The paper proposed the research and implement of text similarity system based on power spectrum analysis. It is not difficult to imagine that the signals of brain are closely linked with writing process. So we build t...The paper proposed the research and implement of text similarity system based on power spectrum analysis. It is not difficult to imagine that the signals of brain are closely linked with writing process. So we build text modeling and set pulse signal function to get the power spectrum of the text. The specific detail is getting power spectrum from economic field to build spectral library, and then using the method of power spectrum matching algorithm to judge whether the test text belonged to the economic field. The method made text similarity system finish the function of text intelligent classification efficiently and accurately.展开更多
J-TEXT, formerly TEXT-U at the University of Texas at Austin in USA, is a medium-sized tokamak at the Huazhong University of Science and Technology. The toroidal field （TF） power supply of this tokamak should provid...J-TEXT, formerly TEXT-U at the University of Texas at Austin in USA, is a medium-sized tokamak at the Huazhong University of Science and Technology. The toroidal field （TF） power supply of this tokamak should provide a current of 160 kA and a flat duration of 500 ms for the toroidal field coils to generate a maximum toroidal field of 3 T at the geometric center of the vaccum vessel. This paper presents a design of a new control system which takes the real-time feedback control option for the TF power supply operation. The system was tested successfully during the commissioning. In the first experimental campaign of J-TEXT, the system effectively controlled the power supply to provide a fiat current up to 92.5 kA and therein the TF produced reached 1.74 T, which enabled the machine to generate the first plasma successfully.展开更多
In order to advance the research on suppressing tearing modes and driving plasma rotation, a DC power supply （PS） system has been developed for dynamic resonant magnetic perturbation （DRMP） coils and applied in th...In order to advance the research on suppressing tearing modes and driving plasma rotation, a DC power supply （PS） system has been developed for dynamic resonant magnetic perturbation （DRMP） coils and applied in the J-TEXT experiment. To enrich experimental phe- nomena in the J-TEXT tokamak, applying the circulating current four-quadrant operation mode in the DRMP DC PS system is proposed. By using the circulating current four-quadrant oper- ation, DRMP coils can be smoothly controlled without the dead-time when the current polarity reverses. Essential circuit analysis, control optimization and simulation of desired scenarios have been performed for normal current. Relevant simulation and test results are also presented.展开更多
Joint-TEXT （J-TEXT） is a medium-sized tokamak constructed at Huazhong University of Science and Technology. At present, the poloidal field power supply consists of three subsystem sets, namely Ohmic heating field po...Joint-TEXT （J-TEXT） is a medium-sized tokamak constructed at Huazhong University of Science and Technology. At present, the poloidal field power supply consists of three subsystem sets, namely Ohmic heating field power supply, vertical field power supply, and horizontal field power supply. The structure and function of the poloidal field （PF） power supply as well as the updated control system of the power supply is described. The new scheme uses the realtime feedback control method for the poloidal field power supply. The experimental results show that the poloidal field energized by poloidal field power supply meets the J-TEXT operational requirement perfectly.展开更多
当前电网数字化转型升级,电力设备智能运维技术快速发展,在运维过程中积累了大量包含电网重要信息的电力设备缺陷文本。由于文本数据标签稀疏,以及描述语言的模糊性、差异性等问题,电力文本中的运维信息难以被有效挖掘。文章提出了一种...当前电网数字化转型升级,电力设备智能运维技术快速发展,在运维过程中积累了大量包含电网重要信息的电力设备缺陷文本。由于文本数据标签稀疏,以及描述语言的模糊性、差异性等问题,电力文本中的运维信息难以被有效挖掘。文章提出了一种针对电力设备缺陷文本的数据增强方法。首先,使用缺陷文本数据集微调预训练模型ERNIE(enhanced representation through knowledge integration),应用多阶段知识掩码策略将电气领域专业知识集成到对缺陷文本的动态编码中;然后在流形假设的基础上基于降噪自动编码器架构设计破坏函数和重建函数,遵循基于信息价值的掩码单元选择策略构建破坏函数,基于微调过的ERNIE构建重建函数,在“破坏-重建”过程中获得位于原始数据流形范围内的增强样本;其次对增强数据集基于影响函数和多样性度量进行数据选择,过滤掉数据质量差和重复度高的增强样本;最后通过多层训练框架,将增强数据应用于各种缺陷文本挖掘任务。算例基于真实设备巡检、检修记录构建了电力设备缺陷文本等级分类任务。结果表明,所提出的算法对缺陷文本挖掘效果有较大提升,并且可以广泛灵活地应用在多种电力设备缺陷文本挖掘任务中。展开更多
核电装备质量文本描述了核电装备在设计、采购、施工和调试阶段出现的质量缺陷等问题。由于不同阶段质量事件的发生频率不同,且同一装备对应不同阶段的质量文本中存在相同的关键词和相似的表述形式,针对类型数量不均衡和语义描述耦合的...核电装备质量文本描述了核电装备在设计、采购、施工和调试阶段出现的质量缺陷等问题。由于不同阶段质量事件的发生频率不同,且同一装备对应不同阶段的质量文本中存在相同的关键词和相似的表述形式,针对类型数量不均衡和语义描述耦合的质量文本分类问题,提出一种融合正则反馈焦点损失函数的改进循环池化网络分类模型。首先,采用BERT(Bidirectional Encoder Representation from Transformers)将核电装备质量文本转化为词向量;然后,提出一个改进的3层循环池化网络的分类模型结构,通过增加中间层并选择合适权重,扩大参数训练的提取空间,提升表征质量缺陷语义特征的能力;接着,提出正则反馈焦点损失函数来训练提出分类模型的参数,通过正则项使损失函数的梯度变化更稳定,根据反馈项对损失函数进行基于真实值和预测值之间误差的迭代调整,解决了不均衡样本在训练过程中梯度偏向不均衡的问题;最后,通过归一化指数函数计算出核电装备质量事件对应的阶段。在某核电公司真实数据集和公共数据集上,与Fast_Text网络相比,所提模型的F1值分别提高了2个百分点和1个百分点,实验结果表明该模型在文本分类任务中具有较高的准确性。展开更多
文摘The paper proposed the research and implement of text similarity system based on power spectrum analysis. It is not difficult to imagine that the signals of brain are closely linked with writing process. So we build text modeling and set pulse signal function to get the power spectrum of the text. The specific detail is getting power spectrum from economic field to build spectral library, and then using the method of power spectrum matching algorithm to judge whether the test text belonged to the economic field. The method made text similarity system finish the function of text intelligent classification efficiently and accurately.
