Search

His current research interests include operation, management and cyber security analysis of smart grids, microgrid, smart city, electric vehicles, artificial intelligence, and machine learning. In recent years, due to the grid modernizations, high penetration of renewable energies, and using smart sensors in the main structure of the power grids, a large amount of data has been generated that can potentially lead to the complexity of the network. His current research interests include power system operation, reliability, resiliency, renewable energy sources, cybersecurity analysis, machine learning, smart grids, and microgrids. TOPIC 1: OPTIMAL OPERATION AND MANAGEMENT OF MULTI-MICROGRIDS USING BLOCKCHAIN TECHNOLOGY Paper 1 by Misagh Dehghani Ghotbabadi et al. investigates the optimal operation of a networked microgrid from the reliability perspective in a correlated atmosphere for the wind generators using an advanced machine learning technique. [Extracted from the article]

Energy forecasting plays a vital role in mitigating challenges in data rich smart grid (SG) systems involving various applications such as demand‐side management, load shedding, and optimum dispatch. Managing efficient forecasting while ensuring the least possible prediction error is one of the main challenges posed in the grid today, considering the uncertainty in SG data. This paper presents a comprehensive and application‐oriented review of state‐of‐the‐art forecasting methods for SG systems along with recent developments in probabilistic deep learning (PDL). Traditional point forecasting methods including statistical, machine learning (ML), and deep learning (DL) are extensively investigated in terms of their applicability to energy forecasting. In addition, the significance of hybrid and data pre‐processing techniques to support forecasting performance is also studied. A comparative case study using the Victorian electricity consumption in Australia and American electric power (AEP) datasets is conducted to analyze the performance of deterministic and probabilistic forecasting methods. The analysis demonstrates higher efficacy of DL methods with appropriate hyper‐parameter tuning when sample sizes are larger and involve nonlinear patterns. Furthermore, PDL methods are found to achieve at least 60% lower prediction errors in comparison to other benchmark DL methods. However, the execution time increases significantly for PDL methods due to large sample space and a tradeoff between computational performance and forecasting accuracy needs to be maintained. [ABSTRACT FROM AUTHOR]

This paper proposes a novel high-performance deep learning framework for load recognition. The framework consists of a deep-shallow model and a fast backpropagation (FBP) algorithm. In the deep-shallow model, power-related wave patterns are perceived by a convolutional neural network (CNN), and feature statistics of the power consumption data are analyzed by a sparse feed-forward neural network. The new architecture improves model interpretability and prevents information loss problems in CNNs. The architecture also greatly boosts the convergence speed and significantly enhances the test set accuracy of a neural network. Compared with conventional CNN models utilized by many load recognition applications, the FBP algorithm consisting of four sub-algorithms converges faster at the start of the training process and reduces at least 87.5% of the filter gradient computations on average. The deep-shallow-fast model that combines the deep-shallow model and the FBP algorithm attains 97.62% accuracy on the test set in the load recognition task. To fully utilize the training data, a data augmentation technique is invented that transforms the voltage and current time series into an image-like 4-D tensor. Experiments illustrate that the proposed framework is much more accurate and converges considerably faster than the conventional CNN model that many deep-learning-based load recognition applications are based upon. [ABSTRACT FROM AUTHOR]

Situation awareness (SA) has been recognized as a critical guarantee for the stable and secure operation of electric power systems, especially under complex uncertainties after renewable energy integration. In this article, an artificial-intelligence-powered solution is presented to reach a full realization of SA covering perception, comprehension, and prediction, the last of which is more advanced but challenging and hence has not been discussed in any literature before. A novel SA model is proposed by aggregating two powerful deep learning structures: convolutional neural network (CNN) and long short-term memory (LSTM) recurrent neural network. The proposed CNN-LSTM model has superiority to achieve collaborative data mining on spatiotemporal measurement data, i.e., to learn both spatial and temporal features simultaneously from phasor measurement units data. Two functional branches are designed within the SA model: a contingency locator to detect the exact fault location at present and a stability predictor to predict stability status of the system in the future. Test results have shown high performance (accuracy) of the model even on a low level of data adequacy. The proposed SA model can promisingly facilitate very fast postfault actions by the system operators to prevent the power system from any unstable operational status. [ABSTRACT FROM AUTHOR]

This paper presents a deep residual network for improving time-series forecasting models, indispensable to reliable and economical power grid operations, especially with high shares of renewable energy sources. Motivated by the potential performance degradation due to the overfitting of the prevailing stacked bidirectional long short-term memory (Bi-LSTM) layers associated with its linear stacking, we propose a concatenated residual learning by connecting the multi-level residual network (MRN) and DenseNet. This method further integrates long and short Bi-LSTM networks, ReLU, and SeLU for its activating function. Rigorous studies present superior prediction accuracy and parameter efficiency for the widely used temperature dataset as well as the actual wind power dataset. The peak value forecasting and generalization capability, along with the credible confidence range, demonstrate that the proposed model offers essential features of a time-series forecasting, enabling a general forecasting framework in grid operations. The source code of this paper can be found in https://github.com/MinseungKo/DRNet.git. [ABSTRACT FROM AUTHOR]

