Parametric Modeling And Optimization Studying Of Gantry Crane's Jib Structure Based On APDL

With the development of the society, the economical trade is significantly strengthened between China and the other countries in the world. The volume of China's import and export trade increases year by year. The demands of various cranes are growing. Gantry crane is the necessary handling machines in the large port, which have advantages such as flexibility, convenience and wide adaptability etc. Based on these, gantry cranes are playing more and more important roles, technical requirements of them are also getting higher and higher. Especially with the wide use of computer technology, many advanced interdisciplinary design methods appeared, which promotes the new design and manufacture technology of gantry cranes. The strong international market competition rely more and more on technological competition, which promote the technical characteristic of crane getting into a new stage of development. The design calculation method of the port machines need to be constantly updated and amended, and the technique should be go to the way of functionality, economic, reliability and safety.As a designer, production costs should be reduced, and the reliability of products must be guaranteed. Therefore, the optimized research about the gantry crane's jib structure has been done in the paper.The finite analysis model is established based on ANSYS, and the relationship is set up between some important parameters of the jib structure and maximum stress, displacement, which provides the learning samples of the following neural network.The neural network model is established based on the non-linear reflect function; the BP neural network is trained according to the data sample coming from the finite analysis. In the process of establishing the BP neural network, the author discusses many questions, including the normalization of the sample data, the selection of the hidden layer network's number. Through the creation of the BP neural network, the author achieved the forecast of the jib structure's maximum stress and displacement by neural network, and overcame many shortcomings, such as wasting too much time on finite element analysis, and the lack of operation in practice before.The optimal model of jib structure is based on optimal theory, the pipes' thickness and length are variables of optimal design, The minimum weight of the jib structure is the target function. The constraint equations are listed based on it can meet the requirement of rigidity, stability and technology. The jib structure is optimized by means of the MAT Lab's optimization toolbox, and the optimal results are inspected, which proves the feasibility of this optimization method.