Extracting Low-Frequency Spatio-Temporal Patterns in Ambient Power System Data Using Blind Source Separation

With advances in modern worlds technology, huge datasets that show dependencies in space as well as in time occur frequently in practice. As an example, several monitoring stations at different geographical locations track hourly concentration measurements of a number of air pollutants for several years. Such a dataset contains thousands of multivariate observations, thus, proper statistical analysis needs to account for dependencies in space and time between and among the different monitored variables. To simplify the consequent multivariate spatio-temporal statistical analysis it might be of interest to detect linear transformations of the original observations that result in straightforward interpretative, spatio-temporally uncorrelated processes that are also highly likely to have a real physical meaning. Blind source separation (BSS) represents a statistical methodology which has the aim to recover so-called latent processes, that exactly meet the former requirements. BSS was already successfully used in sole temporal and sole spatial applications with great success, but, it was not yet introduced for the spatio-temporal case. In this contribution, a reasonable and innovative generalization of BSS for multivariate space-time random fields (stBSS), under second-order stationarity, is proposed, together with two space-time extensions of the well-known algorithms for multiple unknown signals extraction (stAMUSE) and the second-order blind identification (stSOBI) which solve the formulated problem. Furthermore, symmetry and separability properties of the model are elaborated and connections to the space-time linear model of coregionalization and to the classical principal component analysis are drawn. Finally, the usefulness of the new methods is shown in a thorough simulation study and on a real environmental application.

show abstract

“… 2009 ; de Jesús Nuño Ayón et al. 2018 ). All these methods work in a very different setting with respect to the one considered hereafter.…”

Section: Introductionmentioning

confidence: 99%

Blind recovery of sources for multivariate space-time random fields

Muehlmann

Iaco

Nordhausen

2022

Stoch Environ Res Risk Assess

View full text Add to dashboard Cite

show abstract

“…These methods are commonly applied to a set of simultaneous measurements for the modal parameters identification. [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24] For example, frequency domain decomposition (FDD) applies the singular value decomposition to the power spectrum matrix to identify modal information, but FDD requires long data records. 10 Subspace identification-based techniques use the space-state description of the power system dynamics to estimate the modal properties (frequency, damping ratio, and mode shape) obtained from the state matrix 11,14 ; however, these techniques may require a high computational load.…”

Section: Introductionmentioning

confidence: 99%

“…20 Blind source separation allows extracting temporal and spatial patterns associated with inter-area oscillation modes and observation factors, respectively. 21 However, the damping ratios are not estimated. The output-only observer/Kalman filter identification method, 22 an eigensystem realization algorithmbased data-driven approach, 23 and a Hankel block-enhanced dynamic mode decomposition 24 are novel data analysis techniques that have been improved and proposed to estimate oscillation modes and mode shapes from ambient power system measurements.…”

Section: Introductionmentioning

confidence: 99%

“…Numerous data analysis tools have been proposed to improve the observability of the oscillatory modes. These methods are commonly applied to a set of simultaneous measurements for the modal parameters identification 10‐24 . For example, frequency domain decomposition (FDD) applies the singular value decomposition to the power spectrum matrix to identify modal information, but FDD requires long data records 10 .…”

Section: Introductionmentioning

confidence: 99%

“…Multivariate empirical mode decomposition is developed to analyze nonstationary signals and provide information on the modal frequency and its mode shape, while the damping ratio is not identified 20 . Blind source separation allows extracting temporal and spatial patterns associated with inter‐area oscillation modes and observation factors, respectively 21 . However, the damping ratios are not estimated.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A tensor decomposition‐based approach for the analysis and visualization of ambient power system oscillations

Nuño-Ayón

Bañuelos-Cabral

Sotelo-Castañón

et al. 2021

Int Trans Electr Energ Syst

View full text Add to dashboard Cite

In recent years, more and more data analysis tools have been developed to analyze and visualize ambient power system oscillations that are recorded from modern measurement systems. In this paper, a tensor decompositionbased approach is proposed for characterizing and visualizing the dynamic information contained in the ambient data. The approach exploits the correlation of the data through several covariance matrices that are used to form a third-order tensor. From this tensor, the autocovariance signal corresponding to each modal response and its modal amplitude vector (pseudo-mode shape) can be extracted by the parallel factor decomposition. Subsequently, a matrix pencil algorithm is applied to each autocovariance signal to estimate its natural frequency and damping ratio. The proposed methodology was applied to ambient data obtained from a transient stability model of the New England test system/New York power system, and the effect of various method parameters on the quality of the modal identification was analyzed. The results indicated that the oscillatory frequencies, damping ratios, and pseudo-mode shapes can be accurately estimated using the proposed method. Therefore, it can be concluded that the proposed approach is promising in terms of monitoring the ambient oscillations using measured data.

show abstract