Measurement and prediction of correction factors for very high core compressive strength by using the adaptive neuro-fuzzy techniques
| Autor: | Ramazan Demirboğa, Waleed H. Khushefati, Osman Taylan, Traad Mohammed Al-zharani |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Engineering
Adaptive neuro fuzzy inference system Neuro-fuzzy business.industry Gaussian 0211 other engineering and technologies 02 engineering and technology Building and Construction Structural engineering Backpropagation Core (optical fiber) symbols.namesake Compressive strength Buckling 021105 building & construction 0202 electrical engineering electronic engineering information engineering symbols 020201 artificial intelligence & image processing General Materials Science business Civil and Structural Engineering High strength concrete |
| Zdroj: | Construction and Building Materials. 122:320-331 |
| ISSN: | 0950-0618 |
| Popis: | In this study the relationship between the core compressive strength with respect to reference samples and different core sizes with different slenderness ratio, length to diameter (L/D) were investigated. In addition, a comparative study was carried out by a hybrid neuro-fuzzy (ANFIS) technique, the core correction factors were evaluated by statistical methods for comparing the performance of four different ANFIS approaches. The Gaussian membership functions were used for defining linguistic terms. The back propagation multi layer (BPML) and hybrid learning algorithms with grid partition were employed for the development of the ANFIS models. Experimental results showed that the core strength was increased with the decrease of slenderness ratio and have ranged between 0.95 and 1.21. The ANFIS model results showed that it could be used an efficient tool for the estimation of the core correction factor of very high strength concrete. The ANFIS model in the current study performs sufficiently in the estimation of core correction factor of high strength concrete, which particularly estimates closely following the actual values. The ANFIS model with hybrid learning algorithm of grid partition was able to produce the most accurate model outcomes for estimating the correction factor among the examined models. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect