Hierarchical multihead self-attention for time-series-based fault diagnosis

Fault diagnosis is important for maintaining the safety and effectiveness of chemical process.Considering the multivariate,nonlinear,and dynamic characteristic of chemical process,many time-series-based data-driven fault diagnosis methods have been developed in recent years.However,the existing methods have the problem of long-term dependency and are difficult to train due to the sequential way of training.To overcome these problems,a novel fault diagnosis method based on time-series and the hierarchical multihead self-attention(HMSAN)is proposed for chemical process.First,a sliding window strategy is adopted to construct the normalized time-series dataset.Second,the HMSAN is developed to extract the time-relevant features from the time-series process data.It improves the basic self-attention model in both width and depth.With the multihead structure,the HMSAN can pay attention to different aspects of the complicated chemical process and obtain the global dynamic features.However,the multiple heads in parallel lead to redundant information,which cannot improve the diagnosis performance.With the hierarchical structure,the redundant information is reduced and the deep local time-related features are further extracted.Besides,a novel many-to-one training strategy is introduced for HMSAN to simplify the training procedure and capture the long-term dependency.Finally,the effectiveness of the proposed method is demonstrated by two chemical cases.The experimental results show that the proposed method achieves a great performance on time-series industrial data and outperforms the state-of-the-art approaches.

Missing Value Imputation for Radar-Derived Time-Series Tracks of Aerial Targets Based on Improved Self-Attention-Based Network

The frequent missing values in radar-derived time-series tracks of aerial targets(RTT-AT)lead to significant challenges in subsequent data-driven tasks.However,the majority of imputation research focuses on random missing(RM)that differs significantly from common missing patterns of RTT-AT.The method for solving the RM may experience performance degradation or failure when applied to RTT-AT imputation.Conventional autoregressive deep learning methods are prone to error accumulation and long-term dependency loss.In this paper,a non-autoregressive imputation model that addresses the issue of missing value imputation for two common missing patterns in RTT-AT is proposed.Our model consists of two probabilistic sparse diagonal masking self-attention(PSDMSA)units and a weight fusion unit.It learns missing values by combining the representations outputted by the two units,aiming to minimize the difference between the missing values and their actual values.The PSDMSA units effectively capture temporal dependencies and attribute correlations between time steps,improving imputation quality.The weight fusion unit automatically updates the weights of the output representations from the two units to obtain a more accurate final representation.The experimental results indicate that,despite varying missing rates in the two missing patterns,our model consistently outperforms other methods in imputation performance and exhibits a low frequency of deviations in estimates for specific missing entries.Compared to the state-of-the-art autoregressive deep learning imputation model Bidirectional Recurrent Imputation for Time Series(BRITS),our proposed model reduces mean absolute error(MAE)by 31%~50%.Additionally,the model attains a training speed that is 4 to 8 times faster when compared to both BRITS and a standard Transformer model when trained on the same dataset.Finally,the findings from the ablation experiments demonstrate that the PSDMSA,the weight fusion unit,cascade network design,and imputation loss enhance imputation performance and confirm the efficacy of our design.

Modeling urban redevelopment:A novel approach using time-series remote sensing data and machine learning

Accurate mapping and timely monitoring of urban redevelopment are pivotal for urban studies and decisionmakers to foster sustainable urban development.Traditional mapping methods heavily depend on field surveys and subjective questionnaires,yielding less objective,reliable,and timely data.Recent advancements in Geographic Information Systems(GIS)and remote-sensing technologies have improved the identification and mapping of urban redevelopment through quantitative analysis using satellite-based observations.Nonetheless,challenges persist,particularly concerning accuracy and significant temporal delays.This study introduces a novel approach to modeling urban redevelopment,leveraging machine learning algorithms and remote-sensing data.This methodology can facilitate the accurate and timely identification of urban redevelopment activities.The study’s machine learning model can analyze time-series remote-sensing data to identify spatio-temporal and spectral patterns related to urban redevelopment.The model is thoroughly evaluated,and the results indicate that it can accurately capture the time-series patterns of urban redevelopment.This research’s findings are useful for evaluating urban demographic and economic changes,informing policymaking and urban planning,and contributing to sustainable urban development.The model can also serve as a foundation for future research on early-stage urban redevelopment detection and evaluation of the causes and impacts of urban redevelopment.

