Modelling of Friction Data Obtained from Discrete Element Simulation of Peanut Kernel Grading Machine Using Nonlinear Auto-Regressive Network
Yıl 2025,
Cilt: 40 Sayı: 1, 72 - 92, 30.06.2025
Cem Korkmaz
,
İlyas Kacar
Öz
Time series prediction models were used to model the variation of friction in peanut grading sieves. Discrete element method (DEM)-based simulation was used to analyse the friction. Nonlinear autoregressive neural networks (NARX) method was used for modelling and prediction. The NARX network was trained with the Levenberg-Marquardt algorithm. The number of hidden layers is 2 and the number of neurons in the hidden layers is 100 and 50 respectively. In the obtained peanut classification model, the correlation value was 0.97449 and the root mean square error was 3.584527 N. The percentage of the mean absolute error was found to be 3.304729%. 10 iterations were performed and the calculations took 129.7819 seconds. There is a high linear relationship between the model outputs and the data set. The NARX network eliminated the need for any preprocessing of the data. This study aims to improve peanut classification processes using particle simulations and neural networks.
Kaynakça
- Abd-Elhamid, H. F., El-Dakak, A. M., Zeleňáková, M., Saleh, O., Mahdy, M. ve Ghany, S. H. A. E. (2024). Rainfall forecasting in arid regions in response to climate change using ARIMA and remote sensing. Geomatics, Natural Hazards and Risk, 15(1): 2347414.
- Adamczyk, W., Widuch, A., Morkisz, P., Zhou, M., Myöhänen, K., Klimanek, A., et al. (2024). A machine learning-based simplified collision model for granular flows. Powder Technology: 120006.
- Akcali, İ. D., Mutlu, H. ve Ercan, U. (2014). Mathematical model of a sorting machine. Tarım Makinaları Bilimi Dergisi, 10(3): 229-234.
- Akoglu, H. (2018). User's guide to correlation coefficients. Turkish journal of emergency medicine, 18(3): 91-93.
- Alsharif, M. H., Younes, M. K. ve Kim, J. (2019). Time series ARIMA model for prediction of daily and monthly average global solar radiation: The case study of Seoul, South Korea. Symmetry, 11(2): 240.
- Ansys Rocky DEM©. (2021). The most powerful particle simulation software. (Version 2024 R2). Germany: Ansys©.
- Atsyo, S. (2022). Classification of peanut seeds based on different feed rates using a grading machine. (Master Thesis), Cukurova University, Institute of Science and Technology. (706206).
- Cai, J. ve Yang, H. (2024). Attention mechanism-aided model ensemble method of chiller energy consumption prediction. International Journal of Refrigeration.
- Chu, H., Wang, Z. ve Nie, C. (2024). Monthly streamflow prediction of the source region of the yellow river based on long short-term memory considering different lagged months. Water, 16(4): 593.
- Cornejo, A., Franci, A., Zárate, F. ve Oñate, E. (2021). A fully Lagrangian formulation for fluid-structure interaction problems with free-surface flows and fracturing solids. Computers & Structures, 250: 106532. doi: https://doi.org/10.1016/j.compstruc.2021.106532.
- Cristianini, N. ve Shawe-Taylor, J. (2000). An introduction to support vector machines and other kernel-based learning methods. Cambridge: Cambridge University Press.
- Daghistani, F. ve Abuel-Naga, H. (2024). Advancements in understanding interface friction: A combined experimental and machine learning approach using multiple linear and random forest regressions. Geotechnics, 4(1): 109-126.
- Davydzenka, T. ve Tahmasebi, P. (2022). High-resolution fluid–particle interactions: a machine learning approach. Journal of Fluid Mechanics, 938: A20.
- de Sousa Junior, W. T., Montevechi, J. A. B., de Carvalho Miranda, R. ve Campos, A. T. (2019). Discrete simulation-based optimization methods for industrial engineering problems: A systematic literature review. Computers & Industrial Engineering, 128: 526-540.
- Donea, J., Huerta, A., Ponthot, J.-P. ve Rodríguez‐Ferran, A. (2004). Chapter 14: Arbitrary Lagrangian–Eulerian methods The Encyclopedia of Computational Mechanics (Vol. 1, pp. 413-437): Wiley.
- Doroszuk, B., Król, R. ve Wajs, J. (2021). Simple design solution for harsh operating conditions: redesign of conveyor transfer station with reverse engineering and DEM simulations. Energies, 14(13): 4008.
İç Yer Fıstığı Sınıflama Makinesinin Ayrık Elemanlar Simülasyonundan Elde Edilen Sürtünme Verisinin Doğrusal Olmayan Oto-Gerilemeli Ağ ile Modellenmesi
Yıl 2025,
Cilt: 40 Sayı: 1, 72 - 92, 30.06.2025
Cem Korkmaz
,
İlyas Kacar
Öz
Fıstık sınıflandırma eleklerindeki sürtünmenin zamanla değişimini modellemek için zaman serisi kestirim modelleri kullanılmıştır. Sürtünmenin analizi için ayrık elemanlar yöntemi (DEM) esaslı simülasyon yapılmıştır. Modellenmesi ve kestirimi için doğrusal olmayan otoregresif sinir ağları (NARX) yöntemi kullanılmıştır. NARX ağı Levenberg-Marquardt algoritması ile eğitilmiştir. Saklı katman sayısı 2 olup saklı katmanlardaki nöron sayısı sırası ile 100, 50 olmaktadır. Elde edilen fıstık sınıflandırma modelinde korelasyon değeri 0.97449 ve hata karesinin ortalamasının kökü 3.584527 N olarak bulunmuştur. Ortalama mutlak hatanın yüzdesi ise % 3.304729 olarak bulunmuştur. 13 adet iterasyon yapılmış ve hesaplamalar 129.7819 saniye sürmüştür. Model çıktıları ile veri seti arasında yüksek doğrusal bir ilişki mevcuttur. NARX ağı, veriye herhangi bir ön işlem yapılması gereksinimini ortadan kaldırmıştır. Bu çalışma, parçacık simülasyonları ve sinir ağları kullanarak yer fıstığı sınıflandırma süreçlerinin iyileştirilmesini amaçlamaktadır.
