scispace - formally typeset
Search or ask a question
Journal ArticleDOI

Real-Time Event Classification in Power System With Renewables Using Kernel Density Estimation and Deep Neural Network

TL;DR: A kernel density estimation approach for accurate real-time classification of events in a power system with renewables using synchrophasor data using a diffusion type kernel density estimator (DKDE) to characterize the shape of 3-D voltage and frequency distribution along time in terms of probability density functions (PDFs).
Abstract: Real-time classification of events facilitates corrective control strategies, supervisory protection schemes, and on-line transient stability assessment of a power system. The synchrophasor-based event classification techniques face challenges like similar responses for different classes of events, i.e., inter-class similarity (ICS), applicability to limited classes of events, and moderate real-time performance for a large power system. In addition, the enhanced ICS effect of increased renewable penetration on events classification needs to be addressed. This paper proposes a kernel density estimation approach for accurate real-time classification of events in a power system with renewables using synchrophasor data. The proposed method uses a diffusion type kernel density estimator (DKDE) to characterize the shape of 3-D voltage and frequency distribution along time in terms of probability density functions (PDFs). That have distinct scale, shape, and orientation for different classes of events. Thereafter, a set of statistical features is derived from PDFs to train a multi-layered deep neural network for event classification. The proposed method is validated for renewables in IEEE-39 bus system and real transmission system of India grid using DIgSILENT/PowerFactory and also on a real phasor measurement unit data for India grid, where it showed better performance for ICS and renewable integration cases.
Citations
More filters
Proceedings ArticleDOI
17 Jul 2016
TL;DR: In this article, the effect of wind power on frequency regulation capability at different penetration levels is examined and the analytical and simulation results presented here provide some guidance on determining maximum wind power penetration level given a frequency deviation limit.
Abstract: The integration of renewable energy sources into power systems has gathered significant momentum globally because of its unlimited supply and environmental benefits. Within the portfolio of renewable energy, wind power is expected to have a soaring growth rate in the coming years. Despite its well known benefits, wind power poses several challenges in grid integration. The inherent intermittent and non-dispatchable features of wind power not only inject additional fluctuations to the already variable nature of frequency deviation, they also decrease frequency stability by reducing the inertia and the regulation capability. This paper closely examines these effects as well as the effect on tie-line flows and area control error, which causes a larger and longer frequency deviation in the integrated system. Further, the effect of wind power on frequency regulation capability at different penetration levels is also examined. The analytical and simulation results presented here provide some guidance on determining maximum wind power penetration level given a frequency deviation limit.

30 citations

Journal ArticleDOI
TL;DR: The proposed model is formulated to determine the relationship between the power capacity and wind energy loss, considering the wind curtailment loss and traditional energy power uncertain reserve, and the non-parametric kernel density estimation is adopted.
Abstract: Reasonable energy storage capacity in a high source-to-charge ratio local power grid can not only reduce system costs but also improve local power supply reliability. This paper introduces the capacity sizing of energy storage system based on reliable output power. The proposed model is formulated to determine the relationship between the power capacity and wind energy loss, considering the wind curtailment loss and traditional energy power uncertain reserve. The non-parametric kernel density estimation is adopted to estimate the confidence intervals of wind power prediction error and fluctuation ranges of the actual output of a wind farm under different confidence degrees. The actual historical data of scenery resources in a certain area is used to verify the feasibility of the proposed method. The simulation shows the large-capacity energy storage, the reliable output power of the microgrid wind ensures the feasibility of day-ahead generation plans.

28 citations

Journal ArticleDOI
TL;DR: In this paper , an integrated event identification algorithm for power systems is proposed, where an event detection trigger based on the rate of change of frequency (RoCoF) is presented, and wave arrival time difference-based triangulation method considering the anisotropy of wave propagation speed is utilized to estimate the location of the detected event.
Abstract: Accurate event identification is an essential part of situation awareness ability for power system operators. Therefore, this work proposes an integrated event identification algorithm for power systems. First, to obtain and filter suitable inputs for event identification, an event detection trigger based on the rate of change of frequency (RoCoF) is presented. Then, the wave arrival time difference-based triangulation method considering the anisotropy of wave propagation speed is utilized to estimate the location of the detected event. Next, the two-dimensional orthogonal locality preserving projection (2D-OLPP)-based method, which is suitable for multiple types of measured data, is employed to achieve higher effectiveness in extracting the event features compared with traditional one-dimensional projection and principle component analysis (PCA). Finally, the random undersampling boosted (RUSBoosted) trees-based classifier, which can mitigate the data sample imbalance issue, is utilized to identify the type of the detected event. The proposed approach is demonstrated using the actual measurement data of U.S. power systems from FNET/GridEye. Comparison results show that the proposed event identification algorithm can achieve better performance than existing approaches.

