Neural network approach for grading of maintainability of wet areas in high-rise buildings
Authors
Citation
International journal on architectural science, v.4, no.4, 2003, pp. 150-167
Abstract
A maintainability grading system using artificial neural networks which aids in enhancing decision-making of wet area design is derived in this paper. The model was derived from comprehensive condition surveys of 450 tall buildings and in-depth assessment of a further 120 tall buildings and face-to-face interviews with the relevant building professionals. 16 important risk factors were identified and tested according to their sensitivity in affecting maintainability scoring for wet areas. The system allows complete evaluation of various alternative designs, materials, construction and maintenance practices, so as to achieve best possible solutions of technical attributes that lead to minimum life cycle maintenance cost.
(1) Introduction
(2) Research methodology
(a) Data collection
(b) Modeling maintainability using neural network
(c) Construction of risk conditions - model input
(d) Computation of the level of maintainability - model output
(e) Network design and training
(3) Results and discussion
(a) Network Performance
(b) Forecasting maintainability
(4) Conclusions
(5) References
Description
Subject
Type
Article
Format
Date
2003
Language
en