Power equipment anomaly detection based on spatiotemporal clustering

The timely analysis of deformation monitoring data and reasonable diagnosis of the structural health are key tasks in dam health monitoring studies. This article presents a spatio-temporal clustering and health diagnosis method for super-high concrete arch dams that uses deformation monitoring data obtained from plumb meters. The spatio-temporal expression of the deformation monitoring data is proposed first by upgrading a punctuated time series to a curved panel time series, including cross-sectional, dam axial, and temporal changing directions. Second, a comprehensive similarity indicator on three aspects, namely, the absolute distance, incremental distance, and growth rate distance, is constructed after a deep discussion on deformation similarity characteristics both temporally and spatially. Next, the temporal clustering method is proposed by keeping the key features, namely, extreme points and turning points, while eliminating extraneous details, namely, noise points. Finally, the optimal spatio-temporal clustering of dam deformation is achieved by designing a multi-scale fuzzy C-means method of data mining and its iterative algorithm. The proposed method is applied to the Jinping-I hydraulic structure, which is the highest concrete arch dam in the world. The clustering results is quite sensitive in different weight coefficients of the comprehensive similarity indicator and clustering numbers of fuzzy C-means method. The dam deformation behaviors on high-water-level, water-falling, and low-water-level periods are analyzed and diagnosed. The advanced version of proposed methods is verified by comparative analysis on dam health diagnosis results obtained from ordinary deformation distribution figures and the spatio-temporal clustering figures. The proposed method will facilitate the recognition of abnormal deformation areas and associated safety diagnoses.

The prediction of land subsidence is of significant value for the early warning and prevention of geological disasters. Although numerous land subsidence prediction methods are currently available, two obstacles still exist: (i) spatio-temporal heterogeneity of land subsidence is not well considered, and (ii) the prediction performance of individual models is unsatisfactory when the data do not meet their assumptions. To address these issues, we developed a spatio-temporal heterogeneous ensemble learning method for predicting land subsidence. Firstly, a two-stage hybrid spatio-temporal clustering method was proposed to divide the dataset into internally homogeneous spatio-temporal clusters. Secondly, within each spatio-temporal cluster, an ensemble learning strategy was employed to combine one time series prediction model and three spatio-temporal prediction models to reduce the prediction uncertainty of an individual model. Experiments on a land subsidence dataset from Cangzhou, China, show that the prediction accuracy of the proposed method is significantly higher than that of four individual prediction models.

Improved partitioning technique for density cube-based spatio-temporal clustering method

Abstract. Considering the critical role of trajectory data in Big Data era for dynamic geographical processes, human behaviour analysis and meteorological prediction, trajectory clustering has attracted growing attention. Many literatures have discussed the spatiot emporal clustering method of simple trajectories (i.e., has no branches, e.g. vehicle trajectories), yet there are few researches for clustering complex trajectories (i.e., has at least one split and/or merger and/or split -merger branch, e.g. ocean eddy trajectories, rainstorm trajectories). For addressing this issue, we propose a Process-Oriented Spatiotemporal Clustering Method (POSCM) for clustering complex trajectory data. The POSCM includes three parts: the first uses the semantic of process-sequence-state to represent the complex trajectories; the second proposes a Hierarchical Similarity Measurement Method (HSMM) to get the similarity between any two complex trajectories; in the last step, the complex trajectories clustering pattern is extracted through density-based clustering algorithm. Experiments on simulated trajectories are used to evaluate the POSCM and demonstrate the advantage by comparing against that of the VF2 algorithm. The POSCM is applied to the sea surface temperature abnormal variations trajectories from January 1950 to December 2017 in the Pacific Ocean. As shown in this case study, some new mined spatiotemporal patterns can provide new references for understanding the behaviours of marine abnormal variations under the background of the global change.

This paper proposes a method for clustering asynchronous events generated upon scene activities by a dynamic 3D vision system. The inherent detection of moving objects offered by the dynamic stereo vision system comprising a pair of dynamic vision sensors allows event-based stereo vision in real-time and a 3D representation of moving objects. The clustering method exploits the sparse spatio-temporal representation of sensor's events for real-time detection and separation between moving objects. The method makes use of density and distance metrics for clustering asynchronous events generated by scene dynamics (changes in the scene). It has been evaluated on clustering the events of moving persons across the sensor field of view. Tests on real scenarios with more than 100 persons show that the resulting asynchronous events can be successfully clustered and the persons can be detected.

Monte Carlo simulation-based traffic speed forecasting using historical big data

With the rapid advancements in positioning technologies, the volume of spatio-temporal data has grown significantly. Analyzing the spatial and temporal characteristics of these data is imperative for uncovering underlying associations and deriving insights into natural and societal mechanisms. Clustering is a widely utilized technique for data analysis, which groups data with similar characteristics for further investigation. However, current clustering methodologies usually inadequately address temporal properties that are vital in numerous scenarios. Additionally, traditional spatio-temporal clustering approaches are constrained to standalone environments, which struggle to handle large-scale spatio-temporal datasets. To this end, we introduce DiST, the first distributed spatio-temporal clustering method, which simultaneously considers both temporal and spatial proximity. DiST comprises data partition, local clustering, and global merging stages, along with an auto-tuning framework for parameter optimization. DiST addresses key challenges, including the integration of temporal and spatial attributes, managing data duplication across distributed nodes, and selecting appropriate parameters for diverse data characteristics. Comparative experiments on two real-world datasets validate the performance and scalability of DiST, demonstrating its effectiveness in spatio-temporal data analysis.