22 citations

Journal ArticleDOI
TL;DR: Extensive results obtained on the IEEE 39-bus system as well as a large-scale power system in China with field PMU measurements show that the proposed method achieves the highest classification accuracy and computational efficiency as compared to other deep learning algorithms.
Abstract: Data quality issues exist in practical phasor measurement units (PMUs) due to communication errors or signal interferences. As a result, the performances of existing data-driven disturbance classification methods can be significantly affected. In this article, a fast disturbance classification method that is robust to PMU data quality issues is proposed. The impacts of bad PMU measurements on disturbance classification are investigated by analyzing the feature distributions of deep learning methods. A new feature extraction scheme that uses the univariate temporal convolutional denoising autoencoder (UTCN-DAE) is proposed. It allows encoding and decoding univariate disturbance data through a temporal convolutional network to capture the temporal feature representation and is robust to bad data. Based on the features of the frequency and voltage measurements encoded by the UTCN-DAE, a two-stream enhanced network, i.e., the multivariable temporal convolutional denoising network is proposed to achieve optimal feature extraction of multivariate time series by feature fusion. The classification is performed using a multilayered deep neural network and Softmax classifier. Extensive results obtained on the IEEE 39-bus system as well as a large-scale power system in China with field PMU measurements show that the proposed method achieves the highest classification accuracy and computational efficiency as compared to other deep learning algorithms.

22 citations

Journal ArticleDOI
TL;DR: A novel event detection and classification scheme based on wide area frequency measurement system (WAFMS) that requires less input measurement for decision making and is having a low computational complexity, which is suitable for practical application is proposed.
Abstract: Automated event detection and classification are vital to power system monitoring. This article proposes a novel event detection and classification scheme based on wide area frequency measurement system (WAFMS). Raw frequency measurements obtained from WAFMS are used as the only input to the event detection and classification algorithm (EDCA). Wavelet-based signal pre-processing is used to denoise the data. Afterward, the rate of change of frequency (ROCOF) is estimated from the frequency measurements using the Kalman filter (KF). In the same step, phase angle difference (PAD) across different stations is estimated using WAFMS. Thus, the overall algorithm uses three features such as frequency, ROCOF, and PAD to detect and classify events in the power system. In the first step, an event is detected based on standard deviation (SD) of estimated ROCOF and PAD. In the second step, four types of events are classified using wide area frequency measurements. The suggested algorithms have been validated with real WAFMS data from the Indian Power System, recent 9th August 2019 U.K. blackout data collected from the U.K. power system, and 20th July 2017 oscillation event data obtained from ISO New England (ISO-NE) power system. As a promising tool for power system monitoring, the suggested scheme requires less input measurement for decision making and is having a low computational complexity, which is suitable for practical application.

16 citations


Cites background from "Real-Time Event Classification in P..."

  • ...In [6], a deep neural network (DNN) based real-time event classification scheme is suggested....

    [...]

  • ...schemes reported in [6], [7], and [16] use both voltage and frequency signals from PMUs to detect and classify events....

    [...]

References
More filters
Journal ArticleDOI
28 May 2015-Nature
TL;DR: Deep learning is making major advances in solving problems that have resisted the best attempts of the artificial intelligence community for many years, and will have many more successes in the near future because it requires very little engineering by hand and can easily take advantage of increases in the amount of available computation and data.
Abstract: Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.

