Real-time Disturbance Detection Research Articles

Advanced power quality monitoring with dilated convolutional neural networks for smart nanogrids

This study presents the development of an advanced power quality monitoring (PQM) system using dilated convolutional deep neural networks (DCDNN) for smart nanogrids. Traditional convolutional neural networks (CNNs) face challenges in efficiently handling high-dimensional data and identifying nuanced features due to their limited receptive field. To overcome these limitations, our DCDNN model employs a novel use of dilated convolutions to expand the receptive field without increasing network complexity, enabling the capture of long-range dependencies essential for accurate classification of power quality disturbances (PQDs). In addition, the integration of multi-channel time-series data processing, including voltage, current, and frequency measurements, is unique to our approach. The model processes multi-channel time-series data, including voltage, current, and frequency measurements, achieving an average accuracy of 98.9% across 29 PQD classes and 100% accuracy in 6 main classes, with a reduced detection time of 47 µs per segment. This approach not only improves real-time disturbance detection but also enhances the reliability and efficiency of electrical systems. The results demonstrate significant improvements over existing methods, underscoring the potential of DCDNNs in advancing PQM practices. Future work will focus on validating the model in real-world conditions and optimizing its computational efficiency.

Collaborative Reconfiguration of Supply Networks Based on GNN and ALC

With the prevalence of lean and just-in-time principles, traditional supply chains often exhibit inflexibility, leading to challenges in satisfying extensive customized orders and managing risks during disruptions. Thus, there is a need for a more flexible, resilient, and collaborative network and strategies to tackle the aforementioned challenges. In this study, we introduce a new supply network called the industry supply chain, aimed at enabling collaborative decision-making and dynamic reconfiguration. We create a graph neural network model to promptly identify sudden disturbances and devise a distributed multidisciplinary optimization model to facilitate collaborative reconfiguration. The experimental findings from an air-conditioning industry supply chain show that network reconfiguration under real-time disturbance detection reduces losses and improves operational stability.

Using An Attention-Based LSTM Encoder–Decoder Network for Near Real-Time Disturbance Detection

Accurate prediction of future observations based on past data is the key to near real-time disturbance detection using satellite image time series (SITS). To overcome the limitations of existing methods, we present an attention-based long-short-term memory (LSTM) encoder-decoder model in which the historical time series of a pixel is encoded with a bidirectional LSTM encoder while the future time series is produced by another LSTM decoder. An attention mechanism is integrated into the encoder-decoder model to align the input time series with the output time series and to dynamically choose the most relevant contextual information while forecasting. Based on the proposed model, we develop a framework for near real-time disturbance detection and verify its effectiveness in the case of burned area mapping. The prediction accuracy of the proposed model is evaluated using moderate resolution imaging spectroradiometer (MODIS) time series and compared with state-of-the-art models. Experimental results show that our model achieves the best results in terms of lower prediction error and higher model fitness. We also evaluate the disturbance detection ability of the proposed framework. The proposed approach improves the detection rate of disturbances while suppressing false alarms, and increases the temporal accuracy. We suggest that the proposed methods provide new tools for enhancing current early warning systems in real time.

Improved disturbance detection and load shedding technique for low voltage islanded microgrids

The rate of change of frequency (RoCoF) following any disturbance in low voltage islanded microgrids is not only relatively high compared to conventional systems, but also becomes location specific due to the presence of highly resistive lines. This makes the disturbance detection challenging. A novel real-time disturbance detection technique for low voltage islanded microgrids is proposed that is independent of the system inertia, which can detect any disturbance within 100 ms. A dedicated initial RoCoF calculation is developed over a 30 ms window especially during disturbances. The authors use a support vector machine-based regression model for the estimation of the amount of disturbance with three different feature selections. These models prove that only the RoCoF is not enough to estimate the amount of disturbance in an islanded microgrid and additional features from the frequency and different bus voltages also required. A load shedding algorithm encompassing battery support is also proposed to protect the system from quick frequency collapse by providing power from the battery when the frequency reaches its standard operational limit. All the proposed algorithms are tested with a low voltage islanded microgrid modelled in MATLAB/SIMULINK. The results are also verified with noisy data to reflect real system performance.

