FEASIBILITY REPORT

Feasibility Report

Feasibility Report

Part A

Recently, many spherical domes have been constructed for sports facilities and so on. Thin sheet metal and/or membranes are often utilised for roof cladding because of their power and lightness. Being lightweightweightweight and flexible, such roofing materials are susceptible to dynamic breeze actions. Since wind stresses portraying on domes vary spatially as well as in time, the conceive wind burdens should be determined by considering the dynamic characteristics of breeze stresses in an befitting manner. Fatigue of cladding components may play an significant function in the wind-resistant presentation of cladding systems.

In Japan, top covering cladding is generally conceived founded on the most critical peak force coefficients irrespective of breeze direction. Such a accepted codification supplies a single top conceive pressure coefficient for each top covering zone considering a nominal worst-case scenario. Furthermore, the specifications are generally founded on the anticipated values of the top force coefficients. Neither the likelihood circulation of the peak force coefficients nor the peaks other than the largest one are considered. Therefore, they are not suitable for fatigue and risk-consistent designs. It is essential to develop a new methodology that considers the top pressure coefficients according to fixed risk levels and the stacking sequence for estimating the fatigue impairment of roof cladding and its fixings. Computer replication of breeze force time series may be useful for this purpose.

(Kumar and Stathopoulos, 1999) and (Kumar and Stathopoulos, 2001) proposed a innovative simulating procedure that develops both Gaussian and non-Gaussian breeze force fluctuations on reduced building roofs. The method is effectively applied to fatigue investigation as well as to the evaluation of farthest stresses in a risk-consistent way. Therefore, the procedure is used in the present study and a simplification of this method will be considered later. Gaussian and non-Gaussian force fluctuations can be simulated from the statistics of wind stresses, i.e. the mean, benchmark deviation, skewness, kurtosis and power spectrum. These statistical values change with position as well as with numerous components associated to the dome's geometry and the turbulence characteristics of approach flow. For such a complicated occurrence, in which numerous variables are engaged, artificial neural networks (ANNs) can be used effectively.

ANNs can capture a complex, non-linear relationship via training with informative input-output example data pairs obtained from computations and/or experiments. Among a kind of ANNs developed so far, the cascade association discovering mesh (CCLN) (Fahlman and Lebiere, 1990) is applied to the present problem. The results of comprehensive study conveyed by Hongo (1995) are used for training and testing the neural network after some manipulation. For this purpose, a database of the statistics of wind force coefficients is constructed.

The present study proposes a computer-assisted breeze load evaluation system for the conceive of top covering cladding of spherical domes. The composition of the scheme is schematically illustrated in Fig. 1. This system supplies breeze burdens for the conceive of roof cladding and its fixings without bearing out any breeze tunnel experiment. An aerodynamic database, an ANN and a time-series replication technique are engaged ...