Radar Quantitative Precipitation Estimation Based on the Gated Recurrent Unit Neural Network and Echo-Top Data

The Gated Recurrent Unit(GRU) neural network has great potential in estimating and predicting a variable. In addition to radar reflectivity(Z), radar echo-top height(ET) is also a good indicator of rainfall rate(R). In this study, we propose a new method, GRU_Z-ET, by introducing Z and ET as two independent variables into the GRU neural network to conduct the quantitative single-polarization radar precipitation estimation. The performance of GRU_Z-ET is compared with that of the other three methods in three heavy rainfall cases in China during 2018, namely, the traditional Z-R relationship(Z=300R1.4), the optimal Z-R relationship(Z=79R1.68) and the GRU neural network with only Z as the independent input variable(GRU_Z). The results indicate that the GRU_Z-ET performs the best, while the traditional Z-R relationship performs the worst. The performances of the rest two methods are similar.To further evaluate the performance of the GRU_Z-ET, 200 rainfall events with 21882 total samples during May–July of 2018 are used for statistical analysis. Results demonstrate that the spatial correlation coefficients, threat scores and probability of detection between the observed and estimated precipitation are the largest for the GRU_Z-ET and the smallest for the traditional Z-R relationship, and the root mean square error is just the opposite. In addition, these statistics of GRU_Z are similar to those of optimal Z-R relationship. Thus, it can be concluded that the performance of the GRU_ZET is the best in the four methods for the quantitative precipitation estimation.

A HybridManufacturing ProcessMonitoringMethod Using Stacked Gated Recurrent Unit and Random Forest

This study proposed a new real-time manufacturing process monitoring method to monitor and detect process shifts in manufacturing operations.Since real-time production process monitoring is critical in today’s smart manufacturing.The more robust the monitoring model,the more reliable a process is to be under control.In the past,many researchers have developed real-time monitoring methods to detect process shifts early.However,thesemethods have limitations in detecting process shifts as quickly as possible and handling various data volumes and varieties.In this paper,a robust monitoring model combining Gated Recurrent Unit(GRU)and Random Forest(RF)with Real-Time Contrast(RTC)called GRU-RF-RTC was proposed to detect process shifts rapidly.The effectiveness of the proposed GRU-RF-RTC model is first evaluated using multivariate normal and nonnormal distribution datasets.Then,to prove the applicability of the proposed model in a realmanufacturing setting,the model was evaluated using real-world normal and non-normal problems.The results demonstrate that the proposed GRU-RF-RTC outperforms other methods in detecting process shifts quickly with the lowest average out-of-control run length(ARL1)in all synthesis and real-world problems under normal and non-normal cases.The experiment results on real-world problems highlight the significance of the proposed GRU-RF-RTC model in modern manufacturing process monitoring applications.The result reveals that the proposed method improves the shift detection capability by 42.14%in normal and 43.64%in gamma distribution problems.

Real-time analysis and prediction of shield cutterhead torque using optimized gated recurrent unit neural network

An accurate prediction of earth pressure balance(EPB)shield moving performance is important to ensure the safety tunnel excavation.A hybrid model is developed based on the particle swarm optimization(PSO)and gated recurrent unit(GRU)neural network.PSO is utilized to assign the optimal hyperparameters of GRU neural network.There are mainly four steps:data collection and processing,hybrid model establishment,model performance evaluation and correlation analysis.The developed model provides an alternative to tackle with time-series data of tunnel project.Apart from that,a novel framework about model application is performed to provide guidelines in practice.A tunnel project is utilized to evaluate the performance of proposed hybrid model.Results indicate that geological and construction variables are significant to the model performance.Correlation analysis shows that construction variables(main thrust and foam liquid volume)display the highest correlation with the cutterhead torque(CHT).This work provides a feasible and applicable alternative way to estimate the performance of shield tunneling.

