K-TrAdaBoost迁移学习的压裂泵故障诊断方法研究

Research on fault diagnosis method of fracturing pump based on K-TrAdaBoost transfer learning

压裂泵工作状态复杂,不易满足传统故障诊断中训练数据多且独立同分布的条件,故障诊断准确率不高。基于TrAdaBoost的迁移学习方法可有效解决上述问题,但模型训练时间较长,且只适用于二分类问题。为此,提出一种基于K-TrAdaBoost迁移学习的压裂泵故障诊断方法。该方法将大量带标签的辅助训练集与少量带标签的目标训练集结合构成足够多的训练集,通过选取K近邻算法(KNN)的最优K值,并引入高斯函数,优化惩罚因子C,计算辅助训练集与目标训练集的相似性,得到辅助训练数据集的初始权重,从而降低TrAdaBoost迭代次数,减少训练时间。迭代结束后,在模型内部引入多分类器,改变模型输出机制,实现多种故障类型的诊断。实验结果表明:所提方法可解决传统诊断方法训练数据集不足且无法独立同分布的问题,降低TrAdaBoost模型的训练时间,实现KTrAdaBoost多种故障类型的诊断,提高压裂泵故障诊断的准确率。

The working condition of fracturing pump is complicated.It is not easy to meet the conditions of large and independent and identically distributed training data in traditional fault diagnosis,resulting in low fault diagnosis accuracy.The transfer learning method based on TrAdaBoost solves the above problems,but the model takes a long time to train and is only suitable for binary classification problems. This paper presentsa fracturing pump fault diagnosis method based on K-Tradaboost transfer learning method. A large number ofauxiliary training sets with labels are combined with a small number of target training sets with labels to formenough training sets. The optimal K value of the K-Nearest Neighbor(KNN) is selected. The penalty factor C isoptimized by introducing Gaussian function. The similarity between the auxiliary training set and the targettraining set is calculated, and the initial weight of the auxiliary training data set is obtained. The iteration timesof TrAdaBoost and the training time are reduced. After the iteration, multiple classifiers are introduced into themodel to change the model output mechanism and realize multiple fault type diagnosis. The experimentalresults show that the proposed method solves the traditional diagnostic methods problem. Reduces the trainingtime of TrAdaBoost model. K-TrAdaBoost realizes the diagnosis of various fault types and improves theaccuracy of fracturing pump fault diagnosis.