A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Concrete undamaged detection based on fuzzy neural networks
The accuracy of concrete strength inspection has a great influence on the safety evaluation of the building. In order to improve the accuracy, Fuzzy Neural Network (FNN) was built up to evaluate concrete strength. It takes advantage of the merits of the common concrete testing methods: rebounding and drilling core, and the abilities of FNN including self-learning, generation and fuzzy logic inference. Verification test shows that the max relative error of the predicted results is 1.12%, which meets the need of practical engineering. The approach effectively maps the complex non-linear relationship between rebounding value and concrete strength, and provides a efficient way for the concrete strength detection and evaluation. With the accumulation of training data, ANFIS (neuro-fuzzy inference system) can learn new data and adjust parameters to improve the accuracy of predict. Nonlinear relationship between rebounding value and concrete strength is different in different areas. It is necessary to built specific nonlinear relationship for given area. Then the concrete strength can be obtained only from rebounding value and carbonation depth, which is easy to be determined. The concrete undamaged detection system based on ANFIS has a great importance in structure safety assessment, and can also create significant economical and social benefit.
Concrete undamaged detection based on fuzzy neural networks
The accuracy of concrete strength inspection has a great influence on the safety evaluation of the building. In order to improve the accuracy, Fuzzy Neural Network (FNN) was built up to evaluate concrete strength. It takes advantage of the merits of the common concrete testing methods: rebounding and drilling core, and the abilities of FNN including self-learning, generation and fuzzy logic inference. Verification test shows that the max relative error of the predicted results is 1.12%, which meets the need of practical engineering. The approach effectively maps the complex non-linear relationship between rebounding value and concrete strength, and provides a efficient way for the concrete strength detection and evaluation. With the accumulation of training data, ANFIS (neuro-fuzzy inference system) can learn new data and adjust parameters to improve the accuracy of predict. Nonlinear relationship between rebounding value and concrete strength is different in different areas. It is necessary to built specific nonlinear relationship for given area. Then the concrete strength can be obtained only from rebounding value and carbonation depth, which is easy to be determined. The concrete undamaged detection system based on ANFIS has a great importance in structure safety assessment, and can also create significant economical and social benefit.
Concrete undamaged detection based on fuzzy neural networks
Zerstörungsfreie Prüfung von Beton auf der Basis von neuronalen Fuzzy-Netzen
Yang, Songsen (author) / Zhao, Tiejun (author) / Rong, Jian (author) / Xu, Jing (author)
2004
4 Seiten, 6 Bilder, 2 Tabellen, 7 Quellen
Conference paper
English
Non-destructive evaluation of concrete in damaged and undamaged states
| British Library Online Contents | 2009
Non-destructive evaluation of concrete in damaged and undamaged states
| Elsevier | 2009
Non-destructive evaluation of concrete in damaged and undamaged states
| Elsevier | 2009
Anisotropic Viscoelastic Properties of Undamaged Asphalt Mixtures
| Online Contents | 2012
Behavior of damaged and undamaged concrete strengthened by carbon fiber composite sheets
| British Library Online Contents | 2002