Mapping winter wheat using phenological feature of peak before winter on the North China Plain based on time-series MODIS data

By employing the unique phenological feature of winter wheat extracted from peak before winter （PBW） and the advantages of moderate resolution imaging spectroradiometer （MODIS） data with high temporal resolution and intermediate spatial resolution, a remote sensing-based model for mapping winter wheat on the North China Plain was built through integration with Landsat images and land-use data. First, a phenological window, PBW was drawn from time-series MODIS data. Next, feature extraction was performed for the PBW to reduce feature dimension and enhance its information. Finally, a regression model was built to model the relationship of the phenological feature and the sample data. The amount of information of the PBW was evaluated and compared with that of the main peak （MP）. The relative precision of the mapping reached up to 92% in comparison to the Landsat sample data, and ranged between 87 and 96% in comparison to the statistical data. These results were sufficient to satisfy the accuracy requirements for winter wheat mapping at a large scale. Moreover, the proposed method has the ability to obtain the distribution information for winter wheat in an earlier period than previous studies. This study could throw light on the monitoring of winter wheat in China by using unique phenological feature of winter wheat.

Temperature and Daily Mortality in Shanghai:A Time-series Study

To investigate the association between temperature and daily mortality in Shanghai from June 1, 2000 to December 31, 2001. Methods Time-series approach was used to estimate the effect of temperature on daily total and cause-specific mortality. We fitted generalized additive Poisson regression using non-parametric smooth functions to control for long-term time trend, season and other variables. We also controlled for day of the week. Results A gently sloping V-like relationship between total mortality and temperature was found, with an optimum temperature (e.g. temperature with lowest mortality risk) value of 26.7癈 in Shanghai. For temperatures above the optimum value, total mortality increased by 0.73% for each degree Celsius increase; while for temperature below the optimum value, total mortality decreased by 1.21% for each degree Celsius increase. Conclusions Our findings indicate that temperature has an effect on daily mortality in Shanghai, and the time-series approach is a useful tool for studying the temperature-mortality association.

Review of the SBAS InSAR Time-series algorithms, applications, and challenges

In the past 30 years,the small baseline subset(SBAS)InSAR time-series technique has emerged as an essential tool for measuring slow surface displacement and estimating geophysical parameters.Because of its ability to monitor large-scale deformation with millimeter accuracy,the SBAS method has been widely used in various geodetic fields,such as ground subsidence,landslides,and seismic activity.The obtained long-term time-series cumulative deformation is vital for studying the deformation mecha-nism.This article reviews the algorithms,applications,and challenges of the SBAS method.First,we recall the fundamental principle and analyze the shortcomings of the traditional SBAS algorithm,which provides a basic framework for the following improved time series methods.Second,we classify the current improved SBAS techniques from different perspectives:solving the ill-posed equation,increasing the density of high-coherence points,improving the accuracy of monitoring deformation and measuring the multi-dimensional deformation.Third,we summarize the application of the SBAS method in monitoring ground subsidence,permafrost degradation,glacier movement,volcanic activity,landslides,and seismic activity.Finally,we discuss the difficulties faced by the SBAS method and explore its future development direction.

Clustering Structure Analysis in Time-Series Data With Density-Based Clusterability Measure

Clustering is used to gain an intuition of the struc tures in the data.Most of the current clustering algorithms pro duce a clustering structure even on data that do not possess such structure.In these cases,the algorithms force a structure in the data instead of discovering one.To avoid false structures in the relations of data,a novel clusterability assessment method called density-based clusterability measure is proposed in this paper.I measures the prominence of clustering structure in the data to evaluate whether a cluster analysis could produce a meaningfu insight to the relationships in the data.This is especially useful in time-series data since visualizing the structure in time-series data is hard.The performance of the clusterability measure is evalu ated against several synthetic data sets and time-series data sets which illustrate that the density-based clusterability measure can successfully indicate clustering structure of time-series data.

Spatio-temporal changes of underground coal fires during 2008-2016 in Khanh Hoa coal field(North-east of Viet Nam) using Landsat time-series data

Underground coal fires are one of the most common and serious geohazards in most coal producing countries in the world. Monitoring their spatio-temporal changes plays an important role in controlling and preventing the effects of coal fires, and their environmental impact. In this study, the spatio-temporal changes of underground coal fires in Khanh Hoa coal field(North-East of Viet Nam) were analyzed using Landsat time-series data during the 2008-2016 period. Based on land surface temperatures retrieved from Landsat thermal data, underground coal fires related to thermal anomalies were identified using the MEDIAN+1.5×IQR(IQR: Interquartile range) threshold technique. The locations of underground coal fires were validated using a coal fire map produced by the field survey data and cross-validated using the daytime ASTER thermal infrared imagery. Based on the fires extracted from seven Landsat thermal imageries, the spatiotemporal changes of underground coal fire areas were analyzed. The results showed that the thermalanomalous zones have been correlated with known coal fires. Cross-validation of coal fires using ASTER TIR data showed a high consistency of 79.3%. The largest coal fire area of 184.6 hectares was detected in 2010, followed by 2014(181.1 hectares) and 2016(178.5 hectares). The smaller coal fire areas were extracted with areas of 133.6 and 152.5 hectares in 2011 and 2009 respectively. Underground coal fires were mainly detected in the northern and southern part, and tend to spread to north-west of the coal field.

