Dikey Döner Toprak İşlemede Tork, Çeki Kuvveti ve Güç Tüketiminin Makine Öğrenimi Tabanlı Tahmini

Öz

Toprak işleme makineleri, toprağın havalandırılması, parçalanması ve karıştırılması yoluyla bitkisel üretimde önemli bir role sahiptir ve genellikle yüksek enerji gereksinimiyle çalışırlar. Mevcut ekipmanlar arasında dikey toprak frezeleri, yoğun toprak–alet etkileşimi ve geleneksel toprak işleme ekipmanlarına kıyasla sahip oldukları enerji verimliliği avantajı nedeniyle tohum yatağı hazırlığında yaygın olarak tercih edilmektedir. Ancak dikey toprak frezelerinin tork, çeki kuvveti ve güç tüketimi; rotor devri ve bıçak geometrisi gibi işletim parametrelerinden büyük ölçüde etkilenmekte olup, bu parametrelerin doğru bir şekilde tahmin edilmesi makine tasarımının ve enerji kullanımının optimize edilmesi açısından önemli bir zorluk olmaya devam etmektedir. Bu çalışmada, kontrollü laboratuvar koşullarında bir toprak kanalında yürütülen deneyler ile dikey toprak frezesinde rotor devri (220, 270 ve 310 rpm), yatay bıçak açıları (0°, 5°, 10° ve 15°) ve dikey bıçak açıları (0°, 10°, 20°, 30° ve 40°) parametrelerinin tork, çeki kuvveti ve güç tüketimi üzerindeki etkileri incelenmiştir. Tork ve çeki kuvveti ölçümleri sırasıyla özel olarak tasarlanmış bir tork ölçüm sistemi ve yük hücresi kullanılarak gerçekleştirilmiş, toprak penetrasyon direnci 0.2–0.5 MPa ve toprak nem içeriği %18 ± 1 seviyesinde sabit tutulmuştur. Tam faktöriyel deney tasarımından elde edilen veri seti, Gradient Boosting (GB), Light Gradient Boosting Machine (LightGBM), Karar Ağaçları (DT), K-En Yakın Komşular (KNN), Yapay Sinir Ağları (ANN) ve Destek Vektör Makinesi (SVM) olmak üzere altı farklı makine öğrenmesi modeli ile analiz edilmiştir. Model performansı R², MAPE, MAE ve RMSE metrikleri ile değerlendirilmiş, genelleme yeteneğini değerlendirmek amacıyla 5 katlı çapraz doğrulama yöntemiyle desteklenmiştir. Sonuçlar, Gradient Boosting modelinin tüm çıktı parametrelerinde en yüksek tahmin doğruluğunu sağladığını (R² değeri 0.999'a kadar), Karar Ağaçları ve LightGBM modellerinin de düşük hata değerleriyle tutarlı ve yüksek doğruluk sergilediğini, buna karşın ANN ve SVM modellerinin özellikle çeki kuvveti ve güç tüketimi tahmininde nispeten daha yüksek hata oranları gösterdiğini ortaya koymuştur. Bu bulgular, ağaç tabanlı makine öğrenmesi modellerinin dikey toprak frezelerinde tork, çeki kuvveti ve güç tüketimi gibi temel performans parametrelerinin tahmininde uygulanabilir ve güvenilir olduğunu doğrulamaktadır.

Anahtar Kelimeler

• Dikey toprak freze
• Gradyan artırma regrasyonu
• Karar ağaçları
• Makine öğrenmesi
• Tork
• Çeki kuvveti
• Toprak kanalı

Etik Beyan

There is no need to obtain permission from the ethics committee for this study.

Machine Learning-Based Prediction of Torque, Draft Force and Power Consumption in Vertical Rotary Tillage

