The Discrete Variable Optimization Of Tower Parking Structure And CAD/CAE Software Development

With the increase of the number of city vehicles, parking industry ushered in the great business opportunities and broad market, the traditional design method of three-dimensional garage is slow, which with high redundancy structure. In order to improve the three-dimensional garage design efficiency, reduce the cost of the structure of the stereo garage, the parametric modeling analysis of stereo garage which is built by finite element analysis software ANSYS in this paper, combined with MATLAB neural network prediction function to predict the most dangerous working condition of the garage and the most dangerous section, using the discrete variable optimization and the changing layer optimization to explore different directions of the garage optimization and contrast the two optimization results at last. Finally through the Visual Studio development tool for "three-dimensional garage ANSYS modeling simulation optimization system", implements the garage modeling, forecasting, analysis, optimization, validation, and a series of functions. This provides an important reference value for stereo garage design and optimization. This paper discusses the main contents are:1. Realize the input of garage size, working condition and load applied through the finite element analysis software ANSYS and using APDL language. By analysis nine kinds of working condition which is combine with three kinds of wind and three kinds of load. Obtaining the maximum stress and displacement of each condition, find out the most dangerous working condition of the garage.2. Build the BP neural network training model in MATLAB, training the data obtained by the analysis of the study, looking for the nonlinear relationship, through simple data input to predict strength, stiffness and stability of the garage. And the accuracy of the predicted results is verified.3. Combine with discrete variable optimization and layer optimization on the basis of the continuous optimization. Optimize the garage in different directions. The essence of discrete optimization is the discreteness of design variables, by joining in the continuous optimization of discrete variables, using one dimensional search algorithm, realized the garage discrete optimization. Layer variable optimization is on the basis of the discrete optimization to hierarchical optimization of the garage, for different layers with different cross section.4. Through VS development tools, using C# language, in the.NET environment "stereo garage ANSYS modeling simulation optimization system" development. By mixed programming to realize the garage with ANSYS parametric modeling, analysis, optimization validation and other functions. By means of MATLAB engine calls, implements the BP neural network prediction.