中文摘要：

由于失效样本少，传统的可靠性分析方法并不适用于特种设备的寿命预测．根据产品研制阶段的实验数据和同类产品的相关历史记录确定可靠性评估的验前信息．结合小样本现场试验数据，应用Bayes方法融合产品的验前信息和现场信息，得到产品寿命的验后分布，进而建立产品可靠寿命的预测模型．以桥式起重机为例，针对具有可测量退化量的潜在故障模式——主梁下挠超标，推导了主梁两阶段可靠寿命的退化规律，得出了在给定可信度下主梁从下挠值为0至发生潜在故障和发生潜在故障后继续使用至发生功能故障的工作时间，进而为主梁的检测和维修提供理论基础．该研究对于缺乏大样本试验数据的特种设备的可靠寿命预测与风险预防具有借鉴意义．

英文摘要：

The traditional reliability analysis method of huge data samples cannot be applied to life prediction of the special equipments because of lacking failure samples. Prior information of reliability estimation can be calculated according to the failure data from laboratory during designing and precious records of similar products. Then, the reliability estimation modal was built based on Bayes theorem combining prior information with small samples from field test, and the posterior distribution was obtained. Taking the main beam deflection of overhead travelling crane as one measurable potential failure mode, reliable life prediction of two-cycle is realized, i. e. the work time from deflection zero to potential failures and work time from potential failures to functional failure are calculated. The research provides not only some theoretical basis for the main beam inspection and maintenance but also a reference for reliable life prediction and risk prevention of special equipments which lack failure data from factory and laboratory.