Research On A System For Selecting The Types Of Foundations Based On Artificial Neural Network

Determining system of foundations is an important item in preliminary design of foundations. The selection of the foundational system is affected by many factors. The dissertation emplopys the method of Artificial .Neural Network, takes the designing for foundation in structure engineering as research objects and stresses on the research into the selection of the foundational problem in structure based on BP Neural Network.In this paper, the main principle of BP Neural Network is decribed briefly. Through analyzing the affecting elements of the selection of the foundational in structure and the characteristics of BP Neural Network, A identification model of Neural Network about the selection of the foundational in structure is established based on the Neural Networks Toolbox of MATLAB6.1. Taking advantage of nonlinearity and the ability of learning and memory of Neural Network , this model is fixed by training structure of network with training samples. Based on the above work, an system based on neural network toolbox of MATLAB 6.1 is exploited. The system can help designers to select an appropriate system of foundations. The demonstration shows that this method has an important practical significance and a referential worth in the design of foundations.