ANFIS modelling for blast fragmentation and blast-induced vibrations considering stiffness ratio


Akyıldız Ö., Hüdaverdi T.

ARABIAN JOURNAL OF GEOSCIENCES, cilt.13, sa.21, 2020 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 13 Sayı: 21
  • Basım Tarihi: 2020
  • Doi Numarası: 10.1007/s12517-020-06189-7
  • Dergi Adı: ARABIAN JOURNAL OF GEOSCIENCES
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Agricultural & Environmental Science Database, Aquatic Science & Fisheries Abstracts (ASFA), Geobase, INSPEC
  • Anahtar Kelimeler: Stiffness ratio, Blast fragmentation, Ground vibration, ANFIS, Error metrics, PREDICTION, ASYMMETRY, SIZE
  • İstanbul Teknik Üniversitesi Adresli: Evet

Özet

Two independent ANFIS models were created to predict blast fragmentation and blast-induced ground vibrations. The site investigation was performed in a sandstone quarry in Istanbul Region. The first model predicts peak particle velocity. The input parameters of the ANFIS vibration model are the stiffness ratio and scaled distance. The ANFIS model was compared to classical scaled distance-based predictor equations and a multiple regression equation. Twelve predictor equations were considered to reveal actual capability of the ANFIS vibration model. The second ANFIS model predicts mean fragment size of blast muckpile. The input parameters of the ANFIS fragmentation model are the stiffness ratio and powder factor. The ANFIS fragmentation model was compared to a regression equation and the well-known Kuznetsov equation. This research specially focuses on an important but neglected design parameter, the stiffness ratio. It is believed that stiffness ratio significantly affects both ground vibration and fragmentation. The study was also aimed to perform a comprehensive and detailed model validation. Eleven error measures were used to determine prediction capabilities of the models. Performance of the error metrics was also discussed. The developed ANFIS models show quite promise. The ANFIS models have only two input parameters. Robust and noncomplex models were created.