Recent works have shown that neural networks are promising parameter-free limiters for a variety of numerical schemes(Morgan et al.in A machine learning approach for detect-ing shocks with high-order hydrodynamic meth...Recent works have shown that neural networks are promising parameter-free limiters for a variety of numerical schemes(Morgan et al.in A machine learning approach for detect-ing shocks with high-order hydrodynamic methods.et al.in J Comput Phys 367:166-191.,2018;Veiga et al.in European Conference on Computational Mechanics andⅦEuropean Conference on Computational Fluid Dynamics,vol.1,pp.2525-2550.ECCM.,2018).Following this trend,we train a neural network to serve as a shock-indicator function using simulation data from a Runge-Kutta discontinuous Galer-kin(RKDG)method and a modal high-order limiter(Krivodonova in J Comput Phys 226:879-896.,2007).With this methodology,we obtain one-and two-dimensional black-box shock-indicators which are then coupled to a standard limiter.Furthermore,we describe a strategy to transfer the shock-indicator to a residual distribution(RD)scheme without the need for a full training cycle and large data-set,by finding a mapping between the solution feature spaces from an RD scheme to an RKDG scheme,both in one-and two-dimensional problems,and on Cartesian and unstruc-tured meshes.We report on the quality of the numerical solutions when using the neural network shock-indicator coupled to a limiter,comparing its performance to traditional lim-iters,for both RKDG and RD schemes.展开更多
针对动力电池荷电状态(state of charge,SOC)的估算问题,利用长短期记忆(LSTM)循环神经网络建立SOC估算模型,以实验室恒流放电数据训练模型并测试,测试最大绝对误差为2.7%。进一步以FSEC赛车电池实测数据验证,最大测试误差为3.9%。但在...针对动力电池荷电状态(state of charge,SOC)的估算问题,利用长短期记忆(LSTM)循环神经网络建立SOC估算模型,以实验室恒流放电数据训练模型并测试,测试最大绝对误差为2.7%。进一步以FSEC赛车电池实测数据验证,最大测试误差为3.9%。但在工程应用时,考虑到实际运行过程中的环境复杂性以及不同驾驶习惯对动力电池造成的不一致性,需要根据车辆实际行驶工况数据对其进行训练与测试,但是由于该数据中的SOC直接由BMS报文解析而来,无法确定BMS内的SOC算法是否准确,故不能直接用作训练模型时的标签,此时需计算出正确的训练标签或借助已有标签的模型,在其基础上根据实际运行数据对其模型参数进行动态调整。为解决无标签数据的训练问题,本文采取第二种方法,首次提出将迁移学习中的领域自适应网络(DaNN)与LSTM组合形成LSTM-DaNN的SOC估算算法,利用有标签数据预先训练好LSTM模型,再将其模型参数迁移至LSTM-DaNN,最后综合有标签与无标签数据一起对LSTM-DaNN模型进行训练。测试结果表明LSTM-DaNN可以在没有实际行驶工况标签(SOC)的情况下完成训练,最大测试误差为4.8%,相比模型自适应调整前误差下降了14.1%,且保证绝对误差<5%,满足实际需求。展开更多
Lately,the power demand of consumers is increasing in distribution networks,while renewable power generation keeps penetrating into the distribution networks.Insufficient data make it hard to accurately predict the ne...Lately,the power demand of consumers is increasing in distribution networks,while renewable power generation keeps penetrating into the distribution networks.Insufficient data make it hard to accurately predict the new residential load or newly built apartments with volatile and changing time-series characteristics in terms of frequency and magnitude.Hence,this paper proposes a short-term probabilistic residential load forecasting scheme based on transfer learning and deep learning techniques.First,we formulate the short-term probabilistic residential load forecasting problem.Then,we propose a sequence-to-sequence(Seq2Seq)adversarial domain adaptation network and its joint training strategy to transfer generic features from the source domain(with massive consumption records of regular loads)to the target domain(with limited observations of new residential loads)and simultaneously minimize the domain difference and forecasting errors when solving the forecasting problem.For implementation,the dominant techniques or elements are used as the submodules of the Seq2Seq adversarial domain adaptation network,including the Seq2Seq recurrent neural networks(RNNs)composed of a long short-term memory(LSTM)encoder and an LSTM decoder,and quantile loss.Finally,this study conducts the case studies via multiple evaluation indices,comparative methods of classic machine learning and advanced deep learning,and various available data of the new residentical loads and other regular loads.The experimental results validate the effectiveness and stability of the proposed scheme.展开更多
Detection and recognition of a stairway as upstairs,downstairs and negative(e.g.,ladder,level ground)are the fundamentals of assisting the visually impaired to travel independently in unfamiliar environments.Previous ...Detection and recognition of a stairway as upstairs,downstairs and negative(e.g.,ladder,level ground)are the fundamentals of assisting the visually impaired to travel independently in unfamiliar environments.Previous studies have focused on using massive amounts of RGB-D scene data to train traditional machine learning(ML)-based models to detect and recognize stationary stairway and escalator stairway separately.Nevertheless,none of them consider jointly training these two similar but different datasets to achieve better performance.This paper applies an adversarial learning algorithm on the indicated unsupervised domain adaptation scenario to transfer knowledge learned from the labeled RGB-D escalator stairway dataset to the unlabeled RGB-D stationary dataset.By utilizing the developed method,a feedforward convolutional neural network(CNN)-based feature extractor with five convolution layers can achieve 100%classification accuracy on testing the labeled escalator stairway data distributions and 80.6%classification accuracy on testing the unlabeled stationary data distributions.The success of the developed approach is demonstrated for classifying stairway on these two domains with a limited amount of data.To further demonstrate the effectiveness of the proposed method,the same CNN model is evaluated without domain adaptation and the results are compared with those of the presented architecture.展开更多
文摘Recent works have shown that neural networks are promising parameter-free limiters for a variety of numerical schemes(Morgan et al.in A machine learning approach for detect-ing shocks with high-order hydrodynamic methods.et al.in J Comput Phys 367:166-191.,2018;Veiga et al.in European Conference on Computational Mechanics andⅦEuropean Conference on Computational Fluid Dynamics,vol.1,pp.2525-2550.ECCM.,2018).Following this trend,we train a neural network to serve as a shock-indicator function using simulation data from a Runge-Kutta discontinuous Galer-kin(RKDG)method and a modal high-order limiter(Krivodonova in J Comput Phys 226:879-896.,2007).With this methodology,we obtain one-and two-dimensional black-box shock-indicators which are then coupled to a standard limiter.Furthermore,we describe a strategy to transfer the shock-indicator to a residual distribution(RD)scheme without the need for a full training cycle and large data-set,by finding a mapping between the solution feature spaces from an RD scheme to an RKDG scheme,both in one-and two-dimensional problems,and on Cartesian and unstruc-tured meshes.We report on the quality of the numerical solutions when using the neural network shock-indicator coupled to a limiter,comparing its performance to traditional lim-iters,for both RKDG and RD schemes.
