An opportunistic maintenance model is presented for a continuously deteriorating series system with economical de-pendence. The system consists of two kinds of units, which are respectively subjected to the deteriorat...An opportunistic maintenance model is presented for a continuously deteriorating series system with economical de-pendence. The system consists of two kinds of units, which are respectively subjected to the deterioration failure described by Gamma process and the random failure described by Poisson process. A two-level opportunistic policy defined by three decision parameters is proposed to coordinate the different maintenance actions and minimize the long-run maintenance cost rate of the system. A computable expression of the average cost rate is established by using the renewal property of the stochastic process of the maintained system state. The optimal values of three deci- sion parameters are derived by an iteration approach based on the characteristic of Gamma process. The behavior of the proposed policy is illustrated through a numerical experiment. Comparative study with the widely used corrective maintenance policy demonstrates the advantage of the proposed opportunistic maintenance method in significantly reducing the maintenance cost. Simultane- ously, the applicable area of this opportunistic model is discussed by the sensitivity analysis of the set-up cost and random failure rate.展开更多
为了解决无人机航拍图片玉米植株中心检测所面临的诸多挑战,包括植株遮挡、形态多样、光照变化以及视觉混淆等问题,提升检测精度和模型的鲁棒性,开发了一种基于YOLO-TSCAS(YOLO with triplet-attention,saliencyadaptive,and centroid a...为了解决无人机航拍图片玉米植株中心检测所面临的诸多挑战,包括植株遮挡、形态多样、光照变化以及视觉混淆等问题,提升检测精度和模型的鲁棒性,开发了一种基于YOLO-TSCAS(YOLO with triplet-attention,saliencyadaptive,and centroid awareness for scenes)模型的玉米植株中心检测算法。该算法通过三重注意力模块、显著性裁剪混合数据增强方法、自适应池化技术和选择性多单元激活函数等技术手段,有效提高了检测精度和鲁棒性。采用三重注意力模块解决目标遮挡和视觉混淆问题,使模型能够更好地关注植株中心区域。采用显著性裁剪混合数据增强方法,在训练过程中引入不同的环境和光照变化,增强了模型对复杂场景的适应能力。结合自适应池化技术提高空间分辨率,有助于捕捉更精细的特征信息,提高检测的准确性。采用选择性多单元激活函数进一步增强了网络的学习能力,使模型能够更好地适应各种场景下的植株中心检测任务。实验结果表明,与现有的YOLOX算法相比,YOLO-TSCAS算法在平均准确率和平均F1值上分别提高了22.73个百分点和0.255,并且平均对数漏检率也显著降低了0.35。与其他流行的检测模型相比,在两类植株中心目标检测精度上也取得了最佳效果。展开更多
为了提高利用监控和数据采集(supervisory control and data acquisition,SCADA)多变量长时间序列预测齿轮箱油温的精度,解决不同风电机组因处不同运行环境导致的数据分布不一致的问题,提出了一种基于多分支时间序列预测与迁移学习相结...为了提高利用监控和数据采集(supervisory control and data acquisition,SCADA)多变量长时间序列预测齿轮箱油温的精度,解决不同风电机组因处不同运行环境导致的数据分布不一致的问题,提出了一种基于多分支时间序列预测与迁移学习相结合的齿轮箱状态监测方法。首先,利用极致梯度提升(extreme gradient boosting,XGBoost)算法筛选输入参数组成原始序列,对其进行分解得到季节与趋势序列。其次,提出季节、趋势序列特征提取模块获取季节及趋势特征的序列,将其与经过Informer模型处理后的特征序列进行融合后输入进多层感知机映射成最终的预测值,以构建提出的多分支时间序列预测网络(multi-branch time series prediction network,MBFN)。最后,利用迁移学习并结合一分类向量支持机(one-class support vector machine,OCSVM)模型及滑动窗口构建齿轮箱的健康指数,完成齿轮箱状态监测。实验结果表明,所提出模型的MBFN显著提高了油温预测精度,优于常规时间序列预测模型,所使用的迁移策略能以较少数据适应不同数据的分布,进而实现对齿轮箱的状态监测,并且所提出的模型可以提前18.9 d发出齿轮箱故障预警。展开更多
基金supported by the National Natural Science Foundation of China(6090400271201166)
文摘An opportunistic maintenance model is presented for a continuously deteriorating series system with economical de-pendence. The system consists of two kinds of units, which are respectively subjected to the deterioration failure described by Gamma process and the random failure described by Poisson process. A two-level opportunistic policy defined by three decision parameters is proposed to coordinate the different maintenance actions and minimize the long-run maintenance cost rate of the system. A computable expression of the average cost rate is established by using the renewal property of the stochastic process of the maintained system state. The optimal values of three deci- sion parameters are derived by an iteration approach based on the characteristic of Gamma process. The behavior of the proposed policy is illustrated through a numerical experiment. Comparative study with the widely used corrective maintenance policy demonstrates the advantage of the proposed opportunistic maintenance method in significantly reducing the maintenance cost. Simultane- ously, the applicable area of this opportunistic model is discussed by the sensitivity analysis of the set-up cost and random failure rate.
文摘为了解决无人机航拍图片玉米植株中心检测所面临的诸多挑战,包括植株遮挡、形态多样、光照变化以及视觉混淆等问题,提升检测精度和模型的鲁棒性,开发了一种基于YOLO-TSCAS(YOLO with triplet-attention,saliencyadaptive,and centroid awareness for scenes)模型的玉米植株中心检测算法。该算法通过三重注意力模块、显著性裁剪混合数据增强方法、自适应池化技术和选择性多单元激活函数等技术手段,有效提高了检测精度和鲁棒性。采用三重注意力模块解决目标遮挡和视觉混淆问题,使模型能够更好地关注植株中心区域。采用显著性裁剪混合数据增强方法,在训练过程中引入不同的环境和光照变化,增强了模型对复杂场景的适应能力。结合自适应池化技术提高空间分辨率,有助于捕捉更精细的特征信息,提高检测的准确性。采用选择性多单元激活函数进一步增强了网络的学习能力,使模型能够更好地适应各种场景下的植株中心检测任务。实验结果表明,与现有的YOLOX算法相比,YOLO-TSCAS算法在平均准确率和平均F1值上分别提高了22.73个百分点和0.255,并且平均对数漏检率也显著降低了0.35。与其他流行的检测模型相比,在两类植株中心目标检测精度上也取得了最佳效果。
文摘为了提高利用监控和数据采集(supervisory control and data acquisition,SCADA)多变量长时间序列预测齿轮箱油温的精度,解决不同风电机组因处不同运行环境导致的数据分布不一致的问题,提出了一种基于多分支时间序列预测与迁移学习相结合的齿轮箱状态监测方法。首先,利用极致梯度提升(extreme gradient boosting,XGBoost)算法筛选输入参数组成原始序列,对其进行分解得到季节与趋势序列。其次,提出季节、趋势序列特征提取模块获取季节及趋势特征的序列,将其与经过Informer模型处理后的特征序列进行融合后输入进多层感知机映射成最终的预测值,以构建提出的多分支时间序列预测网络(multi-branch time series prediction network,MBFN)。最后,利用迁移学习并结合一分类向量支持机(one-class support vector machine,OCSVM)模型及滑动窗口构建齿轮箱的健康指数,完成齿轮箱状态监测。实验结果表明,所提出模型的MBFN显著提高了油温预测精度,优于常规时间序列预测模型,所使用的迁移策略能以较少数据适应不同数据的分布,进而实现对齿轮箱的状态监测,并且所提出的模型可以提前18.9 d发出齿轮箱故障预警。
