Intelligent modelling of sandstone deformation behaviour using fuzzy logic and neural network systems
| Autor: | Ahmad Fahimifar, Behnam Yazdani Bejarbaneh, Danial Jahed Armaghani, Mohd For Mohd Amin, Elham Yazdani Bejarbaneh, Muhd Zaimi Abd Majid |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Artificial neural network
business.industry 0211 other engineering and technologies Modulus Geology Young's modulus 02 engineering and technology Structural engineering Geotechnical Engineering and Engineering Geology Fuzzy logic symbols.namesake Schmidt hammer Ranking symbols Deformation (engineering) Rock mass classification business 021106 design practice & management 021101 geological & geomatics engineering |
| Zdroj: | Bulletin of Engineering Geology and the Environment. 77:345-361 |
| ISSN: | 1435-9537 1435-9529 |
| DOI: | 10.1007/s10064-016-0983-2 |
| Popis: | A realistic analysis of rock deformation in response to any change in stresses is heavily dependent on the reliable determination of the rock properties as analysis inputs. Young’s modulus (E) provides great insight into the magnitude and characteristics of the rock mass/material deformation, but direct determination of Young’s modulus in the laboratory is time-consuming and costly. Therefore, basic rock properties such as point load strength index, P-wave velocity and Schmidt hammer rebound number have been used to estimate Young’s modulus. These rock properties can be easily measured in the laboratory. The main aim of this study was to develop two intelligent models based upon fuzzy logic and biological nervous systems in order to estimate Young’s modulus of sandstone for a set of known index properties drawn from laboratory tests. The database required to construct these models comprised a series of drill cores (96 samples of sandstone) from site investigation operations for a hydroelectric roller-compacted concrete (RCC) dam located in the Malaysian state of Sarawak. In the final stage of the present study, using the same data sets, multiple regression (MR) analysis was also proposed for comparison with the prediction results of both the fuzzy inference system (FIS) and artificial neural network (ANN) models. The ANN model was found to be far superior to FIS and MR in terms of several performance indices including root-mean-square error and ranking. Thus, from the results of this study, it was concluded that the models proposed herein could be utilised to estimate the E of similar rock types in practice. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink