台风条件下海上风电场风速变化大、无明显周期性,这对海上风电场的风速预测造成了极大的困难。针对此问题,提出台风条件下海上风电场风速多步预测方法。首先,针对台风预报信息与风电场风速数据时间尺度不统一的问题,提出用嵌入层网络对...台风条件下海上风电场风速变化大、无明显周期性,这对海上风电场的风速预测造成了极大的困难。针对此问题,提出台风条件下海上风电场风速多步预测方法。首先,针对台风预报信息与风电场风速数据时间尺度不统一的问题,提出用嵌入层网络对台风预报信息进行动态插值。其次,基于Holland气压场模型和Batts梯度风模型构建融合物理信息的神经网络,将Holland模型和Batts模型中的经验参数替换成网络可学习的参数,并针对网络训练过程中可能出现的数值问题引入适当的近似方法。最后,对含时序模式注意力机制的长短期记忆网络(temporal pattern attention long short-term memory,TPA-LSTM)进行改进,嵌入融合物理信息的神经网络,利用近40年台风期间的数据进行训练和测试。结果表明,在引入较少参数的情况下,物理信息神经网络能减少TPA-LSTM网络的训练迭代次数以及提高预测精度,所提模型相比序列到序列(sequence to sequence,Seq2Seq)模型和TPA-LSTM网络具有更高的预测精度。展开更多
This paper proposedmethod that combined transmission path analysis(TPA)and empirical mode decomposition(EMD)envelope analysis to solve the vibration problemof an industrial robot.Firstly,the deconvolution filter timed...This paper proposedmethod that combined transmission path analysis(TPA)and empirical mode decomposition(EMD)envelope analysis to solve the vibration problemof an industrial robot.Firstly,the deconvolution filter timedomain TPA method is proposed to trace the source along with the time variation.Secondly,the TPA method positioned themain source of robotic vibration under typically different working conditions.Thirdly,independent vibration testing of the Rotate Vector(RV)reducer is conducted under different loads and speeds,which are key components of an industrial robot.The method of EMD and Hilbert envelope was used to extract the fault feature of the RV reducer.Finally,the structural problems of the RV reducer were summarized.The vibration performance of industrial robots was improved through the RV reducer optimization.From the whole industrial robot to the local RV Reducer and then to the internal microstructure of the reducer,the source of defect information is traced accurately.Experimental results showed that the TPA and EMD hybrid methods were more accurate and efficient than traditional time-frequency analysis methods to solve industrial robot vibration problems.展开更多
Since the existing prediction methods have encountered difficulties in processing themultiple influencing factors in short-term power load forecasting,we propose a bidirectional long short-term memory(BiLSTM)neural ne...Since the existing prediction methods have encountered difficulties in processing themultiple influencing factors in short-term power load forecasting,we propose a bidirectional long short-term memory(BiLSTM)neural network model based on the temporal pattern attention(TPA)mechanism.Firstly,based on the grey relational analysis,datasets similar to forecast day are obtained.Secondly,thebidirectional LSTM layermodels the data of thehistorical load,temperature,humidity,and date-type and extracts complex relationships between data from the hidden row vectors obtained by the BiLSTM network,so that the influencing factors(with different characteristics)can select relevant information from different time steps to reduce the prediction error of the model.Simultaneously,the complex and nonlinear dependencies between time steps and sequences are extracted by the TPA mechanism,so the attention weight vector is constructed for the hidden layer output of BiLSTM and the relevant variables at different time steps are weighted to influence the input.Finally,the chaotic sparrow search algorithm(CSSA)is used to optimize the hyperparameter selection of the model.The short-term power load forecasting on different data sets shows that the average absolute errors of short-termpower load forecasting based on our method are 0.876 and 4.238,respectively,which is lower than other forecastingmethods,demonstrating the accuracy and stability of our model.展开更多
文摘台风条件下海上风电场风速变化大、无明显周期性,这对海上风电场的风速预测造成了极大的困难。针对此问题,提出台风条件下海上风电场风速多步预测方法。首先,针对台风预报信息与风电场风速数据时间尺度不统一的问题,提出用嵌入层网络对台风预报信息进行动态插值。其次,基于Holland气压场模型和Batts梯度风模型构建融合物理信息的神经网络,将Holland模型和Batts模型中的经验参数替换成网络可学习的参数,并针对网络训练过程中可能出现的数值问题引入适当的近似方法。最后,对含时序模式注意力机制的长短期记忆网络(temporal pattern attention long short-term memory,TPA-LSTM)进行改进,嵌入融合物理信息的神经网络,利用近40年台风期间的数据进行训练和测试。结果表明,在引入较少参数的情况下,物理信息神经网络能减少TPA-LSTM网络的训练迭代次数以及提高预测精度,所提模型相比序列到序列(sequence to sequence,Seq2Seq)模型和TPA-LSTM网络具有更高的预测精度。
基金supported by Natural Science Foundation of Hunan Province,(Grant No.2022JJ30147)the National Natural Science Foundation of China (Grant No.51805155)the Foundation for Innovative Research Groups of National Natural Science Foundation of China (Grant No.51621004).
文摘This paper proposedmethod that combined transmission path analysis(TPA)and empirical mode decomposition(EMD)envelope analysis to solve the vibration problemof an industrial robot.Firstly,the deconvolution filter timedomain TPA method is proposed to trace the source along with the time variation.Secondly,the TPA method positioned themain source of robotic vibration under typically different working conditions.Thirdly,independent vibration testing of the Rotate Vector(RV)reducer is conducted under different loads and speeds,which are key components of an industrial robot.The method of EMD and Hilbert envelope was used to extract the fault feature of the RV reducer.Finally,the structural problems of the RV reducer were summarized.The vibration performance of industrial robots was improved through the RV reducer optimization.From the whole industrial robot to the local RV Reducer and then to the internal microstructure of the reducer,the source of defect information is traced accurately.Experimental results showed that the TPA and EMD hybrid methods were more accurate and efficient than traditional time-frequency analysis methods to solve industrial robot vibration problems.
基金supported by the Major Project of Basic and Applied Research in Guangdong Universities (2017WZDXM012)。
文摘Since the existing prediction methods have encountered difficulties in processing themultiple influencing factors in short-term power load forecasting,we propose a bidirectional long short-term memory(BiLSTM)neural network model based on the temporal pattern attention(TPA)mechanism.Firstly,based on the grey relational analysis,datasets similar to forecast day are obtained.Secondly,thebidirectional LSTM layermodels the data of thehistorical load,temperature,humidity,and date-type and extracts complex relationships between data from the hidden row vectors obtained by the BiLSTM network,so that the influencing factors(with different characteristics)can select relevant information from different time steps to reduce the prediction error of the model.Simultaneously,the complex and nonlinear dependencies between time steps and sequences are extracted by the TPA mechanism,so the attention weight vector is constructed for the hidden layer output of BiLSTM and the relevant variables at different time steps are weighted to influence the input.Finally,the chaotic sparrow search algorithm(CSSA)is used to optimize the hyperparameter selection of the model.The short-term power load forecasting on different data sets shows that the average absolute errors of short-termpower load forecasting based on our method are 0.876 and 4.238,respectively,which is lower than other forecastingmethods,demonstrating the accuracy and stability of our model.