sport pitch when incorporating a nonwoven, three plots were built on the fields of the experimental station Heidfeldhof at the University of Hohenheim (Figure 2). The plots were constructed following the German standard DIN 18035-4, with a 120 mm high drainage layer and a 130 mm high root zone layer. The plots were 4 m wide and 11 m long. The turf was seeded in October 2018 and mowed with a cutting height of 30 mm. The mowing was performed before the measurements, to avoid measurement errors due to the turf. The plots are surrounded by headland areas of natural soil, only seeded with turf. The plots were numbered from 1 to 3 ascending northwards. The terrain had a slope towards the north. In each of these plots, seven tracks of biodegradable nonwoven were incorporated vertically on 26 August 2019. The dimensions of the incorporated nonwoven were 150 mm times 5 mm (Figure 3). For this purpose, a machine was developed at the University of Hohenheim. It precuts the turf, forms a slit with a box colter, inserting the nonwoven into the slit. Finally, a pressure roller reconsolidates the surface. To evaluate the impact of the device on the surface, the roughness was measured before and after incorporating the nonwoven.

2.2. Method of Measurement

Ten measurement tracks were carried out on each plot with the measurement frame right-angled to the working direction of the device. The measurement tracks were spaced 0.7 m apart and 4 m long. In addition, four measurement tracks were measured in the same direction as the working direction of the device at a distance of 0.7 m from each other and a length of 11 m. The measurement frame was pulled by hand at a speed of about 500 m h⁻¹. The results presented showed the data collected by the rectangular and the longitudinal measurement tracks. Measurements were conducted on 25 August 2019 and 26 August 2019 after incorporating the nonwoven.

2.3. Postprocessing of the Data

The fused coordinates of the measurement frame were used to calculate the profiles of the surface for each sensor. First, the distance between the measured points was calculated according to the plane. This raster was performed with R Studio (Version 1.1.463). Afterward, the resulting profile was used for statistical analyses. According to Liu et al. [18] and Gadelmawla et al. [17], the root mean square roughness was calculated and used for analysis. The authors used the arithmetic mean line to calculate roughness. In the present study, the root mean square (RMS) is calculated with a linear regression because the plots had a slope. First, a linear regression was calculated by least-square method, and then the root mean square of the residuals was calculated and presented in the results. For the validation of the sensor frame, the RMS error was calculated following Equation (1):

$$RM{S}_{error}=\sqrt{\frac{{{{\displaystyle \sum }}_{i=1}^n\left(z_{r,i}-z_{tr,i}\right)}^2}{n}}$$

(1)

where $z_{r,i}$ is the measured profile height, $z_{tr,i}$ the true profile height of the trapezoidal bump at the position I and n is the number of measurement points. The $RM{S}_{error}$ was only calculated at those positions where the trapezoidal bump was located.

The mean $RM{S}_{roughness}$ for the roughness of the measured profiles was calculated using the following equation:

$$RM{S}_{roughness}=\sqrt{\frac{{{{\displaystyle \sum }}_{t=1}^n\left({\widehat{z}}_t-z_t\right)}^2}{n}}.$$

(2)

The $RM{S}_{roughness}$ was calculated with ${\widehat{z}}_t$, the estimated profile height of the linear regression, and $z_t$, the measured profile height at the position t. The number of the measured point is presented as n. The results presented were calculated from n = 750 measurement points. The arithmetic means and the standard deviation of the $RM{S}_{roughness}$ were calculated for the 14 measurement tracks of each of the three plots.

3. Results

For validation of the sensor frame, the measured profiles and the real profile of the trapezoidal bump are presented (Figure 4). The means of the measurements (n = 5) recorded by the feeler wheel, the distance sensor and the ultrasonic sensor are shown. Using Equation (1), the $RM{S}_{error}$ was calculated for the location of the trapezoidal bump. The $RM{S}_{error}$ was 1.8 mm for the profile measured by the feeler wheel. The results of the distance sensor showed an $RM{S}_{error}$ of 7.7 mm according to the profile of the trapezoidal bump. The ultrasonic sensor measured the highest variation to the real profile of the trapezoidal bump, with an $RM{S}_{error}$ of 9.9 mm. The ultrasonic sensor showed a high amount of variation due to the low reflection. However, the error was lower than 10 mm. The presented profiles are the results of one measuring track of the first plot measured by the feeler wheel before and after the incorporation of the nonwoven (Figure 5). The changes in the roughness are obviously lower than 20 mm, and a direction of the variation is not visible in the presented data.

Impact of Incorporating the Nonwoven in the Trial Area

In the presented study, the impact of incorporating a nonwoven in relation to the surface roughness was also investigated. Figure 6 presents the variation in surface roughness caused by incorporation. The $RM{S}_{roughness}$ calculated according to Equation (2) was used to compare the results of the feeler wheel. Each bar of the bar chart presents the mean $RM{S}_{roughness}$ of 14 replications measured on each of the three plots. The error bars indicate a standard deviation of the $RM{S}_{roughness}$ for the 14 replications. The black-colored bar describes the $RM{S}_{roughness}$ before the incorporation of the nonwoven, whereas the grey bar stands for the mean $RM{S}_{roughness}$ after the incorporation. A paired Tukey test for the roughness of all replications was performed. The p-value of 0.0375 indicates a significant change in the mean values of the $RM{S}_{roughness}$ after incorporation. Figure 6 allows the conclusion to be drawn that the surface roughness has been increased, affected by the installation of the nonwoven. For example, the $RM{S}_{roughness}$ in plot 3 increased from 0.008 to 0.011 due to the installation of the fleece. The high standard deviation in the repetitions indicates that the measured plots have a high heterogeneity of surface roughness. Figure 6 shows that the values for $RM{S}_{roughness}$ in plot 1 are lower than those in plots 2 and 3.

