目前在使用模式识别理论进行复杂桥梁地震损伤识别时仍存在许多困难，如损伤模式不易确定、损伤指标难以提取和无法构建足够数量损伤样本等。针对上述问题，提出了一种基于地震易损性分析的桥梁地震损伤模式识别新方法。首先，将地震激励分为主震和余震两类。使用主震对桥梁进行地震易损性分析以确定其在地震激励下可能发生的损伤情形，进而确定地震损伤模式和有策略地布置传感器。使用余震激励已损伤桥梁获得其动力响应信息，以构建足够数量的数据样本。然后，使用希尔伯特黄变换(Hilbert-Huang Transform, HHT)对结构响应数据进行处理，提取对损伤敏感的指标(Damage Index, DI)。最后，采用模式识别理论建立DI与损伤模式之间的对应关系，实现对桥梁地震损伤模式的识别。将所提方法应用于连续刚构桥的地震损伤识别，结果表明该方法可准确识别桥梁地震损伤，且算法具有较强的抗噪声干扰能力。

There are still many difficulties in applying the pattern recognition theory to the complex bridge seismic damage identification, for example, it is uneasy to determine damage patterns, and difficult to extract the damage index, and even it is a big challenge to obtain sufficient damage samples. To counter the problems above, a novel bridge seismic damage pattern recognition method based on seismic fragility analysis was proposed. First, seismic excitations were classified into main-shocks and after-shocks. Main-shocks were used to carry out the bridge seismic fragility analysis in order to obtain the potential damage of the bridge, and as a consequent the layout of the sensors and the patterns of seismic damage were determined strategically. After-shocks were used to excite the damaged bridges to obtain their structural responses, and hence the sufficient data samples were acquired. Then, Hilbert-Huang Transform (HHT) was adopted to deal with the structural response data to extract the Damage Index (DI), which was sensitive to structural damage. Finally, the pattern recognition theory was applied to map the relationships between DI and the damage patterns, and the bridge seismic damage identification was achieved. The proposed method was adopted to identify the seismic damage of a continuous rigid frame girder bridge, the results show that the method can identify the bridge seismic damage patterns precisely and the noise immunity of the method is favorable.