1. Introduction
Agricultural tractor performance is greatly affected by the draft force during tillage operations. The draft force is one of the most widely used indicators for evaluating the performance of agricultural tractors [
1]. Studies analyzing and predicting the traction performance of agricultural tractors have used either the analysis method or the empirical method [
2]. Of the two methods, the empirical method based on soil physical properties is relatively widely used because it can easily measure the mobility of the vehicle compared to the analysis method [
3]. In general, many parameters affect the draft force of an agricultural tractor during tillage operations, such as soil texture, soil water content, traveling speed with slip ratio, and tillage depth, as well as soil strength in terms of cone index [
4].
Many studies have been conducted on the factors affecting the draft force of agricultural tractors. In particular, various studies have been conducted on the effect of soil physical properties on draft force. A field experiment was conducted to measure the specific draught and energy use under different tillage implements and soil conditions [
5,
6]. The results showed that the specific draught was the lowest for the moldboard plow; conversely, it was the highest in chisel plows, and increased with decreasing soil water content. In another study, the specific draft estimation model for offset disc harrows was developed considering the front gang angle, cone index, and forward speed [
7]; the results showed that a front gang angle of 35 degrees was required to minimize drafts. The effect of shank geometry and gear selection on tractor performance was analyzed during deep tillage operations [
8]; the results showed that as the gear selection increased, the axle torque, draft force, and fuel consumption increased sharply. The study investigated the effect of soil texture and soil water content on the performance of agricultural tractors based on the USDA (U.S. Department of Agriculture) standard in upland fields [
9]. The results showed that the soil texture parameter was between 1.4 and 1.5, which is higher than the guidelines set by the ASABE (American Society of Agricultural and Biological Engineers) standard, which prescribes 0.45–0.7 for the same draft force. In another study, an experiment was conducted to analyze the effect of implement surface coating on draft force during tillage operations [
10]; the results showed that the modification of the furrower tines with an ultra-high molecular weight polyethylene (UHMW-PE) coating can significantly reduce the draft force. However, analytical studies on the influence of tillage depth on draft force and workload have been limited due to the difficulty of measurement.
Among the many factors that affect the loading of agricultural machinery, the tillage depth is key in the performance evaluation of the tractor because an actual paddy field has different soil properties according to the soil layers. The tillage depth affects the agricultural ecosystem and has a significant impact on crop yields and quality [
11,
12,
13,
14]. In theory, the target tillage depth is usually determined by the distribution of the hardpan. During plow tillage, the tractor attachment implement must penetrate deeper than the top hardpan depth during tillage. The formation of hardpan in the soil layer affects the soil compaction, which disturbs the root growth and permeation. Therefore, the plow tillage must be performed at the depth of the hardpan layer that is formed within the plow layer to determine that the tillage operation was performed at the appropriate depth. By analyzing the results of the cone penetration test, the depth of the cone index value within the plow layer is determined. This is assumed to be the average depth where the hardpan layer begins to be distributed, which is called the depth of the top hardpan [
15,
16].
Conventionally, tillage depth has been measured manually using a steel ruler. Some parameter studies including tillage depth related to tillage operation have been limited in some soil bin tests. Experiments were conducted to evaluate the performance of the soil strength profile sensor using soil bin and field data from different depth conditions [
17]; the test results showed that the prismatic soil strength index was higher at locations with lower EC, lower water content, and greater bulk density. In other studies, experiments were performed to evaluate the effect of the design parameters of agricultural machinery on the draft force, rolling resistance, and soil compaction [
18,
19,
20,
21]. Although some studies have been conducted on the tillage depth performed in actual farm-land, it was not possible to conduct a precise test by roughly classifying the depth as shallow (0–0.15 m) and deep (0.15–0.3 m) or according to the soil layer [
22,
23]. In previous studies, tillage depth measurement was very labor-intensive and time-consuming. The bigger problem was that finding the average depth was not possible because continuous real-time measurement was not possible; the tillage depth was only available at some points [
24]. As a result, it was not possible to analyze the effects of momentary changes in tillage depth, which would have been carried out under very rough test conditions. Until now, it has been difficult to measure tillage depth in real-time, so we were limited to analyzing the effect of tillage depth on draft force during tillage operations. So, the development and application of a real-time measurement system for tillage depth is needed. However, since it is difficult to verify the accuracy of continuous measurements during actual tillage, it is necessary to verify it using a formula for the interaction between the tillage depth and the draft force. Therefore, in this study, verification of the accuracy of the tillage depth measurement system was performed through an analysis of the effect of tillage depth on draft force during plow tillage.
The objectives of this paper are to propose improved measurement methods for tillage depth and to verify the accuracy of the developed measurement system through field experiments. The specific objectives were: (1) to develop a real-time tillage depth measurement system based on the sensor fusion method; (2) to continuously measure the tillage depth and the draft force during plow tillage; and (3) to confirm the accuracy of the real-time tillage depth measurement system by comparing it with the predicted draft force calculated by the ASABE standards equation.
4. Discussion
The accuracy of the real-time tillage depth measurement system was analyzed during plow tillage by comparing the measured draft force through the field test and the predicted value using the ASABE standard equation. The results of a linear regression using a F
i value of 0.7 for the two gear stages are shown in
Figure 10. The accuracy of the tillage depth measurement system, defined as the coefficient of determination and indicated by the regression analysis between measured and predicted draft force values, tended to vary depending on the working conditions during plow tillage. In M3–low, results of linear regression showed a high coefficient of determination value of 0.847 and low RMSE (Root Mean Square Error) of 1.139, as shown in
Figure 10a. In addition, the coefficient of determination value was close to 1 in the deep tillage depth section where high draft force is generated. The high accuracy in the deep tillage depth range is influenced by the fact that the traveling speed is not significantly reduced even in the deep tillage depth section in M3–low. On the other hand, results of linear regression at M3–high showed lower system accuracy than M3–low with a coefficient of determination of 0.549 and RMSE of 1.497, as shown in
Figure 10b. This has a relatively faster theoretical speed compared to M3–low at M3–high, and is believed to have been affected by the rapid deceleration of the traveling speed due to the large soil resistance at the deep tillage depth section.
