Predicting the Compressive Strength of Reactive Powder Concretes by ANFIS

Creative Commons License

TOPÇU İ. B., GÜLBANDILAR E., Koca A. B.

JOURNAL OF POLYTECHNIC-POLITEKNIK DERGISI, cilt.21, sa.1, ss.165-171, 2018 (ESCI) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 21 Sayı: 1
  • Basım Tarihi: 2018
  • Doi Numarası: 10.2339/politeknik.385923
  • Dergi Adı: JOURNAL OF POLYTECHNIC-POLITEKNIK DERGISI
  • Derginin Tarandığı İndeksler: Emerging Sources Citation Index (ESCI), TR DİZİN (ULAKBİM)
  • Sayfa Sayıları: ss.165-171
  • Eskişehir Osmangazi Üniversitesi Adresli: Evet

Özet

This study is designed to investigate the experimental results of Reactive Powder Concretes (RPC) with the adaptive neural fuzzy inference systems (ANFIS) prediction model. In order to construct this model, the compressive strengths of 42 samples on the 7th and 28th days were obtained from the experiments for the training stages of the ANFIS model. This data was used on ANFIS model as hydration day, 8 input parameters including Portland cement, silica smoke, quartz sand, sand, water, super plasticizer and steel fiber and compressive strength of concrete as output parameter. The ANFIS model presented training performance with 0.015 error. When ANFIS test results compared with experimental results, it found that R-2, RMS and MAPE were statistically accurate to 0.9909, 0.027 and 0.0004, respectively. The test results show that ANFIS model is a convenient to use and simple model for estimating the compressive strengths of 7th and 28th days of RPC