Speech Separation Algorithm Using Gated Recurrent Network Based on Microphone Array

Speech separation is an active research topic that plays an important role in numerous applications,such as speaker recognition,hearing pros-thesis,and autonomous robots.Many algorithms have been put forward to improve separation performance.However,speech separation in reverberant noisy environment is still a challenging task.To address this,a novel speech separation algorithm using gate recurrent unit(GRU)network based on microphone array has been proposed in this paper.The main aim of the proposed algorithm is to improve the separation performance and reduce the computational cost.The proposed algorithm extracts the sub-band steered response power-phase transform(SRP-PHAT)weighted by gammatone filter as the speech separation feature due to its discriminative and robust spatial position in formation.Since the GRU net work has the advantage of processing time series data with faster training speed and fewer training parameters,the GRU model is adopted to process the separation featuresof several sequential frames in the same sub-band to estimate the ideal Ratio Masking(IRM).The proposed algorithm decomposes the mixture signals into time-frequency(TF)units using gammatone filter bank in the frequency domain,and the target speech is reconstructed in the frequency domain by masking the mixture signal according to the estimated IRM.The operations of decomposing the mixture signal and reconstructing the target signal are completed in the frequency domain which can reduce the total computational cost.Experimental results demonstrate that the proposed algorithm realizes omnidirectional speech sep-aration in noisy and reverberant environments,provides good performance in terms of speech quality and intelligibility,and has the generalization capacity to reverberate.

Ensembling Neural Networks for User’s Indoor Localization Using Magnetic Field Data from Smartphones

Predominantly the localization accuracy of the magnetic field-based localization approaches is severed by two limiting factors:Smartphone heterogeneity and smaller data lengths.The use of multifarioussmartphones cripples the performance of such approaches owing to the variability of the magnetic field data.In the same vein,smaller lengths of magnetic field data decrease the localization accuracy substantially.The current study proposes the use of multiple neural networks like deep neural network(DNN),long short term memory network(LSTM),and gated recurrent unit network(GRN)to perform indoor localization based on the embedded magnetic sensor of the smartphone.A voting scheme is introduced that takes predictions from neural networks into consideration to estimate the current location of the user.Contrary to conventional magnetic field-based localization approaches that rely on the magnetic field data intensity,this study utilizes the normalized magnetic field data for this purpose.Training of neural networks is carried out using Galaxy S8 data while the testing is performed with three devices,i.e.,LG G7,Galaxy S8,and LG Q6.Experiments are performed during different times of the day to analyze the impact of time variability.Results indicate that the proposed approach minimizes the impact of smartphone variability and elevates the localization accuracy.Performance comparison with three approaches reveals that the proposed approach outperforms them in mean,50%,and 75%error even using a lesser amount of magnetic field data than those of other approaches.

Short-Term Traffic Flow Prediction Based on Road Network Topology

Accurate short-term traffic flow prediction plays a crucial role in intelligent transportation system (ITS), because it can assist both traffic authorities and individual travelers make better decisions. Previous researches mostly focus on shallow traffic prediction models, which performances were unsatisfying since short-term traffic flow exhibits the characteristics of high nonlinearity, complexity and chaos. Taking the spatial and temporal correlations into consideration, a new traffic flow prediction method is proposed with the basis on the road network topology and gated recurrent unit (GRU). This method can help researchers without professional traffic knowledge extracting generic traffic flow features effectively and efficiently. Experiments are conducted by using real traffic flow data collected from the Caltrans Performance Measurement System (PEMS) database in San Diego and Oakland from June 15, 2017 to September 27, 2017. The results demonstrate that our method outperforms other traditional approaches in terms of mean absolute percentage error (MAPE), symmetric mean absolute percentage error (SMAPE) and root mean square error (RMSE).

融合CNN-BiGRU和注意力机制的网络入侵检测模型

为提高网络入侵检测模型特征提取能力和分类准确率,提出了一种融合双向门控循环单元(CNN-BiGRU)和注意力机制的网络入侵检测模型.使用CNN有效提取流量数据集中的非线性特征;双向门控循环单元(BiGRU)提取数据集中的时序特征,最后融合注意力机制对不同类型流量数据通过加权的方式进行重要程度的区分,从而整体提高该模型特征提取与分类的性能.实验结果表明:其整体精确率比双向长短期记忆网络(BiLSTM)模型提升了2.25%.K折交叉验证结果表明:该模型泛化性能良好,避免了过拟合现象的发生,印证了该模型的有效性与合理性.