基金National Basic Research Program of China(973 Program)(No.2008CB717805)
文摘J-TEXT, formerly TEXT-U at the University of Texas at Austin in USA, is a medium-sized tokamak at the Huazhong University of Science and Technology. The toroidal field （TF） power supply of this tokamak should provide a current of 160 kA and a flat duration of 500 ms for the toroidal field coils to generate a maximum toroidal field of 3 T at the geometric center of the vaccum vessel. This paper presents a design of a new control system which takes the real-time feedback control option for the TF power supply operation. The system was tested successfully during the commissioning. In the first experimental campaign of J-TEXT, the system effectively controlled the power supply to provide a fiat current up to 92.5 kA and therein the TF produced reached 1.74 T, which enabled the machine to generate the first plasma successfully.
基金supported by the National ITER Project Foundation of China(No.2014GB118000)National Natural Science Foundation of China(No.11275080)
文摘In order to advance the research on suppressing tearing modes and driving plasma rotation, a DC power supply （PS） system has been developed for dynamic resonant magnetic perturbation （DRMP） coils and applied in the J-TEXT experiment. To enrich experimental phe- nomena in the J-TEXT tokamak, applying the circulating current four-quadrant operation mode in the DRMP DC PS system is proposed. By using the circulating current four-quadrant oper- ation, DRMP coils can be smoothly controlled without the dead-time when the current polarity reverses. Essential circuit analysis, control optimization and simulation of desired scenarios have been performed for normal current. Relevant simulation and test results are also presented.
基金supported by National Basic Research Program 973 Program of China(No.2008CB717805)
文摘Joint-TEXT （J-TEXT） is a medium-sized tokamak constructed at Huazhong University of Science and Technology. At present, the poloidal field power supply consists of three subsystem sets, namely Ohmic heating field power supply, vertical field power supply, and horizontal field power supply. The structure and function of the poloidal field （PF） power supply as well as the updated control system of the power supply is described. The new scheme uses the realtime feedback control method for the poloidal field power supply. The experimental results show that the poloidal field energized by poloidal field power supply meets the J-TEXT operational requirement perfectly.
文摘当前电网数字化转型升级,电力设备智能运维技术快速发展,在运维过程中积累了大量包含电网重要信息的电力设备缺陷文本。由于文本数据标签稀疏,以及描述语言的模糊性、差异性等问题,电力文本中的运维信息难以被有效挖掘。文章提出了一种针对电力设备缺陷文本的数据增强方法。首先,使用缺陷文本数据集微调预训练模型ERNIE(enhanced representation through knowledge integration),应用多阶段知识掩码策略将电气领域专业知识集成到对缺陷文本的动态编码中;然后在流形假设的基础上基于降噪自动编码器架构设计破坏函数和重建函数,遵循基于信息价值的掩码单元选择策略构建破坏函数,基于微调过的ERNIE构建重建函数,在“破坏-重建”过程中获得位于原始数据流形范围内的增强样本;其次对增强数据集基于影响函数和多样性度量进行数据选择,过滤掉数据质量差和重复度高的增强样本;最后通过多层训练框架,将增强数据应用于各种缺陷文本挖掘任务。算例基于真实设备巡检、检修记录构建了电力设备缺陷文本等级分类任务。结果表明,所提出的算法对缺陷文本挖掘效果有较大提升,并且可以广泛灵活地应用在多种电力设备缺陷文本挖掘任务中。
文摘核电装备质量文本描述了核电装备在设计、采购、施工和调试阶段出现的质量缺陷等问题。由于不同阶段质量事件的发生频率不同,且同一装备对应不同阶段的质量文本中存在相同的关键词和相似的表述形式,针对类型数量不均衡和语义描述耦合的质量文本分类问题,提出一种融合正则反馈焦点损失函数的改进循环池化网络分类模型。首先,采用BERT(Bidirectional Encoder Representation from Transformers)将核电装备质量文本转化为词向量;然后,提出一个改进的3层循环池化网络的分类模型结构,通过增加中间层并选择合适权重,扩大参数训练的提取空间,提升表征质量缺陷语义特征的能力;接着,提出正则反馈焦点损失函数来训练提出分类模型的参数,通过正则项使损失函数的梯度变化更稳定,根据反馈项对损失函数进行基于真实值和预测值之间误差的迭代调整,解决了不均衡样本在训练过程中梯度偏向不均衡的问题;最后,通过归一化指数函数计算出核电装备质量事件对应的阶段。在某核电公司真实数据集和公共数据集上,与Fast_Text网络相比,所提模型的F1值分别提高了2个百分点和1个百分点,实验结果表明该模型在文本分类任务中具有较高的准确性。