文摘针对动力电池荷电状态(state of charge,SOC)的估算问题,利用长短期记忆(LSTM)循环神经网络建立SOC估算模型,以实验室恒流放电数据训练模型并测试,测试最大绝对误差为2.7%。进一步以FSEC赛车电池实测数据验证,最大测试误差为3.9%。但在工程应用时,考虑到实际运行过程中的环境复杂性以及不同驾驶习惯对动力电池造成的不一致性,需要根据车辆实际行驶工况数据对其进行训练与测试,但是由于该数据中的SOC直接由BMS报文解析而来,无法确定BMS内的SOC算法是否准确,故不能直接用作训练模型时的标签,此时需计算出正确的训练标签或借助已有标签的模型,在其基础上根据实际运行数据对其模型参数进行动态调整。为解决无标签数据的训练问题,本文采取第二种方法,首次提出将迁移学习中的领域自适应网络(DaNN)与LSTM组合形成LSTM-DaNN的SOC估算算法,利用有标签数据预先训练好LSTM模型,再将其模型参数迁移至LSTM-DaNN,最后综合有标签与无标签数据一起对LSTM-DaNN模型进行训练。测试结果表明LSTM-DaNN可以在没有实际行驶工况标签(SOC)的情况下完成训练,最大测试误差为4.8%,相比模型自适应调整前误差下降了14.1%,且保证绝对误差<5%,满足实际需求。
基金supported by the National Natural Science Foundation of China(No.52177087)。
文摘Lately,the power demand of consumers is increasing in distribution networks,while renewable power generation keeps penetrating into the distribution networks.Insufficient data make it hard to accurately predict the new residential load or newly built apartments with volatile and changing time-series characteristics in terms of frequency and magnitude.Hence,this paper proposes a short-term probabilistic residential load forecasting scheme based on transfer learning and deep learning techniques.First,we formulate the short-term probabilistic residential load forecasting problem.Then,we propose a sequence-to-sequence(Seq2Seq)adversarial domain adaptation network and its joint training strategy to transfer generic features from the source domain(with massive consumption records of regular loads)to the target domain(with limited observations of new residential loads)and simultaneously minimize the domain difference and forecasting errors when solving the forecasting problem.For implementation,the dominant techniques or elements are used as the submodules of the Seq2Seq adversarial domain adaptation network,including the Seq2Seq recurrent neural networks(RNNs)composed of a long short-term memory(LSTM)encoder and an LSTM decoder,and quantile loss.Finally,this study conducts the case studies via multiple evaluation indices,comparative methods of classic machine learning and advanced deep learning,and various available data of the new residentical loads and other regular loads.The experimental results validate the effectiveness and stability of the proposed scheme.
文摘Detection and recognition of a stairway as upstairs,downstairs and negative(e.g.,ladder,level ground)are the fundamentals of assisting the visually impaired to travel independently in unfamiliar environments.Previous studies have focused on using massive amounts of RGB-D scene data to train traditional machine learning(ML)-based models to detect and recognize stationary stairway and escalator stairway separately.Nevertheless,none of them consider jointly training these two similar but different datasets to achieve better performance.This paper applies an adversarial learning algorithm on the indicated unsupervised domain adaptation scenario to transfer knowledge learned from the labeled RGB-D escalator stairway dataset to the unlabeled RGB-D stationary dataset.By utilizing the developed method,a feedforward convolutional neural network(CNN)-based feature extractor with five convolution layers can achieve 100%classification accuracy on testing the labeled escalator stairway data distributions and 80.6%classification accuracy on testing the unlabeled stationary data distributions.The success of the developed approach is demonstrated for classifying stairway on these two domains with a limited amount of data.To further demonstrate the effectiveness of the proposed method,the same CNN model is evaluated without domain adaptation and the results are compared with those of the presented architecture.
