融合GA优化算法的数字孪生模型在石油旋转机械诊断中的应用

Application of digital twin model with GA optimization algorithm in diagnosis of petroleum rotating machinery

摘要:

为根据模态信号频率水平确定石油旋转机械当前故障行为所属类别,实现对机械设备故障行为的准确诊断,针对融合GA优化算法的数字孪生模型在石油旋转机械诊断中的应用展开研究。定义GA算法优化规则,并在此基础上,建立数字孪生模型,再联合相关故障行为数据,完成对石油旋转机械运行数据的聚类运算,实现基于数字孪生模型的石油旋转机械运行数据聚类处理。计算运行数据损失情况,通过模态分解描述性样本的方式,将核心诊断信息重新耦合在一起,联合求解所得的超参数指标,定义具体的数据样本集中训练模式,实现对石油旋转机械的诊断。实验结果表明,上述诊断方法的应用,对于每一类故障行为模态信号频率的诊断都属于该信号的标准频率数值区段之内,符合100%精准诊断机械故障行为的应用需求,在准确诊断石油旋转机械故障行为方面的可行性能力较为突出。

Abstract:

In order to determine the current fault behavior of petroleum rotating machinery according to the modal signal frequency level, and realize the accurate diagnosis of mechanical equipment fault behavior, the application of digital twin model with fusion GA optimization algorithm in the diagnosis of petroleum rotating machinery is studied. The optimization rules of GA algorithm are defined, and on this basis, the digital twin model is established, and then the relevant fault behavior data are combined to complete the clustering operation of the oil rotating machinery operating data, and the oil rotating machinery operating data cluster processing based on the digital twin model is realized. The loss of operating data is calculated, the core diagnostic information is recoupled together by means of modal decomposition of descriptive samples, the obtained hyperparameter index is solved jointly, and the specific training mode of data samples is defined to realize the diagnosis of petroleum rotating machinery. The experimental results show that the application of the above diagnosis method belongs to the standard frequency value range of the signal for each type of fault behavior mode signal, which meets the application demand of 100% accurate diagnosis of mechanical fault behavior, and has outstanding feasibility in accurately diagnosing the fault behavior of petroleum rotating machinery.