基于改进粒子滤波的重型燃气轮机跳机故障预测

Tripping Fault Prediction of Heavy-duty Gas Turbines Based on Improved Particle Filter

重型燃气轮机是清洁发电的重要装备,其轴系的振动水平是机组运行状态的直观表征。跳机故障是由于振动加大而触发的非计划突然停机,会对燃气轮机的核心部件(如叶片、拉杆等)产生较大冲击,造成设备损伤。提出基于改进粒子滤波的重型燃气轮机振动趋势预测方法,通过对粒子滤波方法的分析,提出一种二次重采样策略,使得改进粒子滤波对粒子匮乏现象更具抵抗力,具有更好的适应性。所提方法在某300 MW重型燃气轮机的跳机故障中得到验证,能够准确预测跳机的故障时刻,为燃气轮机的控制策略调整提供指导。

Heavy-duty gas turbine was the significant equipment in clear energy,and the vibration level of the shafting system is a visual representation of the operating states.Tripping faults were as a kind of unplanned sudden shutdown triggered by increasing vibrations,which would cause a large impact on the core components of the gas turbine,such as blades and tie rods,resulting in equipment damages.A method for predicting the vibration trend of heavy-duty gas turbines was proposed based on improved particle filter.By analyzing the particle filter,a secondary resampling strategy was proposed to make the improved particle filter more resistant to particle degeneracy and improve the adaptability of particle filter.The improved method was verified in a tripping fault of a 300 MW heavy-duty gas turbine,which shows a superior prediction accuracy of tripping fault time.The proposed approach may guide the control strategy of gas turbines.