为解决风电机组故障诊断中故障样本不足而导致模型准确率不高的问题,将当下备受关注的数据增强方法-去噪扩散概率模型(denoising diffusion probability model,DDPM)引入到故障诊断领域以生成大量高质量的故障样本数据集。因此,结合Tran...为解决风电机组故障诊断中故障样本不足而导致模型准确率不高的问题,将当下备受关注的数据增强方法-去噪扩散概率模型(denoising diffusion probability model,DDPM)引入到故障诊断领域以生成大量高质量的故障样本数据集。因此,结合Transformer网络,提出了一种DDPM-Transformer风电机组故障样本生成方法。首先,将用于计算机视觉图像生成领域的DDPM模型应用于风电机组故障诊断领域中,通过前向加噪过程将数据逐渐转化为噪声,再通过逆向去噪过程将噪声逐步恢复为原始数据,实现从噪声中生成故障数据,解决数据不平衡问题;其次,通过对原始DDPM中使用的U-net模块进行改进,使用Transformer模型替换U-net网络,利用扩散后的数据和添加的噪声训练Transformer模型,实现噪声预测,以提高故障数据的生成质量;最后,使用多种生成模型评价指标对生成的故障数据进行评价,在监督控制和数据采集系统(supervisory control and data acquisition,SCADA)故障数据生成中论证改进DDPM-Transformer模型的性能。通过试验证明,所提DDPM-Transformer模型与现有的生成模型相比,最大均值异(maximum mean discrepancy,MMD)最大提升0.13,峰值信噪比(peak signal to noise ratio,PSNR)最大提升7.8。所提模型可以有效地生成质量更高的风电机组故障样本,从而基于该样本集辅助训练基于深度学习的故障诊断模型,可以使诊断模型具有更高精度和良好的稳定性。展开更多
Aiming at node deployment in the monitoring area of the field observation instrument network in the cold and arid regions,we propose a virtual force algorithm based on Voronoi diagram(VFAVD),which adopts probabilistic...Aiming at node deployment in the monitoring area of the field observation instrument network in the cold and arid regions,we propose a virtual force algorithm based on Voronoi diagram(VFAVD),which adopts probabilistic sensing model that is more in line with the actual situation.First,the Voronoi diagram is constructed in the monitoring area to determine the Thiessen polygon of each node.Then,the virtual force on each node is calculated,and the node update its position according to the direction and size of the total force,so as to achieve the purpose of improving the network coverage rate.The simulation results show that the proposed algorithm can effectively improve the coverage rate of the network,and also has a good effect on the coverage uniformity.展开更多
针对在滚动轴承故障诊断领域中存在的故障样本较少,健康样本丰富所导致的故障类别失衡问题以及环境中存在噪声与人为噪声标签干扰等问题,提出了一种基于混合裁剪失衡数据增强与SwinNet网络相结合的故障诊断模型(fault diagnosis model c...针对在滚动轴承故障诊断领域中存在的故障样本较少,健康样本丰富所导致的故障类别失衡问题以及环境中存在噪声与人为噪声标签干扰等问题,提出了一种基于混合裁剪失衡数据增强与SwinNet网络相结合的故障诊断模型(fault diagnosis model combining mixed-cutout imbalance data augmentation and SwinNet,SwinNet-MCIDA)。首先,借鉴图像分类数据增强方法,利用混合裁剪失衡数据增强算法对失衡类别的数据进行裁剪、混合处理生成新的故障样本来增加样本量,构造出增强数据集,然后对增强数据集进行小波变换转换成时频图像,将所得图像输入到卷积神经网络与Swin Transformer编码器相结合的SwinNet网络模型中,进行特征提取和故障分类,从而实现滚动轴承故障的高效诊断。试验结果表明,该文所提出的SwinNet-MCIDA故障诊断方法不仅可以很好地解决滚动轴承故障诊断领域故障类别失衡问题,而且也可以很好地应对故障数据中存在环境噪声问题与人为噪声标签干扰问题。展开更多
