ANALYSIS OF ABOVEGROUND PIPELINES ON UNILATERAL SUPPORTS - A NEURAL-NETWORK APPROACH

A neural network approach is proposed for the effective numerical trea tment of the analysis problem of above-ground pipelines on unilateral supports. As is well known, the structural response of such pipelines can be effectively described by monotone problems having the form of q uadratic minimization problems with inequality constraints. In order t o numerically treat these quadratic problems, the methods of the theor y of neural networks are applied. Within such a neural network computi ng framework, the Hopfield model is first appropriately generalized, a nd then a neural network model covering the case of inequalities is pr oposed. A numerical application illustrates the proposed neural networ k method. (C) 1998 Elsevier Science Ltd. All rights reserved.