Wavelet matrix transform for time-series similarity measurement

A time-series similarity measurement method based on wavelet and matrix transform was proposed,and its anti-noise ability,sensitivity and accuracy were discussed. The time-series sequences were compressed into wavelet subspace,and sample feature vector and orthogonal basics of sample time-series sequences were obtained by K-L transform. Then the inner product transform was carried out to project analyzed time-series sequence into orthogonal basics to gain analyzed feature vectors. The similarity was calculated between sample feature vector and analyzed feature vector by the Euclid distance. Taking fault wave of power electronic devices for example,the experimental results show that the proposed method has low dimension of feature vector,the anti-noise ability of proposed method is 30 times as large as that of plain wavelet method,the sensitivity of proposed method is 1/3 as large as that of plain wavelet method,and the accuracy of proposed method is higher than that of the wavelet singular value decomposition method. The proposed method can be applied in similarity matching and indexing for lager time series databases.

A Hybrid Neural Network Model for Marine Dissolved Oxygen Concentrations Time-Series Forecasting Based on Multi-Factor Analysis and a Multi-Model Ensemble

Dissolved oxygen(DO)is an important indicator of aquaculture,and its accurate forecasting can effectively improve the quality of aquatic products.In this paper,a new DO hybrid forecasting model is proposed that includes three stages:multi-factor analysis,adaptive decomposition,and an optimizationbased ensemble.First,considering the complex factors affecting DO,the grey relational(GR)degree method is used to screen out the environmental factors most closely related to DO.The consideration of multiple factors makes model fusion more effective.Second,the series of DO,water temperature,salinity,and oxygen saturation are decomposed adaptively into sub-series by means of the empirical wavelet transform(EWT)method.Then,five benchmark models are utilized to forecast the sub-series of EWT decomposition.The ensemble weights of these five sub-forecasting models are calculated by particle swarm optimization and gravitational search algorithm(PSOGSA).Finally,a multi-factor ensemble model for DO is obtained by weighted allocation.The performance of the proposed model is verified by timeseries data collected by the pacific islands ocean observing system(PacIOOS)from the WQB04 station at Hilo.The evaluation indicators involved in the experiment include the Nash–Sutcliffe efficiency(NSE),Kling–Gupta efficiency(KGE),mean absolute percent error(MAPE),standard deviation of error(SDE),and coefficient of determination(R^(2)).Example analysis demonstrates that:①The proposed model can obtain excellent DO forecasting results;②the proposed model is superior to other comparison models;and③the forecasting model can be used to analyze the trend of DO and enable managers to make better management decisions.

Monitoring of winter wheat distribution and phenological phases based on MODIS time-series： A case study in the Yellow River Delta, China

Accurate winter wheat identification and phenology extraction are essential for field management and agricultural policy making. Here, we present mechanisms of winter wheat discrimination and phenological detection in the Yellow River Delta（YRD） region using moderate resolution imaging spectroradiometer（MODIS） time-series data. The normalized difference vegetation index（NDVI） was obtained by calculating the surface reflectance in red and infrared. We used the Savitzky-Golay filter to smooth time series NDVI curves. We adopted a two-step classification to identify winter wheat. The first step was designed to mask out non-vegetation classes, and the second step aimed to identify winter wheat from other vegetation based on its phenological features. We used the double Gaussian model and the maximum curvature method to extract phenology. Due to the characteristics of the time-series profiles for winter wheat, a double Gaussian function method was selected to fit the temporal profile. A maximum curvature method was performed to extract phenological phases. Phenological phases such as the green-up, heading and harvesting phases were detected when the NDVI curvature exhibited local maximum values. The extracted phenological dates then were validated with records of the ground observations. The spatial patterns of phenological phases were investigated. This study concluded that, for winter wheat, the accuracy of classification is 87.07%, and the accuracy of planting acreage is 90.09%. The phenological result was comparable to the ground observation at the municipal level. The average green-up date for the whole region occurred on March 5, the average heading date occurred on May 9, and the average harvesting date occurred on June 5. The spatial distribution of the phenology for winter wheat showed a significant gradual delay from the southwest to the northeast. This study demonstrates the effectiveness of our proposed method for winter wheat classification and phenology detection.

