Traumatic spinal cord injury(SCI)is a pathological condition that impairs both sensorimotor and cognitive functions.While research has long focused on understanding the pathophysiology of SCI and developing treatments...Traumatic spinal cord injury(SCI)is a pathological condition that impairs both sensorimotor and cognitive functions.While research has long focused on understanding the pathophysiology of SCI and developing treatments,only a few studies have investigated the cellular and molecular consequences that occur in the brain after trauma.From the earliest stages,the injury triggers microglial activation,increased neuronal death,and reduced hippocampal neurogenesis in the dentate gyrus.展开更多
利用精确解析模型生成的数据可辅助构建数值仿真样本集,为代理模型提供高质量训练数据,从而在降低计算成本的同时提升多目标优化效率。但现有解析建模常受电机拓扑约束,适用范围有限。为此,该文提出一种基于几何相似性迁移学习的电机代...利用精确解析模型生成的数据可辅助构建数值仿真样本集,为代理模型提供高质量训练数据,从而在降低计算成本的同时提升多目标优化效率。但现有解析建模常受电机拓扑约束,适用范围有限。为此,该文提出一种基于几何相似性迁移学习的电机代理模型优化方法。首先,依据物理结构之间的几何相似性构建易于精确解析化的相似电机;随后,建立相似电机设计变量-优化目标的解析映射模型并开展灵敏度分析;进而,对设计变量分层,将变量空间划分为高-低灵敏度子空间,以提高相似电机迁移结果与原型优化结果的一致性。少变量的高灵敏度参数空间通过原电机有限元分析(finite element analysis,FEA)数据建立常规代理模型进行优化,而多变量的低灵敏度参数空间则基于充足的相似电机解析数据并结合少量原型电机FEA数据,利用迁移学习训练多重保真代理模型完成最终优化。所提方法突破了精确解析模型拓扑限制,降低了结构复杂电机解析建模难度,并通过分层优化策略结合多重保真迁移显著提升高维优化效率,在保证精度前提下大幅减少计算量。该方法已用于内置式交替极永磁游标电机多目标优化,样机试验验证了有效性。展开更多
Hybrid Excited Flux Switching Machines(HEFSMs)unique feature of high torque density(T_(den))of Permanent Magnet(PM)machines and flux regulation capability of wound field excitation machines.Due to aforesaid unique fea...Hybrid Excited Flux Switching Machines(HEFSMs)unique feature of high torque density(T_(den))of Permanent Magnet(PM)machines and flux regulation capability of wound field excitation machines.Due to aforesaid unique features,stator active HEFSMs are preferred for EV/HEV applications.In this paper a new Segmented PM Consequent Pole HE-FSM(SPMCPHEFSM)with flux bridge is proposed for EV/HEV.The developed SPMCPHEFSM exhibits improved flux modulation and flux regulation capability at reduced PM usage(suppressed PM volume by 46.52%and PM cost by 46.48%)and eliminating stator leakage flux.First,SPMCPHEFSM is geometric optimized(GO)for investigating influence of leading design with key performance indicators such as flux linkage(Ф_(pp)),average torque(T_(avg)),cogging torque(T_(cog)),T_(den),average power(P_(avg))and power density(P_(den))and then proceeded optimized model to structure modification for optimal stator design and position of field excitation coils(FEC).Comprehensive performance analysis reveals that the developed SPMCPHEFSM show improvedФ_(pp)maximum up to 9.11%,improved T_(avg)maximum up to 23.63%,truncate T_(cog)up to 18.9%whereas T_(den)and P_(den)are boost up to 23.55%and 89.72%respectively.展开更多
文摘Traumatic spinal cord injury(SCI)is a pathological condition that impairs both sensorimotor and cognitive functions.While research has long focused on understanding the pathophysiology of SCI and developing treatments,only a few studies have investigated the cellular and molecular consequences that occur in the brain after trauma.From the earliest stages,the injury triggers microglial activation,increased neuronal death,and reduced hippocampal neurogenesis in the dentate gyrus.
文摘利用精确解析模型生成的数据可辅助构建数值仿真样本集,为代理模型提供高质量训练数据,从而在降低计算成本的同时提升多目标优化效率。但现有解析建模常受电机拓扑约束,适用范围有限。为此,该文提出一种基于几何相似性迁移学习的电机代理模型优化方法。首先,依据物理结构之间的几何相似性构建易于精确解析化的相似电机;随后,建立相似电机设计变量-优化目标的解析映射模型并开展灵敏度分析;进而,对设计变量分层,将变量空间划分为高-低灵敏度子空间,以提高相似电机迁移结果与原型优化结果的一致性。少变量的高灵敏度参数空间通过原电机有限元分析(finite element analysis,FEA)数据建立常规代理模型进行优化,而多变量的低灵敏度参数空间则基于充足的相似电机解析数据并结合少量原型电机FEA数据,利用迁移学习训练多重保真代理模型完成最终优化。所提方法突破了精确解析模型拓扑限制,降低了结构复杂电机解析建模难度,并通过分层优化策略结合多重保真迁移显著提升高维优化效率,在保证精度前提下大幅减少计算量。该方法已用于内置式交替极永磁游标电机多目标优化,样机试验验证了有效性。
文摘Hybrid Excited Flux Switching Machines(HEFSMs)unique feature of high torque density(T_(den))of Permanent Magnet(PM)machines and flux regulation capability of wound field excitation machines.Due to aforesaid unique features,stator active HEFSMs are preferred for EV/HEV applications.In this paper a new Segmented PM Consequent Pole HE-FSM(SPMCPHEFSM)with flux bridge is proposed for EV/HEV.The developed SPMCPHEFSM exhibits improved flux modulation and flux regulation capability at reduced PM usage(suppressed PM volume by 46.52%and PM cost by 46.48%)and eliminating stator leakage flux.First,SPMCPHEFSM is geometric optimized(GO)for investigating influence of leading design with key performance indicators such as flux linkage(Ф_(pp)),average torque(T_(avg)),cogging torque(T_(cog)),T_(den),average power(P_(avg))and power density(P_(den))and then proceeded optimized model to structure modification for optimal stator design and position of field excitation coils(FEC).Comprehensive performance analysis reveals that the developed SPMCPHEFSM show improvedФ_(pp)maximum up to 9.11%,improved T_(avg)maximum up to 23.63%,truncate T_(cog)up to 18.9%whereas T_(den)and P_(den)are boost up to 23.55%and 89.72%respectively.
