To implement the performance-based seismic design of engineered structures, the failure modes of members must be classified. The classification method of column failure modes is analyzed using data from the Pacific Earthquake Engineering Research Center (PEER). The main factors affecting failure modes of columns include the hoop ratios, longitudinal reinforcement ratios, ratios of transverse reinforcement spacing to section depth, aspect ratios, axial compression ratios, and flexure-shear ratios. This study proposes a data-driven prediction model based on an artificial neural network (ANN) to identify the column failure modes. In this study, 111 groups of data are used, out of which 89 are used as training data and 22 are used as test data, and the ANN prediction model of failure modes is developed. The results show that the proposed method based on ANN is superior to traditional methods in identifying the column failure modes.

为实现工程结构的基于性能的抗震设计，必须对构件的失效模式进行分类。使用太平洋地震工程研究中心 (PEER) 的数据分析柱失效模式的分类方法。影响柱破坏模式的主要因素有环箍比、纵筋比、横筋间距与截面深度比、纵横比、轴压比和弯剪比。本研究提出了一种基于人工神经网络 (ANN) 的数据驱动预测模型来识别柱失效模式。本研究使用111组数据，其中89组作为训练数据，22组作为测试数据，建立了ANN失效模式预测模型。