基于KPCA-CNN-DBiGRU模型的短期负荷预测方法

本文针对已有神经网络模型在短期负荷预测中输入维度过高、预测误差较大等问题,提出了一种结合核主成分分析、卷积神经网络和深度双向门控循环单元的短期负荷预测方法。先运用核主成分分析法对原始高维输入变量进行降维,再通过卷积深度双向门控循环单元网络模型进行负荷预测。以第九届全国电工数学建模竞赛试题A题中的负荷数据作为实际算例,结果表明所提方法较降维之前预测误差大大降低,与已有预测方法相比也有大幅的误差降低。

基于CEEMD-SE的CNN&LSTM-GRU短期风电功率预测

为进一步提升短期风电功率的预测精度,提出了一种基于互补集合经验模态分解-样本熵(complementary ensemble empirical mode decomposition-sample entropy,CEEMD-SE)的卷积神经网络(convolutional neural network,CNN)和长短期记忆-门控循环单元(longshorttermmemory-gatedrecurrentunit,LSTM-GRU)的短期风电功率预测模型。首先,利用互补集合经验模态分解将原始风电功率序列分解为若干本征模态函数(intrinsic mode function,IMF)分量和一个残差(residual,RES)分量,利用样本熵算法将相近的分量进行重构;其次,搭建卷积神经网络和长短期记忆网络的并行网络结构,提取数据的局部特征和时序特征,并将特征融合后输入门控循环单元网络中进行学习预测;最后,通过算例进行验证,结果表明采用该模型后预测精度得到了有效提升,其均方根误差降低了15.06%、平均绝对误差降低了15.22%、决定系数提高了1.91%。

基于双层注意力和深度自编码器的时间序列异常检测模型

目前时间序列通常具有弱周期性以及高度复杂的相关性特征,传统的时间序列异常检测方法难以检测此类异常。针对这一问题,提出了一种新的无监督时间序列异常检测模型(DA-CBG-AE)。首先,使用新型滑动窗口方法,针对时间序列周期性设置滑动窗口大小;其次,采用卷积神经网络提取时间序列高维度空间特征;然后,提出具有堆叠式Dropout双向门循环单元网络作为自编码器的基本结构,从而捕捉时间序列的相关性特征;最后,引入双层注意力机制,进一步提取特征,选择更加关键的时间序列,从而提高异常检测准确率。为了验证该模型的有效性,将DA-CBG-AE与6种基准模型在8个数据集上进行比较。最终的实验结果表明,DA-CBG-AE获得了最优的F1值(0.863),并且其检测性能相比最新的基准模型Tad-GAN高出25.25%。

基于CNN-GRU-ISSA-XGBoost的短期光伏功率预测

针对光伏功率随机性及波动性大,单一预测模型往往难以准确分析历史数据波动规律,从而导致预测精度不高的问题,提出一种基于卷积神经网络-门控循环单元(CNN-GRU)和改进麻雀搜索算法(ISSA)优化的极限梯度提升(XGBoost)模型的短期光伏功率预测组合模型.首先去除历史数据中的异常值并对其进行归一化处理,利用主成分分析法(PCA)进行特征选取,以便更好地识别影响光伏功率的关键因素.然后采用CNN网络提取数据的空间特征,再经过GRU网络提取时间特征,针对XGBoost模型手动配置参数困难、随机性大的问题,利用ISSA对模型超参数寻优.最后对两种方法预测的结果用误差倒数法减小误差的同时对权重进行更新,得到新的预测值,从而完成对光伏功率的预测.实验结果表明,所提出的CNN-GRU-ISSA-XGBoost组合模型具有更强的适应性和更高的精度.

基于特征选择及ISSA-CNN-BiGRU的短期风功率预测

针对风电功率随机性大、平稳性低,以及直接输入预测模型往往难以取得较高精度等问题,提出了一种基于特征选择及改进麻雀搜索算法(ISSA)优化卷积神经网络-双向门控循环单元(CNN-BiGRU)的短期风电功率预测方法。首先,利用变分模态分解(VMD)将原始功率分解为一组包含不同信息的子分量,以降低原始功率序列的非平稳性,提升可预测性,同时通过观察中心频率方式确定模态分解数。其次,对每一分量采用随机森林(RF)特征重要度的方法进行特征选择,从风速、风向、温度、空气密度等气象特征因素中,选取对各个分量预测贡献度较高的影响因素组成输入特征向量。然后,建立各分量的CNN-BiGRU预测模型,针对神经网络算法参数难调、手动配置参数随机性大的问题,利用ISSA对模型超参数寻优,自适应搜寻最优参数组合。最后,叠加各分量的预测值,得到最终的预测结果。以中国内蒙古某风电场实际数据进行仿真实验,与多种单一及组合预测方法进行对比,结果表明,本文所提方法相比于其他方法具有更高的预测精度,其平均绝对百分比误差值达到2.644 0%;在其他4个数据集上进行的模型准确性及泛化性验证结果显示,模型平均绝对百分比误差值分别为4.385 3%、3.174 9%、1.576 1%和1.358 8%,均保持在5.000 0%以内,证明本文所提方法具有较好的预测精度及泛化能力。