This paper presents a practical and effective non-intrusive load monitoring (NILM) solution to estimate the energy consumption for common multi-functional home appliances (type II appliances). Type II home appliances are usually the most challenging cases in load disaggregation because they usually have multiple power consumption states, complex state transitions, and multiple operational modes. The practicality of the proposed deep convolutional neural networks-based approach comes from the minimum prerequisite information from the previously unseen customers. That means no sub-metered information for the target appliances in the NILM service subscriber’s house is needed to provide appliance level identification and estimate under the proposed architecture. Our solution also includes a novel post-processing technique that boost the performance significantly on type II home appliances. The effectiveness of the solution is evaluated on a public dataset to allow comparison with the previous works. [ABSTRACT FROM AUTHOR]

As the power system is facing a transition toward a more intelligent, flexible, and interactive system with higher penetration of renewable energy generation, load forecasting, especially short-term load forecasting for individual electric customers plays an increasingly essential role in the future grid planning and operation. Other than aggregated residential load in a large scale, forecasting an electric load of a single energy user is fairly challenging due to the high volatility and uncertainty involved. In this paper, we propose a long short-term memory (LSTM) recurrent neural network-based framework, which is the latest and one of the most popular techniques of deep learning, to tackle this tricky issue. The proposed framework is tested on a publicly available set of real residential smart meter data, of which the performance is comprehensively compared to various benchmarks including the state-of-the-arts in the field of load forecasting. As a result, the proposed LSTM approach outperforms the other listed rival algorithms in the task of short-term load forecasting for individual residential households. [ABSTRACT FROM AUTHOR]

Residential load forecasting has been playing an increasingly important role in modern smart grids. Due to the variability of residents’ activities, individual residential loads are usually too volatile to forecast accurately. A long short-term memory-based deep-learning forecasting framework with appliance consumption sequences is proposed to address such volatile problem. It is shown that the forecasting accuracy can be notably improved by including appliance measurements in the training data. The effectiveness of the proposed method is validated through extensive comparison studies on a real-world dataset. [ABSTRACT FROM PUBLISHER]

• Static and dynamic image processing architectures are proposed for solar forecast; • Hybrid sky image-based forecasting models are effectively developed and verified; • Noticeable improvements for solar ramp event forecasting have been achieved. With the ever-increased penetration of solar energy in the power grid, solar photovoltaic forecasting has become an indispensable aspect in maintaining power system stability and economic operation. At the operating stage, the forecasting accuracy of renewables has a direct influence on energy scheduling and dispatching. In this paper, we propose a series of novel approaches based on deep whole-sky-image learning architectures for very short-term solar photovoltaic generation forecasting, of which the lookahead windows concern the scales from 4 to 20 min. In particular, multiple deep learning models with the integration of both static sky image units and dynamic sky image stream are explicitly investigated. Extensive numerical studies on various models are carried out, through which the experimental results show that the proposed hybrid static image forecaster provides superior performance as compared to the benchmarking methods (i.e. the ones without sky images), with up to 8.3% improvement in general, and up to 32.8% improvement in the cases of ramp events. In addition, case studies at multiple time scales reveal that sky-image-based models can be more robust to the ramp events in solar photovoltaic generation. [ABSTRACT FROM AUTHOR]

Most existing data-driven power system short-term voltage stability assessment (STVSA) approaches presume class-balanced input data. However, in practical applications, the occurrence of short-term voltage instability following a disturbance is minimal, leading to a significant class imbalance problem and a consequent decline in classifier performance. This work proposes a Transformer-based STVSA method to address this challenge. By utilizing the basic Transformer architecture, a stability assessment Transformer (StaaT) is developed as a classification model to reflect the correlation between the operational states of the system and the resulting stability outcomes. To combat the negative impact of imbalanced datasets, this work employs a conditional Wasserstein generative adversarial network with gradient penalty (CWGAN-GP) for synthetic data generation, aiding in the creation of a balanced, representative training set for the classifier. Semi-supervised clustering learning is implemented to enhance clustering quality, addressing the lack of a unified quantitative criterion for short-term voltage stability. Numerical tests on the IEEE 39-bus test system extensively demonstrate that the proposed method exhibits robust performance under class imbalances up to 100:1 and noisy environments, and maintains consistent effectiveness even with an increased penetration of renewable energy. Comparative results reveal that the CWGAN-GP generates more balanced datasets than traditional oversampling methods and that the StaaT outperforms other deep learning algorithms. This study presents a compelling solution for real-world STVSA applications that often face class imbalance and data noise challenges. [Display omitted] • The class imbalance issue is addressed using balanced samples generated by CWGAN-GP. • StaaT with multi-head self-attention mechanisms learns important features. • SFCM labels samples undeterminable directly by domain knowledge. • StaaT exceeds other methods, showing robustness amid class imbalances and noises. [ABSTRACT FROM AUTHOR]

