固体火箭发动机对接装配密封圈应力在线预测方法

Online stress prediction method for sealing ringduring SRM docking assembly

在固体火箭发动机燃烧室和喷管对接装配过程中,为准确实时预测密封圈应力,以确保发动机的装配质量,提出了一种基于Kriging模型的密封圈对接装配应力预测方法。首先,分析装配工况,利用有限元分析方法计算出多种工况下密封圈的应力-应变;其次,使用生成对抗网络的方法扩大数据样本空间;之后,利用拉丁超立方抽样法选取一定数量的应力-应变数据构建Kriging模型;最后,根据定义的加点准则迭代优化Kriging模型,实现主动学习,由此得到密封圈应力预测的数字孪生模型。装配时,通过六自由度并联平台的力位传感器实时采集的信号数据,作为数字孪生预测模型的输入;通过实时读取模型输出,实现对接过程中的装配质量实时在线预测反馈。

In order to accurately predict the stress of sealing ring during the docking assembly of a SRM chamber and nozzle in real time and also ensure the assembly quality,a prediction method for docking assembly stress of SRM sealing ring based on Kriging model was proposed.Firstly,the stress and strain of the sealing ring under various working conditions were calculated by means of FEA.Secondly,the method of generating adversarial networks was used to expand the data sample space.Then,a certain number of stress-strain data were selected by Latin hypercube sampling method(LHS)to construct Kriging model.Finally,the Kriging model was iteratively optimized according to the defined adding point criterion to realize active learning,and the digital twin model of seal ring stress prediction was obtained.The signal data collected in real time by the force and position sensor of the six-degree-of-freedom parallel platform during assembly was used as the input of the digital twin prediction model.By reading model output,the real-time online predictive feedback of assembly quality in docking process was realized.