Öz

Tillage machines play a critical role in agricultural production by improving soil structure through aeration, disintegration, and mixing, while typically operating under high energy demand. Among the available implements, vertical rotary tillers are widely preferred for seedbed preparation owing to their intensive soil–tool interaction and favorable energy efficiency compared with conventional tillage equipment. However, the torque, draft force, and power consumption of vertical rotary tillers are strongly influenced by operational parameters such as rotor speed and blade geometry, and accurately predicting these parameters remains a key challenge for optimizing machine design and energy use. This study investigates the effects of rotor speed (220, 270, and 310 rpm), horizontal blade angles (0°, 5°, 10°, and 15°), and vertical blade angles (0°, 10°, 20°, 30°, and 40°) on torque, draft force, and power consumption of a vertical rotary tiller under controlled laboratory soil bin conditions. Torque and draft force were measured using a custom-designed torque measurement system and a load cell, respectively, while soil penetration resistance and moisture content were maintained at 0.2–0.5 MPa and 18% ± 1%. The experimental dataset, obtained from a full factorial design, was analyzed using six trained machine learning models: Gradient Boosting (GB), Light Gradient Boosting Machine (LightGBM), Decision Tree (DT), K-Nearest Neighbors (KNN), Artificial Neural Networks (ANN), and Support Vector Machine (SVM). Model performance was evaluated using R², MAPE, MAE, and RMSE metrics, supported by 5-fold cross-validation to assess generalization capability. The results showed that Gradient Boosting achieved the highest predictive accuracy across all output parameters (R² up to 0.999), while Decision Tree and LightGBM models also produced consistently high accuracy with low error metrics, whereas ANN and SVM exhibited comparatively higher prediction errors, particularly for draft force and power consumption. These results confirm the applicability and reliability of machine learning models, particularly GB and DT, for predicting key performance parameters in vertical rotary tillage under varying operational conditions.

Anahtar Kelimeler

• Vertical rotary tiller
• Gradient boosting regression
• Decision tree
• Machine learning
• Torque
• Draft force
• Soil bin

Etik Beyan

There is no need to obtain permission from the ethics committee for this study.

Kaynakça

1. Aase, J. K. and Siddoway, F. H. (1982). Evaporative flux from wheat and fallow in a semiarid climate. Soil Science Society of America Journal, 46(3): 619–626.
2. Aikins, K. A., Barr, J. B., Ucgul, M., Jensen, T. A., Antille, D. L. and Desbiolles, J. M. A. (2020). No-tillage furrow opener performance: a review of tool geometry, settings and interactions with soil and crop residue. Soil Research, 58(7): 603.
3. Almeida, W. S. de, Panachuki, E., Oliveira, P. T. S. de, Da Silva Menezes, R., Sobrinho, T. A. and Carvalho, D. F. de (2018). Effect of soil tillage and vegetal cover on soil water infiltration. Soil and Tillage Research, 175: 130–138.
4. Alvarez, R. and Steinbach, H. S. (2009). A review of the effects of tillage systems on some soil physical properties, water content, nitrate availability and crops yield in the Argentine Pampas. Soil and Tillage Research, 104(1): 1–15.
5. Arefi, M., Karparvarfard, S. H., Azimi-Nejadian, H. and Naderi-Boldaji, M. (2022). Draught force prediction from soil relative density and relative water content for a non-winged chisel blade using finite element modelling. Journal of Terramechanics, 100: 73–80.
6. Bolton, F. E. and Booster, D. E. (1981). Strip-Till Planting in Dryland Cereal Production. Transactions of the ASAE, 24(1): 59–62.
7. Borin, M., Menini, C. and Sartori, L. (1997). Effects of tillage systems on energy and carbon balance in north-eastern Italy. Soil and Tillage Research, 40(3-4): 209–226.
8. Boydaş, M. G. (2017). Determination of the effects of different wing structures, forward speed, and working depth on draft force in a winged chisel plough. Mediterranean Agricultural Sciences, 30(3): 219–225. (In Turkish)

