Metal Eklemeli ve Geleneksel İmalat Yöntemleriyle Üretilen Alüminyum Profillerin Sıcaklık Davranışının Bulanık Mantık ve Yapay Zekâ Tabanlı Modellemesi

Yıl 2025, Cilt: 37 Sayı: 2, 863 - 875

Koray Özsoy , Bekir Aksoy

https://doi.org/10.35234/fumbd.1739022

Öz

Çalışmada, alüminyum ekstrüzyon profillerinin sıcaklık özellikleri üzerinde üretim teknolojilerinin etkilerinin incelenmesi amaçlanmıştır. Mevcut geleneksel imalat yöntemlerinin üretim kısıtlamalarını aşmak amacıyla, metal eklemeli imalat yöntemiyle üretilen kalıplar kullanılarak profil çıkış sıcaklıkları değerlendirilmiştir. Eklemeli imalat teknolojisinin sunduğu tasarım esnekliği sayesinde, alüminyum profillerin termal analizleri gerçekleştirilmiş ve böylece üretim teknolojisinin süreç performansı üzerindeki etkileri ortaya konmuştur. Çalışma kapsamında, alüminyum ekstrüzyon kalıpları hem geleneksel imalat yöntemiyle hem de EOS© tarafından sağlanan maraging çeliği (MS1) metal tozu kullanılarak Doğrudan Metal Lazer Sinterleme (DMLS) yöntemiyle üretilmiştir. Alüminyum profillerin termal davranışlarını belirlemek amacıyla saha testleri gerçekleştirilmiş ve profil çıkış sıcaklıkları ölçülmüştür. Elde edilen bu deneysel verilere dayanarak, hız ve üretim teknolojileri giriş değişkenleri ile çıkış değişkeni olarak sıcaklık değerleri kullanılarak bulanık mantık ve yapay zeka tabanlı bir modelleme gerçekleştirilmiştir. Bulanık mantık tabanlı model, hız ve üretim teknolojisi parametrelerine bağlı sıcaklık tahminlerinde deneysel verilerle %95’in üzerinde gerçekleşmiştir. Çalışmanın son aşamasında, bu çalışmaya özgü hazırlanan veri seti yapay sinir ağları modeli ile eğitilmiş ve elde edilen sonuçlara göre R² performans değerlendirme metriğinde %93 doğruluk oranına ulaşılmıştır. Sonuçlar, metal eklemeli imalat yöntemiyle üretilen kalıplar kullanılarak elde edilen alüminyum profillerin çıkış sıcaklıklarının, geleneksel imalatla üretilen profillere kıyasla daha düşük olduğunu ortaya koymuştur. Bu durum, özellikle maraging çeliğinden üretilen kalıp bileşenlerinin termal yük altında daha uzun ömürlü olabileceğini ve kalıp verimliliğinin önemli ölçüde artırılabileceğini göstermektedir.

Anahtar Kelimeler

Eklemeli imalat , metal eklemeli imalat , geleneksel imalat , ekstrüzyon , bulanık mantık , yapay zeka , profil sıcaklığı

Teşekkür

Yazarlar saha destekleri verdikleri için Eksen Aluminium Extrusion Technologies Co. Ltd Araştırma Arfem Aluminium Extrusion Ltd için teşekkür eder.

Modeling the Temperature Behavior of Aluminum Profiles Produced by Metal Additive and Conventional Manufacturing Methods Using Fuzzy Logic and Artificial Intelligence

Öz

The aim of this study is to investigate the effects of production technologies on the thermal properties of aluminum extrusion profiles. To overcome the limitations of conventional manufacturing methods, profile exit temperatures were evaluated using molds produced via metal additive manufacturing. Leveraging the design flexibility offered by additive manufacturing technology, thermal analyses of aluminum profiles were conducted, thereby revealing the influence of production technology on process performance. Within the scope of the study, aluminum extrusion molds were manufactured using both conventional methods and the Direct Metal Laser Sintering (DMLS) process, employing maraging steel (MS1) metal powder provided by EOS©. To determine the thermal behavior of the aluminum profiles, field tests were conducted, and profile exit temperatures were measured. Based on the experimental data obtained, a fuzzy logic and artificial intelligence-based model was developed, with speed and production technology as input variables and temperature as the output variable. The fuzzy logic-based model achieved more than 95% agreement with the experimental data in predicting temperatures depending on production speed and manufacturing technology parameters. In the final stage of the study, a dataset specifically prepared for this research was used to train an artificial neural network model, and the results indicated that the R² performance evaluation metric reached an accuracy level of 93%. The results indicate that the exit temperatures of aluminum profiles produced using molds manufactured via metal additive manufacturing are lower than those produced using conventional methods. This suggests that mold components—particularly those made of maraging steel—can exhibit longer service life under thermal loads, leading to a significant improvement in mold efficiency.