随着全球环境问题不断恶化,人们越来越意识到非污染和可再生能源的重要性。风能作为一种分布广泛的洁净能源,在全世界范围内得到广泛开发和利用。我国是风机制造和生产大国,也是风力发电大国。风电机组常常面临机组工况多变和条件恶劣...随着全球环境问题不断恶化,人们越来越意识到非污染和可再生能源的重要性。风能作为一种分布广泛的洁净能源,在全世界范围内得到广泛开发和利用。我国是风机制造和生产大国,也是风力发电大国。风电机组常常面临机组工况多变和条件恶劣的工作环境,导致其在服役期间容易因故障而停机,带来安全隐患,影响风电场经济效益。本数据集为山西晋城泽州风电场10台风电机组运行及故障数据集,采用自动观测的方法,由数据采集与监视控制系统(Supervisory Control And Data Acquisition,SCADA)生成导出,时间范围为2021年5月19日到2022年5月18日,数据集包含了采样间隔10 min的SCADA数据和故障信息。本数据集时间范围长,收录较多风机故障数据,有效数据比例为97.92%,可为研究风机运行策略、风机故障诊断与预测等科学问题提供数据支持。展开更多
The imbalance of energy consumption in wireless sensor networks(WSNs)easily results in the“hot spot”problem that the sensor nodes in a particular area die due to fast energy consumption.In order to solve the“hot s...The imbalance of energy consumption in wireless sensor networks(WSNs)easily results in the“hot spot”problem that the sensor nodes in a particular area die due to fast energy consumption.In order to solve the“hot spot”problem in WSNs,we propose an unequal clustering routing algorithm based on genetic algorithm(UCR-GA).In the cluster head election phase,the fitness function is constructed based on the residual energy,density and distance between nodes and base station,and the appropriate node is selected as the cluster head.In the data transmission phase,the cluster head selects single-hop or multi-hop communication mode according to the distance to the base station.After we comprehensively consider the residual energy of the cluster head and its communication energy consumption with the base station,an appropriate relay node is selected.The designed protocal is simulated under energy homogeneous and energy heterogeneity conditions,and the results show that the proposed routing protocal can effectively balance energy consumption,prolong the life cycle of network,and is appicable to heterogeneous networks.展开更多
文摘为解决风电机组故障诊断中故障样本不足而导致模型准确率不高的问题,将当下备受关注的数据增强方法-去噪扩散概率模型(denoising diffusion probability model,DDPM)引入到故障诊断领域以生成大量高质量的故障样本数据集。因此,结合Transformer网络,提出了一种DDPM-Transformer风电机组故障样本生成方法。首先,将用于计算机视觉图像生成领域的DDPM模型应用于风电机组故障诊断领域中,通过前向加噪过程将数据逐渐转化为噪声,再通过逆向去噪过程将噪声逐步恢复为原始数据,实现从噪声中生成故障数据,解决数据不平衡问题;其次,通过对原始DDPM中使用的U-net模块进行改进,使用Transformer模型替换U-net网络,利用扩散后的数据和添加的噪声训练Transformer模型,实现噪声预测,以提高故障数据的生成质量;最后,使用多种生成模型评价指标对生成的故障数据进行评价,在监督控制和数据采集系统(supervisory control and data acquisition,SCADA)故障数据生成中论证改进DDPM-Transformer模型的性能。通过试验证明,所提DDPM-Transformer模型与现有的生成模型相比,最大均值异(maximum mean discrepancy,MMD)最大提升0.13,峰值信噪比(peak signal to noise ratio,PSNR)最大提升7.8。所提模型可以有效地生成质量更高的风电机组故障样本,从而基于该样本集辅助训练基于深度学习的故障诊断模型,可以使诊断模型具有更高精度和良好的稳定性。
基金supported by National Natural Science Foundation of China(No.61862038)Lanzhou Talent Innovation and Entrepreneurship Technology Plan Project(No.2019-RC-14).