Time-series gas prediction model using LS-SVR within a Bayesian framework

The traditional least squares support vector regression（LS-SVR）model,using cross validation to determine the regularization parameter and kernel parameter,is time-consuming.We propose a Bayesian evidence framework to infer the LS-SVR model parameters.Three levels Bayesian inferences are used to determine the model parameters,regularization hyper-parameters and tune the nuclear parameters by model comparison.On this basis,we established Bayesian LS-SVR time-series gas forecasting models and provide steps for the algorithm.The gas outburst data of a Hebi 10th mine working face is used to validate the model.The optimal embedding dimension and delay time of the time series were obtained by the smallest differential entropy method.Finally,within a MATLAB7.1 environment,we used actual coal gas data to compare the traditional LS-SVR and the Bayesian LS-SVR with LS-SVMlab1.5 Toolbox simulation.The results show that the Bayesian framework of an LS-SVR significantly improves the speed and accuracy of the forecast.

Time-series analysis of the characteristic pressure fluctuations in a conical fluidized bed with negative pressure

The negative pressure conical fluidized bed is widely used in the pharmaceutical industry.In this study,experiments based on the negative pressure conical fluidized bed are carried out by changing the material mass and particle size.The pressure fluctuation signals are analyzed by the time and the frequency domain methods.A method for absolutely characterizing the degree of the energy concentration at the main frequency is proposed,where the calculation is to divide the original power spectrum by the average signal power.A phenomenon where the gas velocity curve temporarily stops growing is observed when the material mass is light,and the particle size is small.The standard deviation and kurtosis both rapidly change at the minimum fluidization velocity and thus can be used to determine the flow regime,and the variation rule of the kurtosis is independent of both the material mass and particle size.In the initial fluidization stage,the dominant pressure signal comes from the material movement;with the increase in the gas velocity,the power of a 2.5 Hz signal continues to increase.A method of dividing the main frequency by the average cycle frequency can conveniently determine the fluidized state,and a novel concept called stable fluidized zone proposed in this paper can be obtained.Controlling the gas velocity within the stable fluidized zone ensures that the fluidized bed consistently remains in a stable fluidized state.

Time-series gene expression prof iles in AGS cells stimulated with Helicobacter pylori

AIM: To extend the knowledge of the dynamic interaction between Helicobacter pylori (H. pylori) and host mucosa. METHODS: A time-series cDNA microarray was performed in order to detect the temporal gene expression prof iles of human gastric epithelial adenocarcinoma cells infected with H. pylori. Six time points were selected to observe the changes in the model. A differential expression prof ile at each time point was obtained by comparing the microarray signal value with that of 0 h. Real-time polymerase chain reaction was subsequently performed to evaluate the data quality. RESULTS: We found a diversity of gene expression patterns at different time points and identifi ed a group of genes whose expression levels were significantly correlated with several important immune response and tumor related pathways. CONCLUSION: Early infection may trigger some important pathways and may impact the outcome of the infection.

Detecting winter canola(Brassica napus) phenological stages using an improved shape-model method based on time-series UAV spectral data

Accurate information about phenological stages is essential for canola field management practices such as irrigation, fertilization, and harvesting. Previous studies in canola phenology monitoring focused mainly on the flowering stage, using its apparent structure features and colors. Additional phenological stages have been largely overlooked. The objective of this study was to improve a shape-model method(SMM) for extracting winter canola phenological stages from time-series top-of-canopy reflectance images collected by an unmanned aerial vehicle(UAV). The transformation equation of the SMM was refined to account for the multi-peak features of the temporal dynamics of three vegetation indices(VIs)(NDVI, EVI, and CI). An experiment with various seeding scenarios was conducted, including four different seeding dates and three seeding densities. Three mathematical functions: asymmetric Gaussian function(AGF), Fourier function, and double logistic function, were employed to fit timeseries vegetation indices to extract information about phenological stages. The refined SMM effectively estimated the phenological stages of canola, with a minimum root mean square error(RMSE) of 3.7 days for all phenological stages. The AGF function provided the best fitting performance, as it captured multiple peaks in the growth dynamics characteristics for all seeding date scenarios using four scaling parameters. For the three selected VIs, CIred-edgeachieved the greatest accuracy in estimating the phenological stage dates. This study demonstrates the high potential of the refined SMM for estimating winter canola phenology.