9. Boydaş, M. (2023). Determination of the effect of cultivator shares with different tip structures on draft force using the finite element method. Journal of Tekirdag Agricultural Faculty, 20(2): 306–317. (In Turkish)
10. Çarman, K. (1997). Effect of different tillage systems on soil properties and wheat yield in Middle Anatolia. Soil and Tillage Research, 40(3-4): 201–207.
11. Çay, A., Özpınar, S. and Aydın, A. (2017). Effects of long-term reduced tillage and green manure practices under different crop rotations on sunflower growth and weed infestation. Journal of Tekirdag Agricultural Faculty, 14(3): 116–126. (In Turkish)
12. Çelik, A. and Erkmen, Y. (1999). Effects of soil tillage with different blade types on field emergence and yield of winter wheat in a rotary tiller. Atatürk University Journal of Agricultural Faculty, 30(1): 55–62. (In Turkish)
13. Daraghmeh, O. A., Petersen, C. T., Munkholm, L. J., Znova, L., Obour, P. B., Nielsen, S. K. and Green, O. (2019). Impact of tillage intensity on clay loam soil structure. Soil Use and Management, 35(3): 388–399.
14. Das, S., Dey, A., Pal, A. and Roy, N. (2015). Applications of artificial intelligence in machine learning: review and prospect. International Journal of Computer Applications, 115(9): 31-41.
15. De Vita, P., Di Paolo, E., Fecondo, G., Di Fonzo, N. and Pisante, M. (2007). No-tillage and conventional tillage effects on durum wheat yield, grain quality and soil moisture content in southern Italy. Soil and Tillage Research, 92(1-2): 69–78.
16. Friedman, J. H. (2001). Greedy function approximation: a gradient boosting machine. Annals of Statistics, 29(5): 1189–1232.
17. Gautam, P. V., Agrawal, K. N., Roul, A. K., Mansuri, S. M. and Subeesh, A. (2024). Predictive modelling of sweep's specific draft using machine learning regression approaches. Soil Use and Management, 40(1): 1-16.
18. He, C., Guo, Y., Guo, X. and Sang, H. (2023). A mathematical model for predicting the draft force of shank-type tillage tine in a compacted sandy loam. Soil and Tillage Research, 228: 1-12.
19. Hendrick, J. G. (1980). A powered rotary chisel. Transactions of the ASAE, 23(6): 1349–1352.
20. Johansen, C., Haque, M. E., Bell, R. W., Thierfelder, C. and Esdaile, R. J. (2012). Conservation agriculture for small holder rainfed farming: Opportunities and constraints of new mechanized seeding systems. Field Crops Research, 132: 18–32.
21. Karlen, D. L., Wollenhaupt, N. C., Erbach, D. C., Berry, E. C., Swan, J. B., Eash, N. S. and Jordahl, J. L. (1994). Long-term tillage effects on soil quality. Soil and Tillage Research, 32(4): 313–327.
22. Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W. and Liu, T. Y. (2017). LightGBM: A Highly Efficient Gradient Boosting Decision Tree. 31st Conference on Neural Information Processing Systems (NIPS 2017). 4-9 December, p. 3149-3157, Long Beach, CA, U. S. A.
23. Kumar, S. P., Tewari, V. K., Chandel, A. K., Mehta, C. R., Pareek, C. M., Chethan, C. R. and Nare, B. (2023). Modelling specific energy requirement for a power-operated vertical axis rotor type intra-row weeding tool using Artificial Neural Network. Applied Sciences, 13(18): 10084.
24. Li, T., Xie, H., Ren, Z., Hou, Y., Zhao, D., Wang, W., Wang, Z., Liu, Y., Wen, X., Han, J., Mo, F. and Liao, Y. (2023). Soil tillage rather than crop rotation determines assembly of the wheat rhizobacterial communities. Soil and Tillage Research, 226: 1-11.
25. Mahore, V., Soni, P., Paul, A., Patidar, P. and Machavaram, R. (2024). Machine learning-based draft prediction for mouldboard ploughing in sandy clay loam soil. Journal of Terramechanics, 111: 31-40.
26. Mandal, S. K., Bhattacharyya, B., Mukherjee, S. and Prasad, A. K. (2016). Design optimization of rotary tiller blade using specific energy requirement. International Journal of Current Engineering and Technology, 6(4): 1257–1263.
27. Manuwa, S. I. (2009). Performance evaluation of tillage tines operating under different depths in a sandy clay loam soil. Soil and Tillage Research, 103(2): 399–405.
28. Nagar, H., Machavaram, R., Ambuj, Soni, P., Saha, S. and Bose, T. S. C. (2024). An integrated cloud system based serverless android app for generalised tractor drawbar pull prediction model using machine learning. Systems Science and Control Engineering, 12(1): 2385332.