基于动态工况实测数据图像和深度学习的锂电池容量估计方法

针对实际应用中基于动态工况下电池状态参数的片段数据进行电池健康状态(state of health,SOH)实时估计的问题,提出基于动态工况下锂离子电池状态参数(电压、电流、温度)实测数据二维特征图像和深度学习的锂离子电池容量估计算法。首先,将动态工况下电池状态参数监测量(电压、电流和温度)的片段数据转化为二维特征图像。其次,提出基于残差卷积神经网络(residual convolutional neural network,Res-CNN)和门控循环单元(gate recurrent unit,GRU)网络结合的多通道深度学习模型Res-CNN-GRU,以构建动态工况下电池状态参数特征图像和SOH之间的复杂非线性关系,其中电压、电流和温度的二维特征图像以三通道的方式输入到Res-CNN-GRU模型中,模型输出为对应电池的相邻参考充放电循环实验所获得容量的差值。研究结果表明:此方法在锂电池随机充放电工况下对电池健康状态估计效果更佳,且Res-CNN-GRU模型的泛化性和全局特征提取能力较强。论文研究为现实工况下电池健康状态估计的进一步深入研究提供了参考。

基于VMD-TCN-GRU模型的水质预测研究

为充分挖掘水质数据在短时震荡中的变化特征,提升预测模型的精度,提出一种基于VMD(变分模态分解)、TCN(卷积时间神经网络)及GRU(门控循环单元)组成的混合水质预测模型,采用VMD-TCN-GRU模型对汾河水库出水口高锰酸盐指数进行预测,并与此类研究中常见的SVR(支持向量回归)、LSTM(长短期记忆神经网络)、TCN和CNN-LSTM(卷积神经网络-长短期记忆神经网络)这4种模型预测结果对比表明:VMD-TCN-GRU模型能更好挖掘水质数据在短时震荡过程中的特征信息,提升水质预测精度;VMD-TCN-GRU模型的MAE(平均绝对误差)、RMSE(均方根误差)下降,R^(2)(确定系数)提高,其MAE、RMSE、R^(2)分别为0.0553、0.0717、0.9351;其预测性能优越,预测精度更高且拥有更强的泛化能力,可以应用于汾河水质预测。

基于深度学习的盾构机土舱压力场预测方法

土舱压力是盾构机受力状态和掌子面稳定等核心问题中的关键因素。土舱压力具有显著的空间变异性,其形成演化机制源于装备与岩土之间的复杂耦合作用,与地质特征、掘进参数等多源参数相关。然而,现有土舱压力预测方法一般未考虑空间分布特征或地质参数影响。针对该问题,提出了一种基于空间分布物理特征函数导引深度学习的盾构机土舱压力场预测方法。该方法构建物理特征函数用于解耦土舱压力空间分布特征,采用卷积神经网络和门控循环单元分别提取多源参数历史信息的空间特征和特征系数的时序特征,结合多源参数实时信息对特征系数进行预测,从而实现土舱压力场的预测。以长沙地铁四号线某区段为案例,利用该方法准确预测了土舱压力空间分布实测数据,准确率高达0.98,验证了所提方法的有效性。敏感性分析表明,不同地层中土舱压力空间分布特征系数的主要敏感参数基本一致,但其敏感度随地层地质条件的变化规律差异显著,可为复杂地层盾构机土舱压力精细化调控提供参考。

Deep Learning Based Channel Estimation in Fog Radio Access Networks

As a promising paradigm of the fifth generation networks,fog radio access network(F-RAN)has attracted lots of attention nowadays.To fully utilize the promising gain of F-RANs,the acquisition of accurate channel state information is significant.However,conventional channel estimation approaches are not suitable in F-RANs due to the large training and feedback overhead.In this paper,we consider the channel estimation in F-RANs with fog access point(F-AP)equipped with massive antennas.Thanks to the computing ability of F-AP and the sparsity of channel matrices in angular domain,Gated Recurrent Unit(GRU),a data-driven based channel estimation is proposed at F-AP to reduce the training and feedback overhead.The GRU-based method can capture the hidden sparsity structure automatically through the network training.Moreover,to further improve the channel estimation,a bidirectional GRU based method is proposed,whose target channel structure is decided by previous and subsequent structures.We compare the performance of our proposed channel estimation with traditional methods(Orthogonal Matching Pursuit(OMP)and Simultaneous OMP(SOMP)).Simulation results show that the proposed approaches have better performance compared with the traditional OMP and SOMP methods.