Anahtar Kelimeler

Additive manufacturing , metal additive manufacturing , conventional manufacturing , extrusion , fuzzy logic , artificial intelligence , profile temperature

Kaynakça

• Aslantas K, Hasçelik A, Erçetin A, Danish M, Alatrushi LK, Rubaiee S, Mahfouz AB. Effect of cutting conditions on tool wear and wear mechanism in micro-milling of additively manufactured titanium alloy. Tribol Int 2024;193:109340.
• Özsoy K. Examining mechanical properties of profiles manufactured aluminium extrusion dies using powder bed fusion. Meas 2021;177, 109266.
• Li D, He J, Tian X, Liu Y, Zhang A, Lian Q, Lu B. Additive manufacturing: integrated fabrication of macro/microstructures. Chin J Mech Eng 2013; 49(6): 129-135.
• Lu B. Additive manufacturing—Current situation and future. Chin Mech Eng 2020; 31(01): 19.
• Katz-Demyanetz A, Popov Jr VV, Kovalevsky A, Safranchik D, Koptioug A. Powder-bed additive manufacturing for aerospace application: Techniques, metallic and metal/ceramic composite materials and trends. Manuf Rev 2019; 6.
• Xiao Y, Yang Y, Wang D, Zhou H, Liu Z, Liu L, Song C. In-situ synthesis of spatial heterostructure Ti composites by laser powder bed fusion to overcome the strength and plasticity trade-off. Int J Mach Tools and Manuf, 2024; 196: 104117.
• Yang Y, Jiang R, Han C, Chen J, Li H, Wang Y, Wang D. Frontiers in laser additive manufacturing technology. Addit. Manuf Front 2024; 200160.
• Dong Z, Han C, Zhao Y, Huang J, Ling C, Hu G, Yang Y. Role of heterogenous microstructure and deformation behavior in achieving superior strength-ductility synergy in zinc fabricated via laser powder bed fusion. Int J Extreme Manuf 2024; 6(4): 045003.
• Han C, Fang Q, Shi Y, Tor SB, Chua CK, Zhou K. Recent advances on high‐entropy alloys for 3D printing. Adv Mater 2020; 32(26): 1903855.
• Sarzyński B, Śnieżek L, Grzelak K. Metal additive manufacturing (MAM) applications in production of vehicle parts and components—a review. Metals 2024; 14(2): 195.
• ISO/ASTM 52900 .Additive manufacturing—general principles—fundamentals and vocabulary. ISO, Geneva, 2021.
• Ansell TY. Current status of liquid metal printing. J Manuf Mater Process 2021; 5(2): 31.
• Qian M. Metal powder for additive manufacturing. Jom 2015; 67(3): 536-537.
• Slotwinski JA, Garboczi EJ, Stutzman PE, Ferraris CF, Watson SS, Peltz MA. Characterization of metal powders used for additive manufacturing. J Res Natl Inst Stand Technol 2014; 119:460.
• Isaza JF, Aumund-Kopp C. Additive Manufacturing with metal powders: Design for Manufacture evolves into Design for Function. Powd Meta Rev 2014; 3(2): 41-50.
• Jha AK, Sreekumar K, Tharian T, Sinha PP. Process optimization for high fracture toughness of maraging steel rings formed by mandrel forging. J Manuf Proce 2010; 12(1): 38-44.
• Samei J, Asgari H, Pelligra C, Sanjari M, Salavati S, Shahriari A, Mohammadi MA hybrid additively manufactured martensitic-maraging stainless steel with superior strength and corrosion resistance for plastic injection molding dies. Addit Manuf 2021; 45: 102068.
• Oter ZC, Coskun M, Akca Y, Sürmen Ö, Yılmaz MS, Özer G, Koc E. Benefits of laser beam based additive manufacturing in die production. Optic 2019; 176: 175-184.
• Sheppard T. Extrusion of Aluminium Alloys. 1nd ed. New York, NY, USA: Springer Science, 2013.
• Hölker R, Haase M, Khalifa NB, Tekkaya AE. Hot extrusion dies with conformal cooling channels produced by additive manufacturing. Mater Today Proc 2015; 2(10): 4838-4846.
• Chung CY, Liu CJ. Metal additive manufacturing of conformal cooling channels for improving the quality of metal injection–molded products. Adv Mech Eng 2024; 16(10).
• Marqués A, Dieste JA, Monzón I, Laguía A, Gracia P, Javierre C, Elduque D. Improvements in Injection Moulds Cooling and Manufacturing Efficiency Achieved by Wire Arc Additive Manufacturing Using Conformal Cooling Concept. Poly 2024; 16(21): 3057.
