为保障电动车辆的可靠性和安全性,提出了一种dropout Monte Carlo(dropout-MC)递归神经网络的锂离子动力电池系统的剩余寿命(RUL)预测方法。以等电压充电时间间隔作为间接健康因子,考虑外部干扰和容量再生现象的影响,以变分模态分解(VMD...为保障电动车辆的可靠性和安全性,提出了一种dropout Monte Carlo(dropout-MC)递归神经网络的锂离子动力电池系统的剩余寿命(RUL)预测方法。以等电压充电时间间隔作为间接健康因子,考虑外部干扰和容量再生现象的影响,以变分模态分解(VMD)来获得电池退化趋势。以改进的递归神经网络模型——长短时间序列(LSTM)来获得剩余寿命预测。以dropout-MC采样方法来表征锂离子电池剩余寿命的不确定性,并获得锂离子电池RUL的95%置信区间。结果表明:相较于传统的极限学习机(ELM)方法和非线性自回归神经网络(NARX)方法,该文方法的剩余寿命预测性能指标均低于2.4%。因而,该方法具有优越的预测性能,且获得预测的置信区间。展开更多
This paper deals with the stochastic stability of networked control systems with the presence of network- induced delay and transmitted data dropout. Based on the Lyapunov approach, sufficient conditions for the mean-...This paper deals with the stochastic stability of networked control systems with the presence of network- induced delay and transmitted data dropout. Based on the Lyapunov approach, sufficient conditions for the mean-square stability of the networked control system are derived subject that the sequence of transmission interval is driven by an identically independently distributed sequence and by a finite state Markov chain, respectively. Stabilization controllers are constructed in terms of linear matrix inequalities correspondingly. An example is provided to illustrate our results.展开更多
低压差线性电压调整器应用广泛,电源纹波抑制比是反映其性能指标的关键参数之一。为解决传统电源纹波抑制比测试方法测量频率范围较小、测试效率低、难以满足高电源纹波抑制比测试等不足,提出基于功率分配器和低频网络分析仪相结合的电...低压差线性电压调整器应用广泛,电源纹波抑制比是反映其性能指标的关键参数之一。为解决传统电源纹波抑制比测试方法测量频率范围较小、测试效率低、难以满足高电源纹波抑制比测试等不足,提出基于功率分配器和低频网络分析仪相结合的电源纹波抑制比测试方法,并采用典型低压差线性电压调整器对基于功率分配器和基于电感电容总和节点法的两种测试方法进行测试验证。实验结果表明:基于功率分配器的电源纹波抑制比测试方法最低测试频率可达30 Hz,可满足70 d B以上电源纹波抑制比的测试需求,具有频率测量范围更宽、测试效率高等特点。展开更多
文摘为保障电动车辆的可靠性和安全性,提出了一种dropout Monte Carlo(dropout-MC)递归神经网络的锂离子动力电池系统的剩余寿命(RUL)预测方法。以等电压充电时间间隔作为间接健康因子,考虑外部干扰和容量再生现象的影响,以变分模态分解(VMD)来获得电池退化趋势。以改进的递归神经网络模型——长短时间序列(LSTM)来获得剩余寿命预测。以dropout-MC采样方法来表征锂离子电池剩余寿命的不确定性,并获得锂离子电池RUL的95%置信区间。结果表明:相较于传统的极限学习机(ELM)方法和非线性自回归神经网络(NARX)方法,该文方法的剩余寿命预测性能指标均低于2.4%。因而,该方法具有优越的预测性能,且获得预测的置信区间。
基金National Natural Science Foundation of China (No.60874021, 60674046)Natural Science Foundation from JiangsuProvince (No.BK2007061)+1 种基金Natural Science Foundation from Jiangsu Provincial Department for Education (No.06KJB120088)Research Fundfor Doctoral Program of Nantong University (No.07B14).
文摘This paper deals with the stochastic stability of networked control systems with the presence of network- induced delay and transmitted data dropout. Based on the Lyapunov approach, sufficient conditions for the mean-square stability of the networked control system are derived subject that the sequence of transmission interval is driven by an identically independently distributed sequence and by a finite state Markov chain, respectively. Stabilization controllers are constructed in terms of linear matrix inequalities correspondingly. An example is provided to illustrate our results.
文摘低压差线性电压调整器应用广泛,电源纹波抑制比是反映其性能指标的关键参数之一。为解决传统电源纹波抑制比测试方法测量频率范围较小、测试效率低、难以满足高电源纹波抑制比测试等不足,提出基于功率分配器和低频网络分析仪相结合的电源纹波抑制比测试方法,并采用典型低压差线性电压调整器对基于功率分配器和基于电感电容总和节点法的两种测试方法进行测试验证。实验结果表明:基于功率分配器的电源纹波抑制比测试方法最低测试频率可达30 Hz,可满足70 d B以上电源纹波抑制比的测试需求,具有频率测量范围更宽、测试效率高等特点。