Research And Development Of CAD System For Space-grid Structure

Determining system of retaining and protecting for deep foundation excavation is an important item in preliminary design. The development of theory and practice promotes great development of retaining and protecting structures. However, the lectotype decision is still mostly using conventional manual method that is lower in science and intellect level. The dissertation employs the method of Back-propagation Neural Network, takes five popular retaining and protecting structure types as research objects and stresses on research into the lectotype decision for retaining and protecting structures. The dissertation starts with the theory of retaining and protecting for deep foundation excavation and abundant case studies of retaining and protecting engineering, profoundly analyses legion influencing factors of retaining and protecting lectotype and abstracts index system. Aim to the characteristics that lectotype involves many fuzziness and non-determinacy influencing factors, based on Back-propagation neural network theory and MATLAB language, designs the lectotype decision model of retaining and protecting structures for deep foundation excavation. Using the sample knowledge base comes from abundant case studies that are handled by series of quantifying method, the lectotype decision model is trained and debugged and comes into being a lectotype decision model of retaining and protecting structures. An engineering practice of this lectotype decision model is introduced. From the aspects of the design of humankind experts, of fuzzy comprehensive evaluation method and of the monitoring data, the feasibility and authenticity of this lectotype decision model are examined. It can be expediently used and proved to be much effective.