46,982 citations

Journal ArticleDOI
01 Mar 1974

3,841 citations

Journal ArticleDOI
20 Nov 2017
TL;DR: In this paper, the authors provide a comprehensive tutorial and survey about the recent advances toward the goal of enabling efficient processing of DNNs, and discuss various hardware platforms and architectures that support DNN, and highlight key trends in reducing the computation cost of deep neural networks either solely via hardware design changes or via joint hardware and DNN algorithm changes.
Abstract: Deep neural networks (DNNs) are currently widely used for many artificial intelligence (AI) applications including computer vision, speech recognition, and robotics. While DNNs deliver state-of-the-art accuracy on many AI tasks, it comes at the cost of high computational complexity. Accordingly, techniques that enable efficient processing of DNNs to improve energy efficiency and throughput without sacrificing application accuracy or increasing hardware cost are critical to the wide deployment of DNNs in AI systems. This article aims to provide a comprehensive tutorial and survey about the recent advances toward the goal of enabling efficient processing of DNNs. Specifically, it will provide an overview of DNNs, discuss various hardware platforms and architectures that support DNNs, and highlight key trends in reducing the computation cost of DNNs either solely via hardware design changes or via joint hardware design and DNN algorithm changes. It will also summarize various development resources that enable researchers and practitioners to quickly get started in this field, and highlight important benchmarking metrics and design considerations that should be used for evaluating the rapidly growing number of DNN hardware designs, optionally including algorithmic codesigns, being proposed in academia and industry. The reader will take away the following concepts from this article: understand the key design considerations for DNNs; be able to evaluate different DNN hardware implementations with benchmarks and comparison metrics; understand the tradeoffs between various hardware architectures and platforms; be able to evaluate the utility of various DNN design techniques for efficient processing; and understand recent implementation trends and opportunities.

2,391 citations


"Real-Time Event Classification in P..." refers background or methods in this paper

  • ...However, rectified linear unit (reLU) and exponential linear unit (eLU) activation functions enable fast network training with less complexity and introduces sparsity in the network, thus preferred for DNN training in method [27]....

    [...]

  • ...DNN in comparison to a conventional neural network has large number of hidden layers that gives it ability to learn high level features with complexity and abstraction, thereby show superior learning and execution accuracy [27]....

    [...]

  • ...numerical or categorical outputs from an output layer [27]....

    [...]

  • ...are also used in DNN training to achieve high classification accuracy [27]....

    [...]

Journal ArticleDOI
TL;DR: A new adaptive kernel density estimator based on linear diffusion processes that builds on existing ideas for adaptive smoothing by incorporating information from a pilot density estimate and a new plug-in bandwidth selection method that is free from the arbitrary normal reference rules used by existing methods.
Abstract: We present a new adaptive kernel density estimator based on linear diffusion processes. The proposed estimator builds on existing ideas for adaptive smoothing by incorporating information from a pilot density estimate. In addition, we propose a new plug-in bandwidth selection method that is free from the arbitrary normal reference rules used by existing methods. We present simulation examples in which the proposed approach outperforms existing methods in terms of accuracy and reliability.

1,410 citations


"Real-Time Event Classification in P..." refers background or methods in this paper

  • ...The optimal value of HD for minimized AMISE [24] is: H∗D = ( E [ γ−1(X) ] 2M √ πθ2D )2/5 (16) To obtain the optimal value of HD, first the EO is determined from bandwidth estimates with Gaussian KDE at ith iteration [24] using: Hig ∼= ⎛ ⎝0.9 ( 1 + 1 2i+0.5 ) (1 × 3 × 5 . . .× (2i − 1)) 3M √ π/2.θ (i)g ⎞ ⎠ 1 3+2i (17) where ith estimate of Gaussian plug-in estimator θ ig is obtained as: θ ig = −1i N2 M∑ k=1 N∑ j=1 κg ( Xk,Xj; 2Hi−1g ) (18) From the estimated Hig, EO and PIE values are obtained for ith iteration....

    [...]

  • ...where γ (x) = √a(x)/p(x) [24] is the diffusion coefficient that affects the DKDE smoothing, X = [x1, x2, ....

    [...]

  • ...The optimal value of HD for minimized AMISE [24] is:...

    [...]

  • ...The process is repeated till the optimization criteria of (ei+1 −ei) ≤ τ is satisfied [24]....

    [...]

  • ...DKDE uses the solution of a linear diffusion partial differential equation (PDE) (8) of a Gaussian kernel [24]....

    [...]

Journal ArticleDOI
TL;DR: In this article, a new adaptive kernel density estimator based on linear diffusion processes is proposed, which builds on existing ideas for adaptive smoothing by incorporating information from a pilot density estimate.
Abstract: We present a new adaptive kernel density estimator based on linear diffusion processes. The proposed estimator builds on existing ideas for adaptive smoothing by incorporating information from a pilot density estimate. In addition, we propose a new plug-in bandwidth selection method that is free from the arbitrary normal reference rules used by existing methods. We present simulation examples in which the proposed approach outperforms existing methods in terms of accuracy and reliability.

1,410 citations