Data for the roughness of the three plots measured with the long-range sensor are presented in Figure 7. The results are shown as the mean $RM{S}_{roughness}$ before and after the incorporation of the nonwoven, following Equation (2). Figure 7 describes that the incorporation of the nonwoven has increased the roughness of the surface in plot 1 and plot 2. Statistics performed by a paired Tukey test describe no significant difference between the means of all measurements before and after the incorporation. This is mainly caused by the high standard deviation of the first plot. The high standard deviation may result from measurement errors due to the reflection of the laser beam on the turf. From the calculated mean $RM{S}_{roughness}$ presented in Figure 8 it is obvious that the profile measured by the ultrasonic sensor detects no changes in the roughness. A day before the incorporation of the nonwoven, the roughness was about 10 to 12 mm, presented as $RM{S}_{roughness}$. After the device had incorporated the nonwoven into the sport pitch, the surface shows a higher roughness with an increase of about 3 mm. The standard deviation is much lower compared to the measurements performed with the long-range sensor. The calculated p-value for the paired Tukey test is about p = 0.24 when comparing the means of the measurements performed before and after the treatment. As a result, no significant changes in the surface roughness can be detected. The ultrasonic sensor has a resolution of 0.18 mm, but due to the vegetation on the sport pitch, it is possible that the measurement errors increased. This assumption can also be transferred to the long-range sensor. In addition, the ultrasonic sensor covers a larger area, increasing the errors of the measurements.

Figure 6. Arithmetic mean $RM{S}_{roughness}$ (m) of the measured profile before and after the incorporation of the nonwoven. n = 14. The measurement was performed by the feeler wheel. The error bar is the standard deviation. The p-value shows the paired Tukey test for mean $RM{S}_{roughness}$ before and after the incorporation calculated for all three plots.

Figure 6. Arithmetic mean $RM{S}_{roughness}$ (m) of the measured profile before and after the incorporation of the nonwoven. n = 14. The measurement was performed by the feeler wheel. The error bar is the standard deviation. The p-value shows the paired Tukey test for mean $RM{S}_{roughness}$ before and after the incorporation calculated for all three plots.

4. Discussion

In this section, the results are analyzed and discussed. Three sensors were compared during the experiment. The feeler wheel showed the lowest standard deviation and a significant difference after incorporating the nonwoven. These results indicate that the feeler wheel is best suited to determine the effects of the incorporation device in relation to the surface. The feeler wheel weighs 6100 g and therefore avoids measurement errors due to vegetation because it presses down the grass leaves. However, the weight is low enough that no compaction of the soil surface occurs. This fact is evident from the significant difference between the treatments. The measurements of the ultrasonic and long-range sensors showed no significant difference and a high standard deviation. In particular, the long-range sensor showed high scattering in the values measured on the first plot (Figure 5). The reason for this could be that the leaf mass of the grasses deflects the laser beam. This ultrasonic sensor had already been tested to measure the working quality of tillage implements such as the chisel plow [19]. The circular measurement area of the ultrasonic sensor has a suitable size. However, a soil surface is significantly rougher than the tested sports surface. Still, a significant increase in the sports turf surface is evident as a result of the incorporation process. This is mainly due to the opening of the soil with the aid of the colter, which pushes the soil open and cannot be completely leveled by the pressure rollers. In the same way, the cutting of the turf cover has an influence on the surface change.

The measured profiles were compared to the $RM{S}_{roughness}$. Considering the measurement results of the feeler wheel, it is evident that there is a significant increase in roughness. The question now arises as to how this change is to be assessed. In the German standard DIN 18035-4, a maximum deviation of 30 mm at 4 m measuring bar length is specified. Since this deviation is only measured at one point, the $RM{S}_{roughness}$ according to Equation (2) is equal to the root of 30 mm. The maximum roughness is about 5.4 mm. According to the results measured before the incorporation, the roughness is higher than the standard. However, the difficulty is the comparability of the two measuring methods. In addition, the vegetation causes high roughness due to the uneven height of the material build-up caused by the turf (e.g., by lawn thatch [20]). To solve this problem, the turf could have been groomed and top-dressed before the trials started. However, to investigate the influence of the device incorporating the nonwoven, the difference between before and after installation is crucial. If comparing the results of the feeler wheel, the difference is significant, but it is less than 5 mm. This indicates that the developed device increases the roughness of the surface. The measurement results indicate that it is necessary to improve the incorporation device and, for example, to equip it with a heavy pressure roller.

5. Conclusions

A low roughness is one of the central requirements for the good quality of sport pitches. Therefore, it is important to develop maintenance equipment that does not cause any negative changes to the surface. There has been little research on the roughness of turf sports pitches so far. A measurement frame was used to determine the surface roughness before and after the vertical incorporation of a nonwoven in three plots, built according to the German standard DIN 18035-4. In common research, a leveling bar is used for measuring surface roughness [21]. Nevertheless, from the presented studies, the following remarks can be provided:

• The validation of the measurement frame showed that the root mean square error between the real surface and the measured surface was below 10 mm. Therefore, the measurement method was considered suitable.
• The measurements performed by the feeler wheel showed a significant increase in roughness after the incorporation. The results indicate that only the feeler wheel can be used to continuously determine the roughness of a sport pitch because a low standard deviation was observed.
• The device for incorporation of the nonwoven in existing sport pitches must be improved to avoid increasing the roughness of the turf surface.