Yapay Sinir Ağ Modeli Kullanılarak Ön Germeli Beton Demeti Maksimum Çekme Mukavemetinin Tahmini
Demir ve çelik endüstrisi, bir ülkenin endüstriyel ve ekonomik kalkınması için vazgeçilmez sektörlerden biridir. Demir ve çelik endüstrisindeki en yaygın sorun, ürünün maksimum çekme mukavemetini belirlemektir. Ön germeli beton demeti (ÖGBD) ürününde kullanılan hammaddeler kuvvet altında deforme olmakta ve karakteristikleri sabit olmadığından şekilleri ve boyutları değişmektedir. Ürünün, akma ve maksimum çekme mukavemeti gibi malzeme özelliklerini anlamak için bazı mekanik testler gerçekleştirilir. Bu mekanik testlerde ortaya çıkan ürün, zaman ve iş gücü kaybı, tahribatsız ölçümlere dayanan bir tahmin metodu geliştirme ihtiyacını ortaya koymaktadır. Bu çalışmada, ön germeli beton demeti ürününün mekanik özellikleri yapay sinir ağları (YSA) kullanılarak tahmin edilmiştir. Mevcut işlem için en doğru ağ tipi olduğundan 'İleri Beslemeli Geri Yayılım (İBGY)' tercih edilmiştir. Maksimum çekme mukavemetini belirlemek için, malzeme üzerine uygulanılan yük (yük hücresi çıkışı), indüksiyon fırınının DC gerilimi ve DC akımı, ÖGBD hattının hızı, indüksiyon fırınının sıcaklığı, soğutma tankının sıcaklığı ve ÖGBD ürününün çapı gibi veriler gerçek bir üretim hattından toplanmakta ve simülasyon ortamında YSA’nın girdi parametreleri olarak kullanılmaktadır. Çalışma, ANN modelinin, ön gerilmeli beton demetinin maksimum çekme mukavemetine dair çok iyi tahminde bulunulduğunu göstermektedir.
Prediction of Ultimate Tensile Strength of Prestressed Concrete Strand Using Artificial Neural Network Model
The iron and steel industry is one of the essential sector for the industrial and economic development of a country. The most common problem in iron and steel industry is to determine the ultimate tensile strength of the product. The raw materials that are used in the Prestressed Concrete (PC) strand product are deformed under force and their shape and size are changed since the characteristics of them are not constant. To understand the material properties of the product such as the yield and the ultimate tensile strength, some mechanical tests are carried out. The product, the time and the labor loss occured in these mechanical tests reveal the need to develop a prediction method based on non-destructive measurement. In this study, the mechanical properties of PC strand product is predicted by using artificial neural networks (ANN). 'Feed-Forward Backpropagation (FFBP)' has been preferred since it is the most accurate network type for the current process. To determine the ultimate tensile strength, the data such as the load applied to the material (loadcell output), the DC voltage and the DC current of the induction furnace, the speed of the PC strand line, the temperature of the induction furnace, the temperature of the quench tank and the diamater of the PC strand product are collected from a real production line and are utilized as the input parameters of the ANN in the simulation environment. The study illustrates that the ANN model give a very good prediction of the ultimate tensile strength of PC strand.
___
- 1. Malinov, S., Sha, W., 2004. Application of
Artificial Neural Networks for Modelling
Correlations in Titanium Alloys. Materials
Science and Engineering: A, 365(1-2),
202-211.
- 2. Fujii, H., Mackay, D.J.C., Bhadeshia,
H.K.D.H., 1996. Bayesian Neural Network
Analysis of Fatigue Crack Growth Rate in
Nickel Base Superalloys. ISIJ International,
36(11), 1373-1382.
- 3. Cool, T., Bhadeshia, H.K.D.H., Mackay,
D.J.C., 1997. The Yield and Ultimate Tensile
Strength of Steel Welds. Materials Science and
Engineering: A, 223(1-2), 186–200.
- 4. Al-Assaf, Y., El-Kadi, H., 2001. Fatigue Life
Prediction of Unidirectional Glass Fiber/epoxy
Composite Laminae Using Neural Networks.
Composite Structures, 53(1), 65-71.
- 5. Akbari, M.K., Shirvanimoghaddam, K., Hai,
Z., Zhuiykov, S., Khayyam, H., 2017. Nano
TiB2 and TiO2 Reinforced Composites: A
Comparative Investigation on Strengthening
Mechanisms and Predicting Mechanical
Properties Via Neural Network Modeling.
Ceramics International, 43, 16799-16810.
- 6. Malinov, S., Sha, W., Mckeown, J.J., 2001.
Modelling the Correlation Between Processing
Parameters and Properties in Titanium Alloys
Using Artificial Neural Network.
Computational Materials Science, 21(3),
375–394.
- 7. Mcbride, J., Malinov, S., Sha, W., 2004.
Modelling Tensile Properties of Gamma-Based
Titanium Aluminides Using Artificial Neural
Network. Materials Science and Engineering:
A, 384(1-2), 129–137.
- 8. Akbari, J., Rakhshan, N., Ahmadvand, M.,
2013. Evaluation of Ultimate Torsional
Strength of Reinforcement Concrete Beams
Using Finite Element Analysis and Artificial
Neural Network. IJE Transactions B:
Applications, 26(5), 501-508.
- 9. Jeyasehar, C.A., Sumangala, K., 2006.
Nondestructive Evaluation of Prestressed
Concrete Beams Using an Artificial Neural
Network (ANN) Approach. Structural Health
Monitoring, 5(4), 313-323.
- 10. Esfahani, M.B., Toroghinejad, M.R, Abbasi,
S., 2009. Artificial Neural Network Modeling
the Tensile Strength of Hot Strip Mill Products.
ISIJ International, 49(10), 1583-1587.
- 11. Lalam, S., Tiwari, P.K., Sahoo, S., Dalal, A.K.,
2017. Online Prediction and Monitoring of
Mechanical Properties of Industrial Galvanised
Steel Coils Using Neural Networks.
Ironmaking & Steelmaking.
- 12. Khayyam, H., Fakhrhoseini, S.M., Church,
J.S., Milani, A.S., Bab-Hadiashar, A., Jazar,
R.N., Naebe M., 2017. Predictive Modelling
and Optimization of Carbon Fiber Mechanical
Properties Through High Temperature Furnace,
Applied Thermal Engineering, 125, 1539-1554.
- 13. http://www.strongwillwire.com/
- 14. ASTM International, 2017. ASTM A416
A416M-17 Standard Specification for Low
Relaxation, Seven-wire Steel Strand for
Prestressed Concrete. West Conshohocken,
PA.
- 15. Saravanakumar, P., Jothimani, V., Sureshbabu,
L., Ayyappan, S., Noorullah, D.,
Venkatakrishnan, P.G., 2012. Prediction of
Mechanical Properties of Low Carbon Steel in
Hot Rolling Process Using Artificial Neural
Network Model. Procedia Engineering, 38,
3418-3425.
- 16. Denizer, B., 2008. Artificial Neural Network
Analysis of the Mechanical Properties of
Tungsten Fiber/Bulk Metallic Glass Matrix
Composites via Neutron Diffraction and Finite
Element Modeling., Iowa State University,
Ames, Iowa, 33.
- 17. Beale, M.H., Hagan, M.T., Demuth, H.B.,
2017. Neural Network Toolbox User's Guide
the MathWorks Inc., Natick, MA, 512.