基于CNN 与BiLSTM 的退化设备剩余使用寿命预测

针对DC-DC 变换器难以建立物理退化模型的问题，提出了一种基于卷积神经网络和双向长短期记忆神经网络的剩余寿命预测算法。首先，通过对DC-DC 变换器中关键器件进行失效模式分析，得到DC-DC 变换器的失效特征参数；然后，采用卷积神经网络对多传感器信号进行融合和特征提取，获得百分比指数退化指标；其次，将获得的健康指标输入到融合注意力机制的双向记忆神经网络中，利用注意力机制对输出结果进行加权融合，利用不同的权重值优化预测模型，最后，利用蒙特卡洛丢弃法获得剩余寿命的区间估计，并通过仿真数据验证了所提出方法的有效性。

Prediction of Remaining Useful Life of Degraded Equipment Basedon CNN and BiLSTM

Aiming at the problem that it is difficult to establish the indoor degradation model of DC-DC converters, aresidual life estimation algorithm based on convolutional neural network and bidirectional long short-term memory neuralnetwork is proposed. First, by analyzing the failure mode of the key components in the DC-DC converter, the failurecharacteristic parameters of the DC-DC converter are obtained; then, the convolutional neural network is used to fuseand extract the multi-sensor signals, and the percentage index degradation index is obtained. ; Secondly, the obtainedhealth indicators are input into the bidirectional memory neural network fused with the attention mechanism, the outputresults are weighted and fused by the attention mechanism, and the prediction model is optimized by using differentweight values. Finally, the Monte Carlo discarding method is used to obtain the interval estimation of remaining life andthe effectiveness of the proposed method is verified by simulation data.