In actual plow tillage, the farmer could use both gears, so the measured draft force data from the two gears used in the field test were integrated and analyzed for further linear regression. As can be seen from
Figure 11, it shows a coefficient of determination of 0.715 and RMSE of 1.375, which is smaller relative error than M3–high but larger than M3–low. The accuracy of the real-time tillage depth measurement system, which shows an R-Square value of 0.715, is improved compared to the accuracy of the ASABE standard prediction equation with an error range of
40%, and it can be used in various studies on tillage depth and draft force. The deviation between measured and prediction can be attributed to: (1) the limitation caused by roughly distinguishing only three dimensionless soil texture adjustment parameter (F
i) values, even though the soil texture is classified into various types, (2) use of different types of agricultural tractors and implements, and (3) soil properties differences present during the field experimental process (e.g., inconsistence of moisture content, hardpan, and soil surface roughness between working paths). Overall, the deeper the cultivation depth in both gear stages, the smaller the difference between the predicted and measured values, resulting in improved system accuracy. It is believed that the system accuracy is reduced in deep tillage depth conditions, where the traveling speed is reduced due to soil resistance. In conclusion, the accuracy of the draft force prediction using the real-time tillage depth measurement system was greatly influenced by the traveling speed and soil penetration depth of the attachment implement. Therefore, the tillage depth should be considered when analyzing the performance agricultural tractors. It is very important to know the depth at which the tillage operation was performed to ensure tractor design reliability. Through this study, system accuracy with an R-square value of 0.715 could be obtained using the developed real-time tillage depth measurement system. This system is expected to be of great help for future research on traction force prediction according to the working conditions.
5. Conclusions
In this study, a real-time tillage depth measurement system was developed for continuous measurement during plow tillage to analyze the effect of tillage depth on draft force using a sensor fusion method. The real-time tillage depth measurement system developed in this study selected the sensor fusion type A (linear potentiometer, inclinometer), suitable for actual tillage operations according to a soil non-penetration test. The selected sensor fusion method was used to continuously measure tillage depth and draft force during plow tillage. Finally, the accuracy of the real-time tillage depth measurement system was evaluated by comparing the measured draft force and predicted draft force based on the ASABE standard. The main results of this study are as follows.
- (1)
The real-time tillage depth measurement system was developed through sensor fusion consisting of a linear potentiometer, an optical distance sensor, and an inclinometer. A simple soil non-penetration test bed was constructed considering the actual tillage depth of moldboard plow. As a result of the accuracy verification test, sensor fusion type A showed a 2% measurement error, which was 9% lower than the sensor fusion type B, using the optical distance sensor and the inclinometer, which had an 11% measurement error. In addition, sensor fusion type A showed only about 3% measurement error in actual moldboard plow operations. Therefore, sensor fusion type A is a suitable method for measuring the tillage depth during actual tillage operations. In addition, the system can only be measured with relative vertical displacement, which makes it possible to use it for any type of tillage operation.
- (2)
To verify the accuracy of the real-time tillage depth measurement system in the plow tillage, a comparative test between the measured draft force and the predicted values based on the ASABE standard equation was conducted. In addition, soil texture analysis of the test field was performed to select Fi (dimensionless soil texture adjustment parameter), a key factor used in the draft force prediction. As a result of the soil texture analysis, the soil of the test field was classified as loam, and it was confirmed that the Fi value of 0.7 should be used when using the draft force prediction equation of ASABE standard as the soil texture of the medium texture group.
- (3)
Field test results showed that the overall average traveling speed and draft force values of the two gear stages were similar, but are greatly affected by the tillage depth. Overall, the higher the gear selection, the deeper the tillage depths, which will result in greater soil resistance, slowing the traveling speed while increasing the average draft force. The regression analysis for system accuracy analysis showed that the deeper the tillage depth in each gear, the closer the coefficient of determination was to 1. Specifically, the lower the gear selection and the deeper the tillage depth during plow tillage, the higher the coefficient of determination as a result of the regression analysis of the measured values. This is thought to be influenced by the relatively rapid decrease in traveling speed due to greater soil resistance in the deep tillage depth section of the high gear stage M3–high. Therefore, in the case of no reduction in traveling speed due to soil resistance even in the deep tillage depth section during plow tillage, it was seen that the deeper the tillage depth and the greater the draft force, the higher the accuracy of the draft force prediction using the real-time tillage depth measurement system.
Although some aspects need to be improved, this tillage measurement method is judged to have a wide utilization range for research into the effect of the tillage depth on the loads of agricultural tractor during tillage operations. It is judged that this will enable the study of the effects of improved measurement precision and machine reliability on the local soil, subject to future study and testing, and that it will be useful for the study of traction performance and design load of tractors considering the agricultural working environment.
In conclusion, the effect of the tillage depth on the draft force of agricultural tractors during moldboard plow operations was confirmed with the proposed measurement system configuration. The results of this study will be useful for field test condition setting and data analysis of the measured workload to be used for the optimal design of agricultural tractors. In future studies, the soil–machine parameters according to tillage depth will be analyzed by field experiments when attaching other tillage implements, and a discrete element method model for agricultural machinery will be developed to conduct a parameter simulation on the factors affecting the design load of agricultural tractor and model verification by field demonstrations.