注意力ConvLSTM模型在RUL预测中的应用

Application of Attention-ConvLSTM Model for Remaining Useful Life Prediction

预测性维护的应用能够极大地降低企业运维成本,而设备剩余使用寿命(Remaining Useful Life,RUL)预测是预测性维护的关键技术之一.针对传统RUL预测算法难以提取时序数据的潜藏特征以及特征权重分配不合理的问题,本文提出一种基于注意力机制(Attention Mechanism)的卷积长短时记忆(Convolution Long-Short Term Memory,ConvLSTM)预测模型,该模型充分利用LSTM网络处理和预测长期时间序列的优势,并引入注意力机制对产生显著影响的特征因子提高权重,极大地优化了模型的时空特征提取能力.为验证模型预测效果,本文以NASA提供的CMAPSS数据集为对象进行实验,以均方根误差(Root Mean Squared Error,RMSE)和数据集自定义的Score为评价指标,将预测结果与其他RUL预测算法作比对,证明了该模型具有更佳的预测准确性.

The application of predictive maintenance technology can greatly reduce the operation and maintenance costs of enterprises,and the remaining useful life(RUL)prediction of equipment is one of the key technologies of predictive maintenance.Aiming at the problem that the traditional RUL prediction algorithm is difficult to extract the hidden features of the time series data and the distribution of feature weights is unreasonable,this paper proposes a Convolution Long-Short Term Memory(ConvLSTM)prediction model based on the Attention Mechanism.This model makes full use of the advantages of LSTM networks to process and predict long-term time series,and introduce the attention mechanism to significantly increase the weight of feature factors,which greatly optimizes the space-time feature extraction capability of the model.In order to verify the prediction effect of the model,this paper uses the CMAPSS data set provided by NASA as an experiment,which take root mean square error(Root Mean Squared Error,RMSE)and the score of the data set as evaluation indicators.The prediction results are compared with other RUL prediction algorithms,which proves that the model has better prediction accuracy.