基于遗传算法和神经网络的塔机结构动态优化设计

Dynamic Optimum Design of Tower Crane Based on Neural Networks and Genetic Algorithms

利用遗传算法和BP神经网络建立复杂结构系统动态优化的计算模型,该模型可代替系统原来的有限元模型,用于振动系统的快速重分析。首先对塔式起重机结构系统进行模态分析及谐响应动力学分析,找出对结构动态特性影响最大的模态频率,再利用灵敏度分析,确定对动态特性较敏感的设计变量作为神经网络的输入变量,并利用正交试验法确定神经网络训练样本,用有限元模型计算出样本点数据,建立反映结构振动特性的人工神经网络模型,最后利用遗传算法对所建立的神经网络模型寻优,得到使结构动态性能最优的设计参数。

A dynamic optimal computational model for the complex structure system with generic algorithm （GA） and BP neural network（NN） was presented. Instead of the traditional finite element model, this model can be used for the fast re-analysis for the vibration system. Firstly, the harmonic response kinetics analysis can be processed on a tower crane structure system and can find out the mode frequency which has the strongest effect on the system dynamic behavior. Secondly, from the sensitivity analysis, the design variables which are more sensitive to the system dynamic behavior can be confirmed as the input variables. Then an orthogonal experimentation was used in choosing the training sample data and the sample data was calculated through the finite element model. The artificial neural network model which presented the dynamic behavior of the structure vibration was established. At last, the neural network model will be optimized through the generic algorithm and the optimal parameters of the structure dynamic behavior will be obtained.