With ever-increasing road mileages worldwide, more pavement deterioration and ageing present great challenges to the maintenance and rehabilitation (M&R) of road pavement. In this study, an intelligent decision-making framework is developed for pavement maintenance using the clustering-PageRank algorithm (CPRA) based on historical big data. The proposed model is applied to a 3.5 km pavement (500 road sections) and leads to recommendations for the optimal pavement maintenance plans with appropriate possibilities. The results indicate that seven plans are the same as those obtained by the experience-based maintenance approach, while the other three are similar. The framework is also verified by comparison with the experience-based maintenance activities and is found to have limited reliability when dealing with a small quantity of solutions. The method and results of this study are expected to serve as a reference for decision makers to make well-informed project decisions on the optimum M&R activities.

Applying Spatio-temporal Scan Statistics and Spatial Autocorrelation Statistics to identify Covid-19 clusters in the world - A Vaccination Strategy?

Some methods for extracting the symptoms of faults in power equipment are presented by analyzing and modeling time series (TS). In this paper, the on-line data are preprocessed at first, which will be helpful to reduce the influence of temperature and humidity. And then the relevant autoregressive-moving average (ARMA) models are established. Furthermore, the characteristic values of the model can be calculated. These values are used as symptoms to estimate the state of power equipment. In this paper the concepts of stationary stochastic process and time series are also introduced. At last some measured data, which from a substation in Guangdong province prove the feasibility of the method.

The need to operate electrical devices the design capabilities of which have been exhausted almost completely under the conditions of more stringent reliability requirements gives rise to increased attention to the development and research of methods for diagnosing the power equipment state. This problem acquires not only a scientific interest, but also practical significance, which follows, in particular, from the mass-scale use of power transformers for operating voltages above 35 kV. One of promising development lines is the method of constructing dielectric absorption spectra, which is based on using the capabilities of isothermal depolarization current measurements. In this case, the information that accumulates as a result of the insulation system’s dielectric medium polarization as a certain number of loosely coupled charges displaced from their locations is adequate (equivalent) to the information obtained as a result of depolarization when measuring the return voltage (RV) value. Owing to the above-mentioned adequacy, it is possible to use the quantity tI(t) as a parameter for monitoring the state of the power equipment insulation system. The spectrum construction algorithm presented in the article is similar to the algorithm for constructing the RV spectrum, but is produced in the course of computer-aided processing of measurement results using the time windows that specify the insulation gap polarization time interval. As a result, it becomes possible not only to produce the dielectric absorption spectrum, but also to calculate the values of the main parameters characterizing the gap state.

Spatio-temporal chaos: A solvable model

Automatic interpolation of precipitation maps combining rain gauge and radar data has been done in the past but considering only the data collected at a given time interval. Since radar and rain gauge data are collected at short intervals, a natural extension of previous works is to account for temporal correlations and to include time into the interpolation process. In this work, rainfall is interpolated using data from the current time interval and the previous one. Interpolation is carried out using kriging with external drift, in which the radar rainfall estimate is the drift, and the mean precipitation is set to zero at the locations where the radar estimate is zero. The rainfall covariance is modeled as non-stationary in time, and the space system of reference moves with the storm. This movement serves to maximize the collocated correlation between consecutive time intervals. The proposed approach is analyzed for four episodes that took place in Catalonia (Spain). It is compared with three other approaches: (i) radar estimation, (ii) kriging with external drift using only the data from the same time interval, and (iii) kriging with external drift using data from two consecutive time intervals but not accounting for the displacement of the storm. The comparisons are performed using cross-validation. In all four episodes, the proposed approach outperforms the other three approaches. It is important to account for temporal correlation and use a Lagrangian system of coordinates that tracks the rainfall movement.

Partial Discharge (PD) detection and its estimation is a vital event for all the power equipment. Acoustic sensor based PD detection is more advantageous than the other available PD measurement technique. Acoustic PD detection technique has noise immunity ability to perform online partial discharge detection. In this work, a real life transformer model (RLTM) has been developed in the laboratory. Experimental procedure using acoustic sensor has been adopted to generate PD within RLTM. Several acoustic PD data have been acquired through acoustic sensors by placing the artificially created PD source at various locations inside the RL TM. Each of the sensor readings can be thought of as a waveform (time-series data) whose structure changes depending on the fault location. The novelty of the work lies in the fact that a number of features have been extracted using the Time Series Feature Extraction on Basis of Scalable Hypothesis tests (tsfresh- A Python Package) from the acoustic partial discharge data followed by classification using Random Forest classifier. Feature Selection is executed using Mutual Information Method. Percentage of identification accuracy for individual and coupled partial discharge sources are 94.4% and 91.5% respectively

Wave height prediction is an important task for ocean and marine resource management. Traditionally, regression techniques are used for this prediction, but estimating continuous changes in the corresponding time series can be very difficult. With the purpose of simplifying the prediction, wave height can be discretised in consecutive intervals, resulting in a set of ordinal categories. Despite this discretisation could be performed using the criterion of an expert, the prediction could be biased to the opinion of the expert, and the obtained categories could be unrepresentative of the data recorded. In this paper, we propose a novel automated method to categorise the wave height based on selecting the most appropriate distribution from a set of well-suited candidates. Moreover, given that the categories resulting from the discretisation show a clear natural order, we propose to use different ordinal classifiers instead of nominal ones. The methodology is tested in real wave height data collected from two buoys located in the Gulf of Alaska and South Kodiak. We also incorporate reanalysis data in order to increase the accuracy of the predictors. The results confirm that this kind of discretisation is suitable for the time series considered and that the ordinal classifiers achieve outstanding results in comparison with nominal techniques.