中文摘要：

基于陕西省某连续刚构桥实桥监测系统采集到的桥梁有效竖向预应力长期变化数据，以实测数据作为初始序建立灰色系统理论预测模型GM（1，1），分析竖向预应力损失随时间的变化规律，进行时间非等步距残差修正，提出竖向预应力长期损失预测模型．采用该模型预测竖向预应力长期损失占控制张拉力的比例为7．92％，与实桥监测的7．8％～8．2％基本吻合，表明若采用公路桥规设计张拉力的0．6倍作为竖向预应力施工控制标准，将导致桥梁后期有效竖向预应力达不到要求．建议施工过程中锚下应力控制应充分考虑长期损失效应影响，压浆前锚下有效应力不小于设计张拉力o．7倍为宜．这为竖向预应力施工及无损检测提供了理论依据和参考，对PC混凝土箱梁竖向预应力体系应用发展具有重要的理论意义及实用价值．

英文摘要：

The paper based on the vertical prestress loss data of a concrete continuous rigid frame bridge collected by the real bridge monitoring system in Shaanxi province, establishes the grey theory forecasting model GM （1,1） with the measured data as the initial order, analyzes the law of vertical prestress loss with time, and puts forward long-term vertical prestress loss prediction formula. Using this model to predict the long-term loss of vertical prestress accounted for 7.92% of the control tension~ and it is consistent with the 7.8 %-8.2 % of the actual bridge monitoring basically. The results show that if using the design of the highway bridge gauge,0.6 times of the tension forces, as vertical prestress construction control standard, will cause the late effective vertical prestress bridge can not meet the requirements. Based on the research outcomes, effective stress under anchor in vertical prestress construction should be more than 0.7 times of the designed stretch-draw force. The results provide a theoretical basis and reference for vertical prestress construction and NDE（nondestruetive examination）, and are of great importance of practical value for the application and development of vertical prestress system in prestressed concrete box-girder bridge.