The SVM-ARIMA Fusion Approach for Forecasting the Structural Integrity of Obsidian Substitution Mortars


Çakmak T., Yılmaz Y., Ustabaş İ.

2024 Innovations in Intelligent Systems and Applications Conference (ASYU), Ankara, Türkiye, 16 - 18 Ekim 2024, ss.1-7, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/asyu62119.2024.10757110
  • Basıldığı Şehir: Ankara
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.1-7
  • Recep Tayyip Erdoğan Üniversitesi Adresli: Evet

Özet

The construction materials used in building design have changed over time, and their properties continue to evolve. Currently, the most widely used construction material is concrete. Concrete possesses important characteristics such as high strength, durability, low cost, and easy accessibility. However, the preparation and consumption of cement, which is the primary basic material for concrete, contribute to CO2 emissions, accounting for 5-8% of total emissions. In recent years, researchers have been conducting various studies to reduce this ratio. One of the most important alternatives is the use of binders with pozzolanic properties alongside cement. In this study, mortar specimens were produced using different proportions of obsidian powder with pozzolanic properties, ranging from 0% to 30%, in conjunction with cement. The compressive strengths of mortar specimens were measured by subjecting them to mechanical tests on different days such as 3, 7, 14, 28, 56, and 90. However, one of the significant challenges concerns the direction and rate of change in compressive strengths of concrete at later ages. To overcome this challenge, the SVM-ARIMA fusion forecasting model is used in this paper for predicting short-term compressive strengths of mortar specimens. Using this model, compressive strength values for the next 30 days were predicted based on the 90-day compressive strength data. The analysis results yielded R2 values of 0.982, 0.974, 0.979, and 0.977, respectively. These results indicate that the use of statistical models is appropriate for predicting the compressive strengths of mortar specimens.