Etik Beyan
Veri ve Kod Kullanılabilirliği
Çalışmada kullanılan veriler şekil veya tablolarda zaten verilmiştir.
Yazar Katkıları
Tüm yazarlar eşit katkıda bulunmuştur.
Çıkar çatışmaları / Rekabet eden çıkarlar
Bu çalışma, kamu, ticari veya kâr amacı gütmeyen sektörlerdeki fon kuruluşlarından herhangi bir özel hibe almamıştır.
Destekleyen Kurum
Bu çalışmayı desteleyen ticari bir kurum yoktur.
Teşekkür
Ayrık eleman yöntemi simülasyonları için kullanılan RockyDEM©’in, makine öğrenme yönteminin kodlanmasında yararlanılan Matlab©'ın eğitsel amaçlı kullanım imkânını sağlayan sırasıyla, Çukurova Üniversitesi Ziraat Fakültesi'ne, Karadeniz Teknik Üniversitesi’ne, Dr. Mehmet Seyhan’a teşekkür ederim. Bu çalışmanın inceleme ve değerlendirme aşamasında yapmış oldukları değerli katkılardan dolayı editör, hakem ve emeği geçenlere içten teşekkür ederim.
Kaynakça
- Abd-Elhamid, H. F., El-Dakak, A. M., Zeleňáková, M., Saleh, O., Mahdy, M. ve Ghany, S. H. A. E. (2024). Rainfall forecasting in arid regions in response to climate change using ARIMA and remote sensing. Geomatics, Natural Hazards and Risk, 15(1): 2347414.
- Adamczyk, W., Widuch, A., Morkisz, P., Zhou, M., Myöhänen, K., Klimanek, A., et al. (2024). A machine learning-based simplified collision model for granular flows. Powder Technology: 120006.
- Akcali, İ. D., Mutlu, H. ve Ercan, U. (2014). Mathematical model of a sorting machine. Tarım Makinaları Bilimi Dergisi, 10(3): 229-234.
- Akoglu, H. (2018). User's guide to correlation coefficients. Turkish journal of emergency medicine, 18(3): 91-93.
- Alsharif, M. H., Younes, M. K. ve Kim, J. (2019). Time series ARIMA model for prediction of daily and monthly average global solar radiation: The case study of Seoul, South Korea. Symmetry, 11(2): 240.
- Ansys Rocky DEM©. (2021). The most powerful particle simulation software. (Version 2024 R2). Germany: Ansys©.
- Atsyo, S. (2022). Classification of peanut seeds based on different feed rates using a grading machine. (Master Thesis), Cukurova University, Institute of Science and Technology. (706206).
- Cai, J. ve Yang, H. (2024). Attention mechanism-aided model ensemble method of chiller energy consumption prediction. International Journal of Refrigeration.
- Chu, H., Wang, Z. ve Nie, C. (2024). Monthly streamflow prediction of the source region of the yellow river based on long short-term memory considering different lagged months. Water, 16(4): 593.
- Cornejo, A., Franci, A., Zárate, F. ve Oñate, E. (2021). A fully Lagrangian formulation for fluid-structure interaction problems with free-surface flows and fracturing solids. Computers & Structures, 250: 106532. doi: https://doi.org/10.1016/j.compstruc.2021.106532.
- Cristianini, N. ve Shawe-Taylor, J. (2000). An introduction to support vector machines and other kernel-based learning methods. Cambridge: Cambridge University Press.
- Daghistani, F. ve Abuel-Naga, H. (2024). Advancements in understanding interface friction: A combined experimental and machine learning approach using multiple linear and random forest regressions. Geotechnics, 4(1): 109-126.
- Davydzenka, T. ve Tahmasebi, P. (2022). High-resolution fluid–particle interactions: a machine learning approach. Journal of Fluid Mechanics, 938: A20.
- de Sousa Junior, W. T., Montevechi, J. A. B., de Carvalho Miranda, R. ve Campos, A. T. (2019). Discrete simulation-based optimization methods for industrial engineering problems: A systematic literature review. Computers & Industrial Engineering, 128: 526-540.
- Donea, J., Huerta, A., Ponthot, J.-P. ve Rodríguez‐Ferran, A. (2004). Chapter 14: Arbitrary Lagrangian–Eulerian methods The Encyclopedia of Computational Mechanics (Vol. 1, pp. 413-437): Wiley.
- Doroszuk, B., Król, R. ve Wajs, J. (2021). Simple design solution for harsh operating conditions: redesign of conveyor transfer station with reverse engineering and DEM simulations. Energies, 14(13): 4008.