【摘要】 介绍了框架结构的连接方法，分析了半刚性节点的研究现状及进展情况，针对目前框架结构的链接方法中存在的问题，以半刚性节点为研究对象，利用神经网络方法建立了粒子群RBF神经网络算法，按照节点聚类分析方法建立了神经网络算法的聚类分析并给出了其分析合理性评判标准。对节点模型初始转动刚度进行了预测分析，预测结果与实际结果较为接近，表明该方法的准确性较好，并对螺栓间距、端板厚度及有效宽度3个参数进行了优化分析，得到了最优解。更多还原

【Abstract】 The connection method of frame structure is introduced, and the research status and development of semi-rigid joints are analyzed. Aiming at the problems existing in the link method of frame structure, taking the semi-rigid joints as the research object, the particle swarm RBF neural network algorithm is established by using the neural network method, the node clustering analysis of the neural network algorithm is established according to the node clustering analysis method, and the evaluation standard of its analysis rationality is given. The initial rotational stiffness of the node model is predicted and analyzed. The predicted results are close to the actual results, which shows that the accuracy of this method is good. The three parameters of bolt spacing, end plate thickness and effective width are optimized and analyzed, and the optimal solution is obtained.更多还原