文摘Aiming at node deployment in the monitoring area of the field observation instrument network in the cold and arid regions,we propose a virtual force algorithm based on Voronoi diagram(VFAVD),which adopts probabilistic sensing model that is more in line with the actual situation.First,the Voronoi diagram is constructed in the monitoring area to determine the Thiessen polygon of each node.Then,the virtual force on each node is calculated,and the node update its position according to the direction and size of the total force,so as to achieve the purpose of improving the network coverage rate.The simulation results show that the proposed algorithm can effectively improve the coverage rate of the network,and also has a good effect on the coverage uniformity.
文摘针对在滚动轴承故障诊断领域中存在的故障样本较少,健康样本丰富所导致的故障类别失衡问题以及环境中存在噪声与人为噪声标签干扰等问题,提出了一种基于混合裁剪失衡数据增强与SwinNet网络相结合的故障诊断模型(fault diagnosis model combining mixed-cutout imbalance data augmentation and SwinNet,SwinNet-MCIDA)。首先,借鉴图像分类数据增强方法,利用混合裁剪失衡数据增强算法对失衡类别的数据进行裁剪、混合处理生成新的故障样本来增加样本量,构造出增强数据集,然后对增强数据集进行小波变换转换成时频图像,将所得图像输入到卷积神经网络与Swin Transformer编码器相结合的SwinNet网络模型中,进行特征提取和故障分类,从而实现滚动轴承故障的高效诊断。试验结果表明,该文所提出的SwinNet-MCIDA故障诊断方法不仅可以很好地解决滚动轴承故障诊断领域故障类别失衡问题,而且也可以很好地应对故障数据中存在环境噪声问题与人为噪声标签干扰问题。
基金National Natural Science Foundation of China(No.61862038)Gansu Province Science and Technology Program(No.20JR10RA213)+1 种基金Gansu Province Science and Technology Program-Innovation Fund for Small and Medium-sized Enterprises(No.21CX6JA150)Foundation of a Hundred Youth Talents Training Program of Lanzhou Jiaotong University。
文摘随着全球环境问题不断恶化,人们越来越意识到非污染和可再生能源的重要性。风能作为一种分布广泛的洁净能源,在全世界范围内得到广泛开发和利用。我国是风机制造和生产大国,也是风力发电大国。风电机组常常面临机组工况多变和条件恶劣的工作环境,导致其在服役期间容易因故障而停机,带来安全隐患,影响风电场经济效益。本数据集为山西晋城泽州风电场10台风电机组运行及故障数据集,采用自动观测的方法,由数据采集与监视控制系统(Supervisory Control And Data Acquisition,SCADA)生成导出,时间范围为2021年5月19日到2022年5月18日,数据集包含了采样间隔10 min的SCADA数据和故障信息。本数据集时间范围长,收录较多风机故障数据,有效数据比例为97.92%,可为研究风机运行策略、风机故障诊断与预测等科学问题提供数据支持。
基金National Natural Science Foundation of China(No.61862038)Lanzhou Talent Innovation and Entrepreneurship Technology Plan Project(No.2019-RC-14)Foundation of a Hundred Youth Talents Training Program of Lanzhou Jiaotong University。
文摘The imbalance of energy consumption in wireless sensor networks(WSNs)easily results in the“hot spot”problem that the sensor nodes in a particular area die due to fast energy consumption.In order to solve the“hot spot”problem in WSNs,we propose an unequal clustering routing algorithm based on genetic algorithm(UCR-GA).In the cluster head election phase,the fitness function is constructed based on the residual energy,density and distance between nodes and base station,and the appropriate node is selected as the cluster head.In the data transmission phase,the cluster head selects single-hop or multi-hop communication mode according to the distance to the base station.After we comprehensively consider the residual energy of the cluster head and its communication energy consumption with the base station,an appropriate relay node is selected.The designed protocal is simulated under energy homogeneous and energy heterogeneity conditions,and the results show that the proposed routing protocal can effectively balance energy consumption,prolong the life cycle of network,and is appicable to heterogeneous networks.
