烟气轮机涡轮盘高温合金剩余寿命预测

Remaining Life Prediction of Superalloy for Turbine Disk Using Artificial Neural Network

本文针对已运行60000h的Waspaloy合金烟气轮机涡轮盘进行剩余寿命预测分析,采用不同试验条件下得到的waspaloy合金持久寿命数据对人工神经网络模型进行训练,得到预测精度较高的模型参数,建立温度、应力等服役条件与持久断裂寿命之间的人工神经网络模型,并利用该模型对Waspaloy合金涡轮盘的剩余寿命进行预测分析。结果表明,中间层节点个数为15时,所建立的人工神经网络模型对Waspaloy合金的持久断裂寿命具有最好的统计预测精度,并可以良好地表征Waspaloy合金剩余持久寿命与服役条件间复杂的非线性关系。

Remaining life prediction for Waspaloy alloy turbine disk after a long service period （above 60,000 hours） was studied by Artificial Neural Network （ANN） model. The model was established to describe the relationship between the remaining rupture life and such service conditions as temperature and stress. At the same time, the remaining rupture life tests under different temperatures and stresses were carried out. Based on the test data, the model was trained to get appropriate parameters and used to predict the remaining life of the turbine disk. The results show that the ANN model with 15 middle layer nodes can represent the best prediction accuracy, and can suitably depict the nonlinear relationship between the remaining rupture life and the different service conditions.