基于卷积自编码神经网络的锂离子电池健康状况评估方法研究 |
| |
引用本文: | 侯瑞磊,范秋华.基于卷积自编码神经网络的锂离子电池健康状况评估方法研究[J].计算机测量与控制,2020,28(8):265-269. |
| |
作者姓名: | 侯瑞磊 范秋华 |
| |
作者单位: | 青岛大学电气工程学院,山东青岛 266071;青岛大学电气工程学院,山东青岛 266071 |
| |
摘 要: | 目前锂离子电池已被广泛用作能量存储系统,在手机、电动汽车和飞机中均有广泛的应用。然而锂离子电池在使用过程中存在一定的危险性,若不能及时对电池健康状态评估(SOH)发现危险将会导致十分严重的后果。因此,研究一种基于卷积神经网络的锂离子电池健康状况评估方法,该方法通过使用卷积自编码神经网络对电池状态数据进行特征提取,有效提升了评估的准确率,并且神经网络能够在使用过程中不断进行学习,具有较高的灵活性,最后通过使用NASA公开的锂电池数据集测试,评估准确率达到93.6%,相比传统方法有较大提升。
|
关 键 词: | 锂电池 SOH 卷积 自编码 Softmax |
收稿时间: | 2019/12/26 0:00:00 |
修稿时间: | 2020/1/16 0:00:00 |
HealthAssessmentMethodofLithiumIonBatteryBasedonConvolutionalSelf-EncodingNeuralnetwork |
| |
Abstract: | At present, lithium-ion batteries have been widely used as energy storage systems, and they are widely used in mobile phones, electric vehicles and aircraft. However, there are certain dangers in the use of lithium ion batteries. If the battery health status (SOH) is not found in time, the danger will lead to very serious consequences. Therefore, a method for assessing the health of lithium-ion batteries based on a convolutional neural network is studied. This method uses a convolutional self-encoding neural network to extract the characteristics of the battery state data, effectively improving the accuracy of the evaluation, and the neural network can Continuous learning during the use process has high flexibility. Finally, by using the lithium battery data set published by NASA, the evaluation accuracy rate is 93.6%, which is greatly improved compared with the traditional method. |
| |
Keywords: | Lithium Battery SOH Convolution Self-Encoding Softmax |
本文献已被 万方数据 等数据库收录! |
| 点击此处可从《计算机测量与控制》浏览原始摘要信息 |
|
点击此处可从《计算机测量与控制》下载全文 |
|