Generating Time-Series Data Using Generative Adversarial Networks for Mobility Demand Prediction

The increasing penetration rate of electric kickboard vehicles has been popularized and promoted primarily because of its clean and efficient features.Electric kickboards are gradually growing in popularity in tourist and education-centric localities.In the upcoming arrival of electric kickboard vehicles,deploying a customer rental service is essential.Due to its freefloating nature,the shared electric kickboard is a common and practical means of transportation.Relocation plans for shared electric kickboards are required to increase the quality of service,and forecasting demand for their use in a specific region is crucial.Predicting demand accurately with small data is troublesome.Extensive data is necessary for training machine learning algorithms for effective prediction.Data generation is a method for expanding the amount of data that will be further accessible for training.In this work,we proposed a model that takes time-series customers’electric kickboard demand data as input,pre-processes it,and generates synthetic data according to the original data distribution using generative adversarial networks(GAN).The electric kickboard mobility demand prediction error was reduced when we combined synthetic data with the original data.We proposed Tabular-GAN-Modified-WGAN-GP for generating synthetic data for better prediction results.We modified The Wasserstein GAN-gradient penalty(GP)with the RMSprop optimizer and then employed Spectral Normalization(SN)to improve training stability and faster convergence.Finally,we applied a regression-based blending ensemble technique that can help us to improve performance of demand prediction.We used various evaluation criteria and visual representations to compare our proposed model’s performance.Synthetic data generated by our suggested GAN model is also evaluated.The TGAN-Modified-WGAN-GP model mitigates the overfitting and mode collapse problem,and it also converges faster than previous GAN models for synthetic data creation.The presented model’s performance is compared to existing ensemble and baseline models.The experimental findings imply that combining synthetic and actual data can significantly reduce prediction error rates in the mean absolute percentage error(MAPE)of 4.476 and increase prediction accuracy.

Time-Series Characteristics of Wind Power and Its Impact on Jilin Power Grid

With the rapid development of wind power, the large-scale wind power integration brings a new range of issues in dispatching operation. In order to gain a better grasp of the influence caused by wind power combined to the grid, the paper first establishes the impact characteristic indexes, and then analyzes the regularity of wind power time series in different spatial and temporal scales. At last, according to the analysis results, this paper assesses the impact of time-series characteristics of wind power on power grid, such as the frequency regulation, peak load regulation, which can provide the reference for wind power optimal dispatching of Jilin Power Grid.

Classification of Vegetation in North Tibet Plateau Based on MODIS Time-Series Data

Based on the 16d-composite MODIS （moderate resolution imaging spectroradiometer）-NDVI（normalized difference vegetation index） time-series data in 2004, vegetation in North Tibet Plateau was classified and seasonal variations on the pixels selected from different vegetation type were analyzed. The Savitzky-Golay filtering algorithm was applied to perform a filtration processing for MODIS-NDVI time-series data. The processed time-series curves can reflect a real variation trend of vegetation growth. The NDVI time-series curves of coniferous forest, high-cold meadow, high-cold meadow steppe and high-cold steppe all appear a mono-peak model during vegetation growth with the maximum peak occurring in August. A decision-tree classification model was established according to either NDVI time-series data or land surface temperature data. And then, both classifying and processing for vegetations were carried out through the model based on NDVI time-series curves. An accuracy test illustrates that classification results are of high accuracy and credibility and the model is conducive for studying a climate variation and estimating a vegetation production at regional even global scale.

TIME-SERIES MODELI NG AND FAULT FORECAST STUDY ON SPECTRAL ANALYSIS OF LUBRICATING OIL

The application of ti me-series modeling and forecasting method to the spectral analysis for lubricat ing oil of mechanical equipment is discussed. The AR model is used to perform a time-series modeling and forecasting analysis for the spectral analysis data co llected from aero-engines. In the oil condition monitoring field of mechanical equipment, the use of the method of time-series analysis has rarely been report ed. As indicated in the satisfactory example, a practical method for condition m onitoring and fault forecasting of mechanical equipment has been achieved.

Study and application of monitoring plane displacement of a similarity model based on time-series images

In order to compensate for the deficiency of present methods of monitoring plane displacement in similarity model tests,such as inadequate real-time monitoring and more manual intervention,an effective monitoring method was proposed in this study,and the major steps of the monitoring method include:firstly,time-series images of the similarity model in the test were obtained by a camera,and secondly,measuring points marked as artificial targets were automatically tracked and recognized from time-series images.Finally,the real-time plane displacement field was calculated by the fixed magnification between objects and images under the specific conditions.And then the application device of the method was designed and tested.At the same time,a sub-pixel location method and a distortion error model were used to improve the measuring accuracy.The results indicate that this method may record the entire test,especially the detailed non-uniform deformation and sudden deformation.Compared with traditional methods this method has a number of advantages,such as greater measurement accuracy and reliability,less manual intervention,higher automation,strong practical properties,much more measurement information and so on.