Real-time industrial carbon emission estimation aims to estimate emissions more accurately to promote carbon reduction and mitigate climate change. Compared with input-output-based (IOA) analysis methods, the process-based analysis (PA) methods provide more specific information to decision-makers based on extensive detailed data. However, the required data is hard to obtain and normally contains missing data. To address these challenges, this paper proposes a novel deep learning-based carbon emission estimation framework to track the emissions of industrial customers in smart grids with smart meter data. The proposed framework encompasses three pivotal stages: data imputation, device recognition, and emission estimation—collectively referred to as DI-DR-EE. Specifically, the Data Imputation Network (DINet) based on super-resolution perception (SRP) is first introduced to recover the missing smart meter data. Then the recovered data is used to recognize the device states through the Device Recognition Network (DRNet), which thrives upon subspace blueprint separable convolutions (BSConv-S) to elevate the accuracy of device recognition with low-frequency data, all the while optimizing computational efficiency. Finally, the direct emission estimation is conducted based on the device states, and the indirect emission is estimated based on the power consumption. Case studies with five factories connected to the IEEE 57-bus system have verified the effectiveness of the proposed framework. The model training process was executed using Python with PyTorch version 1.8.1, coupled with Cuda 11.1 for accelerated computations. Results underscore that DINet and DRNet outperform established benchmarks, while DI-DR-EE remarkably maintains its capacity to attain estimations within a 10% margin of error, even when grappling with up to 90% missing meter data. • A novel deep learning-based carbon emission estimation framework is proposed to track the emissions of industrial customers. • A Data Imputation Network (DINet) is introduced to reconstruct missing data more efficiently. • A Device Recognition Network (DRNet) is proposed to identify the device states more accurately. • Real-time emission estimation is realized even with incomplete and low-frequency smart meter data. [ABSTRACT FROM AUTHOR]

Constrained optimization (CO) has made a profound impact in solving many real-world problems. Due to the high computation burden in exact solvers, data-driven CO based on machine learning techniques is recently receiving extensive research interests for its capability to solve CO problems in real time. The existing data-driven CO approaches only serve for optimization problems with rather simple constraints that can be directly incorporated into model training. However, constraints that are computationally infeasible or burdensome to evaluate are commonly experienced in realistic optimization applications, especially in the engineering sector. This paper proposes an athlete–referee dual learning system (ARDLS) for end-to-end CO with large-scale complex constraints, where an athlete model is trained as the main optimizer while a referee model is trained as a probabilistic constraint classifier to guide the athlete training. A risk-based constrained loss function is designed to fine-tune the athlete model for constraint satisfaction. A case study on electric power system emergency control application is conducted to validate the proposed ARDLS, where the testing results demonstrate the excellent capability of ARDLS to improve the likelihood of satisfying large-scale complex constraints in CO. [ABSTRACT FROM AUTHOR]

• A risk-averse deep learning method for real-time emergency load shedding. • Deep learning algorithm to enhance prediction accuracy on load shedding strategy. • A shedding amount inference model to speculate the stability after load shedding. • A risk-averse loss function to minimize the risk of load under-cutting events. • Experiments on multiple power systems to verify the practical values of the method. Emergency load shedding is an effective and frequently used emergency control action for power system transient stability. Solving the full optimization models for load shedding is computational burdensome and thus slow react to the intense system variations from the increasing renewable energy sources and the more active demand-side behavior. Other sensitivity-based methods impair the control accuracy and may not guarantee global optimality. Artificial intelligence methods, as the data-driven approaches, have recently been well-recognized for its real-time decision-making capability to tackle the system variations. The existing artificial intelligence methods for emergency load shedding are based on shallow learning algorithms and can lead to both load under-cutting and over-cutting events. However, when the loads are under-cut, the power system will be exposed to a high risk of post-control instability that can propagate into cascading events, which incurs significantly higher cost than an over-cutting event. Being aware of such unbalanced control costs, this paper proposes a risk-averse deep learning method for real-time emergency load shedding, which trains deep neural network towards the reluctance to load under-cutting events, so as to avoid the huge control cost incurred by control failure. The case studies on two renewable power systems demonstrate that, compared to the state-of-the-art methods, the proposed risk-averse method can significantly improve the control success rate with negligible increase in prediction error, ending up with lower overall control cost. The results verify the enhanced control performance and the practical values of the proposed method. [ABSTRACT FROM AUTHOR]