29. Ogle, S. M., Breidt, F. J. and Paustian, K. (2005). Agricultural management impacts on soil organic carbon storage under moist and dry climatic conditions of temperate and tropical regions. Biogeochemistry, 72(1): 87–121.
30. Oliveira, I. N. de, Souza, Z. M. de, Bolonhezi, D., Totti, M. C. V., Moraes, M. T. de, Lovera, L. H., Lima, E. d. S., Esteban, D. A. A. and Oliveira, C. F. (2022). Tillage systems impact on soil physical attributes, sugarcane yield and root system propagated by pre-sprouted seedlings. Soil and Tillage Research, 223: 105460.
31. Öztürk, K. and Şahin, M. E. (2018). An overview of artificial neural networks and artificial intelligence. Takvim-i Vekayi, 6(2): 25–36. (In Turkish)
32. Roul, A. K., Raheman, H., Pansare, M. S. and Machavaram, R. (2009). Predicting the draught requirement of tillage implements in sandy clay loam soil using an artificial neural network. Biosystems engineering, 104(4): 476-485.
33. Šarauskis, E., Buragienė, S., Romaneckas, K., Masilionytė, L., Kriaučiūnienė, Z., Sakalauskas, A., Jasinskas, A. and Karayel, D. (2014). Deep, shallow and no-tillage effects on soil compaction parameters. Engineering for Rural Development, 13: 31–36.
34. Sarker, I. H., (2021). Machine learning: Algorithms, real-world applications and research directions. SN Computer Science, 2(3): 160.
35. Selvi, K., Sağlam, M. and Dengiz, O. (2017). Geostatistical approach to determine the effects of different soil tillage methods on penetration resistance in a clayey soil. INMATEH Agricultural Engineering, 52(2): 19-26.
36. Smola, A. J. and Schölkopf, B. (2004). A tutorial on support vector regression. Statistics and Computing, 14: 199-222.
37. Song, Y., Liang, J., Lu, J. and Zhao, X. (2017). An efficient instance selection algorithm for k nearest neighbor regression. Neurocomputing, 251: 26-34.
38. Srivastava, A. K., Goering, C. E., Rohrbach, R. P. and Buckmaster, D. R. (2006). Soil Tillage. In: Engineering Principles of Agricultural Machines, Ed(s): Srivastava, A. K., Carroll, E. Rohrbach, R. P., Buckmaster, D. R., American Society of Agricultural and Biological Engineers, U. S. A.
39. Sukcharoenvipharat, W., Usaborisut, P. and Choedkiatphon, S. (2017). Comparative study on soil tillage using rotary tiller and power harrow. International Journal of Agricultural and Biosystems Engineering, 11(10): 744-747.
40. Tafesse, M., Intaravichai, S., Saitthiti, B. and Kiatiwat, T. (2007). The optimum design parameters in terms of total specific energy requirements for the rotary blade power tiller under unsaturated sandy clay loam soil condition. Kasetsart Journal: Natural Science, 41: 775-781.
41. Tapela, M. and Colvin, T. S. (2002). Quantifying seedbed condition using soil physical properties. Soil and Tillage Research, 64(3-4): 203-210.
42. Varani, M., Mattetti, M., Molari, G., Biglia, A. and Comba, L. (2023). Correlation between power harrow energy demand and tilled soil aggregate dimensions. Biosystems Engineering, 225: 54–68.
43. Wang, X., Zhou, H., Wang, S., Zhou, H. and Ji, J. (2023). Methods for reducing the tillage force of subsoiling tools: A review. Soil and Tillage Research, 229: 1-14.
44. Wesołowska, S., Daniłkiewicz, D., Gawęda, D., Haliniarz, M., Rusecki, H. and Łukasz, J. (2022). The effect of tillage systems and weed control methods on the yield and quality of spelt grain (Triticum aestivum ssp. spelta L.). Agriculture, 12(9): 1-17.
45. Xu, M., Watanachaturaporn, P., Varshney, P. K. and Arora, M. K. (2005). Decision tree regression for soft classification of remote sensing data. Remote Sensing of Environment, 97(3): 322-336.
46. Yao, R., Gao, Q., Liu, Y., Li, H., Yang, J., Bai, Y., Zhu, H., Wang, X., Xie, W. and Zhang, X. (2023). Deep vertical rotary tillage mitigates salinization hazards and shifts microbial community structure in salt-affected anthropogenic-alluvial soil. Soil and Tillage Research, 227: 1-17.
47. Zhai, S., Shi, Y., Zhou, J., Liu, J., Huang, D., Zou, A. and Jiang, P. (2022). Simulation optimization and experimental study of the working performance of a vertical rotary tiller based on the discrete element method. Actuators, 11(12): 1-16.
48. Zhang, Y., Wang, S., Wang, H., Ning, F., Zhang, Y., Dong, Z., Wen, P., Wang, R., Wang, X. and Li, J. (2018). The effects of rotating conservation tillage with conventional tillage on soil properties and grain yields in winter wheat-spring maize rotations. Agricultural and Forest Meteorology, 263: 107–117.