• Butdee SA. Prediction approach for aluminum extrusion processing using neuro-fuzzy based decision making. In International Scientific-Technical Conference Manufacturıng. Switzerland AG: Springer International Publishing, 2022. pp. 237-249.
• Zadeh LA. Fuzzy sets as a basis for a theory of possibility.” Fuzzy sets and systems 1999; 100: 9-34.
• Zadeh LA. Is there a need for fuzzy logic?. Infor Sci 2008; 178(13): 2751-2779.
• Al-Saadi T, Rossiter JA, Panoutsos G. Control of selective laser melting processes: existing efforts, challenges, and future opportunities. In 2021 29th Mediterranean Conference on Control and Automation (MED), IEEE. pp.89-94.
• Hua Y, Choi J. Adaptive direct metal/material deposition process using a fuzzy logic-based controller. J Las Appl 2005; 17(4): 200-210.
• Farshidianfar M.H, Khajepour A, Zeinali M, Gelrich A. System identification and height control of laser cladding using adaptive neuro-fuzzy inference systems. 32nd International Congress on Laser Materials Processing, Laser Microprocessing and Nanomanufacturing, 6–10 October 2013; Miami, Florida, USA: AIP Publishing. pp. 615-623.
• Li Y, Li X, Zhang G, Horv´ath I, Han Q. Interlayer closed-loop control of forming geometries for wire and arc additive manufacturing based on fuzzylogic inference. J Manuf Proc 2020; 63: 35–47.
• Lamarche-Gagnon MÉ, Molavi-Zarandi M, Raymond V, Ilinca F. Additively manufactured conformal cooling channels through topology optimization. Struc Mult Opt 2024; 67(8): 138.
• Çalışkan Cİ, Özer G, Koç E, Sarıtaş US, Yıldız CF, Çiçek ÖY. Efficiency research of conformal channel geometries produced by additive manufacturing in plastic injection mold cores (inserts) used in automotive industry. 3D Prin Addi Manuf 2023; 10(2): 213-225.
• Sagias VD, Zacharia P, Tempeloudis A, Stergiou C. Adaptive neuro-Fuzzy inference system-based predictive modeling of mechanical properties in additive manufacturing. Mach 2024; 12(8): 523.
• Toprak CB, Dogruer CU. Neuro-fuzzy modelling methods for relative density prediction of stainless steel 316L metal parts produced by additive manufacturing technique. J Mech Sci Tech 2023; 37(1): 107-118.
• Kodaloğlu M, Akarslan KF, Evaluation of Thermal Comfort In Terms of Occupational Safety In Weaving Facilities By Fuzzy Logic”, Int J 3D Prin Tech Dig Ind 2022; 6 (2): 273-279.
• McCulloch WS, Pitts W. A logical calculus of the ideas immanent in nervous activity. Bul Math Bio 1943; 5(4): 115-133.
• Rosenblatt F. The perceptron: A probabilistic model for information storage and organization in the brain. Psych Rev 1958; 65(6): 386.
• Rumelhart DE, Hinton, G. E., & Williams, R. J. Learning representations by back-propagating errors. Nat 1986; 323(6088): 533-536.
• Werbos PJ. Beyond regression: New tools for prediction and analysis in the behavioral sciences, Ph. D., Harvard University. USA, 1974.
• Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Adv Neu Infor Proce Sys 2012, pp. 1097-1105.
• Mikolov T, Chen K, Corrado G, Dean J. Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781. 2013.
• Hinton G, Deng L, Yu D, Dahl G, Mohamed A, Jaitly N, Kingsbury B. Deep neural networks for acoustic modeling in speech recognition. IEEE Sig Proce Mag 2012; 29(6): 82-97.
• Kohara K, Ishikawa T, Yoshikawa Y, Morita Y. Financial forecasting with neural networks. Neu Comp Appl 1997; 5(4): 213-220.
• Chai T, Draxler RR. Root mean square error (RMSE) or mean absolute error (MAE)?–Arguments against avoiding RMSE in the literature. Geosc Mod Devel 2014; 7(3): 1247-1250.
• Chicco D, Warrens MJ, Jurman G. The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. Peerj Comp Sci 2021; 7: e623.
• Özer D, Aksoy B. Yapay Zeka Uygulaması ile Güneş Paneli Sistemi Enerji Üretimi Tahmini. Yalvaç Aka Der 2024; 9(2): 138-151.
• Hodson TO. Root-mean-square error (RMSE) or mean absolute error (MAE): when to use them or not, Geosci. Mod Dev 2022; 15: 5481–5487.