Boundary Wavelet Coefficients for Real-Time Detection of Transients Induced by Faults and Power-Quality Disturbances

In the last few years, much research has been focused on wavelet coefficients for high-speed detection of transients induced by faults, high impedance faults (HIFs), and some power-quality (PQ) disturbances. However, the performance of the wavelet coefficient-based analysis in the real-time disturbance detection substantially changes with the choice of the mother wavelet, presents time delay for long mother wavelets, and may fail when the transients are overdamped, which limits the applicability of the conventional wavelet coefficients in actual digital relays. Therefore, in order to overcome all of the aforementioned drawbacks of the wavelet coefficients, this paper proposes a new wavelet decomposition process which provides additional wavelet coefficients with border distortions of a one-cycle sliding window (boundary wavelet coefficients), yielding the real-time detection of the transients scarcely affected by the choice of the mother wavelet, with no time delay for compact and long wavelets, as well as a more reliable disturbance detection than the conventional wavelet coefficients, even in cases of damped transients. In addition, a boundary wavelet coefficient requires equivalent computational burden of a conventional wavelet coefficient. The performance of the proposed boundary wavelet coefficients was assessed through the real-time analysis of records with faults, HIFs, and some PQ disturbances as well as with simulated data from transmission and distribution power systems, taking into account the effects of the fault inception angle, resistance, and distance.

Near real-time disturbance detection using satellite image time series

Near real-time monitoring of ecosystem disturbances is critical for rapidly assessing and addressing impacts on carbon dynamics, biodiversity, and socio-ecological processes. Satellite remote sensing enables cost-effective and accurate monitoring at frequent time steps over large areas. Yet, generic methods to detect disturbances within newly captured satellite images are lacking. We propose a multi-purpose time-series-based disturbance detection approach that identifies and models stable historical variation to enable change detection within newly acquired data. Satellite image time series of vegetation greenness provide a global record of terrestrial vegetation productivity over the past decades. Here, we assess and demonstrate the method by applying it to (1) simulated time series of vegetation greenness data from satellite data, (2) real-world satellite greenness image time series between February 2000 and July 2011 covering Somalia to detect drought-related vegetation disturbances. First, simulation results illustrate that disturbances are successfully detected in near real-time while being robust to seasonality and noise. Second, major drought-related disturbance corresponding with most drought-stressed regions in Somalia are detected from mid-2010 onwards. The method can analyse in-situ or satellite data time series of biophysical indicators from local to global scale since it is fast, does not depend on thresholds and does not require time series gap filling. While the data and methods used are appropriate for proof-of-concept development of global scale disturbance monitoring, specific applications (e.g., drought or deforestation monitoring) mandate integration within an operational monitoring framework (e.g., http://www.fews.net/).

A Simple Method For Real-Time Detection Of Voltage Sags and Swells in Practical Loads

This paper presents an algorithm for real time detection of short duration voltage disturbances that occur in single phase power supply systems. The method used involves the use of a sliding, overlapping window of fixed size for measuring the RMS(root mean square) value of the voltage signal. The RMS value of a fixed number of voltage samples, is computed over a cycle, and this value is updated over a sample period by the overlapping window. A trigger signal is generated in response to the comparison of the RMS values between two consecutive cycles of the voltage signal. The trigger thus obtained, can be used to initiate the operation of the main control system of a Dynamic Voltage Restorer (DVR) to be used in conjunction with the detection unit to mitigate the sag or swell, or as a triggering function to a digital recorder to record the occurrence of the disturbance. The modeling of the algorithm and test of its functioning was carried out in Matlab - Simulink environment. Reliability of the functioning of the algorithm under conditions of noise and change in the nature of the sag and swell in the voltage signal was tested on real data acquired during events such as starting of induction motor, intermittent loading of welding transformer, and dynamic loading of a captive diesel generator. TMS320F6713 Digital Signal Processor used for the implementation of the algorithm has been programmed through the Matlab embedded link for code composer studio. The experimental set up consists of a digital signal processor (DSP) TMS320C6713, for generation of trigger, a personal computer for record of the parameters and is successfully tested for real time disturbance detection.