基于门控循环单元网络的钻井井漏智能监测方法

井漏是钻井过程中常见的钻井风险,若对该风险发现、处理不及时,极易导致井塌事故,轻则延长施工周期,重则危害现场人员人身安全。为了提高油气井钻井过程中井漏风险识别的准确性,降低风险识别对人为经验的依赖,结合钻井参数的非线性以及长时依赖特征,提出了一种基于门控循环单元(Gated Recurrent Unit,GRU)网络的井漏风险智能识别方法。该模型以池体积、出口流量和立管压力作为监测参数构建GRU网络,能够提取监测参数的时间序列特征,以实现对井漏风险的准确识别。利用现场实测钻井数据对模型进行了实验测试,结果表明,该方法对井漏风险的识别准确率达到了90.1%,优于长短期记忆网络的识别结果。

基于时序生成对抗网络的居民用户非侵入式负荷分解

现有的非侵入式负荷分解算法往往需要大量电器设备级的负荷数据才能保证分解精度,但由于用户对隐私性的考虑以及安装成本过高等问题,很难获取这些数据。因此,构建一种能深度挖掘电力负荷数据时序特性和电器相关性的时序生成对抗网络。利用降维网络对所有电器有功功率序列组成的高维向量进行降维以降低计算的复杂度,通过复原网络将结果还原为高维向量。基于电器运行状态和深度学习的非侵入式分解方法,运用卷积神经网络-双向门控循环单元构建状态复杂电器的负荷分解回归模型,对状态简单电器利用深度神经网络构建负荷识别分类模型。通过对比其他数据生成方法,以及改变典型公开数据集中生成数据比例所得的负荷分解结果验证了所提方法的有效性。

一种基于改进门控循环单元的叠前时变子波提取方法

子波的精确提取是地震勘探后续反演与成像的前提,针对传统时变子波提取方法受到的各类假设限制,且需分别提取子波振幅谱与相位谱的问题,本文提出了一种基于改进门控循环单元(GRU)网络的叠前时变地震子波提取方法.根据实际叠前地震数据分布特征与非平稳性质,本方法首先建立非平稳地震记录与添加随机噪声的时变子波训练数据集;为对提取出的时序特征进行拓展,提升传统GRU网络对长时序列的处理能力,本方法搭建起含多层GRU模块与全连接神经网络的改进门控循环单元网络模型;利用建立的训练数据集对网络模型进行训练使网络具备提取时变子波的能力;为提高训练效率与提取精度,本方法在训练的反向传播过程中应用自定义WaveLoss损失函数衡量误差,最终实现叠前时变子波的估计.经合成数据仿真实验与不同方法对比验证,本文提出的叠前时变子波提取方法具有更高的准确度;经对中国西部不同地区实际叠前地震资料处理与反褶积验证分析,该方法可有效提高目标区叠前地震剖面分辨率.

基于CNN-BiGRU-Attention的短期电力负荷预测

针对目前电力负荷数据随机性强,影响因素复杂,传统单一预测模型精度低的问题,结合卷积神经网络(Convolutional neural network,CNN)、双向门控循环单元(Bi-directional gated recurrent unit,BiGRU)以及注意力机制(Attention)在短期电力负荷预测上的不同优点,提出一种基于CNN-BiGRU-Attention的混合预测模型。该方法首先通过CNN对历史负荷和气象数据进行初步特征提取,然后利用BiGRU进一步挖掘特征数据间时序关联,再引入注意力机制,对BiGRU输出状态给与不同权重,强化关键特征,最后完成负荷预测。试验结果表明,该模型的平均绝对百分比误差(Mean absolute percentage error,MAPE)、均方根误差(Root mean square error,RMSE)、判定系数(R-square,R~2)分别为0.167%、0.057%、0.993,三项指标明显优于其他模型,具有更高的预测精度和稳定性,验证了模型在短期负荷预测中的优势。