基于立体视觉测量系统的锅炉管道剩余寿命预测

Remaining Life Prediction of Boiler Pipes Based on Stereoscopic Vision Measurement System

锅炉管道在高温高压的环境长期工作,在蠕变效应的作用下导致其在低于屈服极限的情况下失效。深度调峰下电厂锅炉主蒸汽、冷再热管道的蠕变效应更为明显,但因其管系复杂、工况恶劣,难以实现有效监测。基于Larson-Miller参数法对电厂主汽、冷再热管道蠕变寿命进行研究,以CAESARⅡ分析不同工况下的管系的位移、应力状态,通过CNN-LSTM建立全管系实时位移、应力在线模型,以立体视觉测量系统对所建立的数据集进行验证。结果证明所提出的CNN-LSTM网络能准确重建管系位移与应力值,重建误差控制在±3%以内;根据管系蠕变分布结果可以看出管道弯折部分容易产生应力集中,并且主汽管道产生的应力集中多于冷再热管道;蠕变寿命预测结果发现管系部分节点的蠕变寿命下降速率高于其余部分,在进行检修时需要对该部分进行重点监测。

Boiler pipes operate under high temperature and pressure for extended periods.Due to the effects of creep,they can fail even when stresses are below the yield limit.Under deep load cycling,the creep effects in power plant main steam and reheat pipes are more pronounced.However,due to the complexity of the piping system and harsh working conditions,effective monitoring is challenging.Larson-Miller parameter method is utilized to research the creep life of main steam and cold reheat pipes in power plants.By using CAESAR II,the displacement and stress states of the piping system under different conditions are analyzed.A real-time displacement and stress online model for the entire piping system is established by using CNN-LSTM and validated with a stereoscopic vision measurement system.The results show that the proposed CNN-LSTM network can accurately reconstruct the displacement and stress values of the piping system,with reconstruction errors controlled within±3%.Based on the creep distribution results,it's observed that the bent parts of the pipes are prone to stress concentration,and the main steam pipes exhibit more stress concentration than the cold reheat pipes.Predictions on creep life reveal that some nodes in the piping system have a faster decline in creep life than others.These particular nodes should be a focus for monitoring during maintenance.