A Bayesian Learning Based Scheme for Online Dynamic Security Assessment and Preventive Control

The traditional security-constrained optimal power flow (SCOPF) model under the classical N-1 criterion is implemented in the power industry to ensure the secure operation of a power system. However, with increasing uncertainties from renewable energy sources (RES) and loads, the existing SCOPF model has difficulty meeting the practical requirements of the industry. This paper proposed a novel chance-constrained preventive SCOPF model that considers the uncertainty of power injections, including RES and load, and contingency probability. The chance constraint is used to constrain the overall line flow within the limits with high probabilistic guarantees and to significantly reduce the constraint scales. The cumulant and Johnson systems were combined to accurately approximate the cumulative distribution functions, which is important in solving chance-constrained optimization problems. The simulation results show that the model proposed in this paper can achieve better performance than traditional SCOPF.

In the following work, A machine learning-based Online Dynamic security assessment (DSA) scheme is proposed for which Energy-based Kinetic energy terms are applied as input features. The aim is to develop simple pre-processed input features that are easier to calculate and when applied as inputs, require fewer samples of data to develop a trained model. The model is trained using the current system topological condition. The proposed scheme is implemented using a Bayesian neural network (BNN), which is developed by the Bayes-by-Backprop algorithm. The performance of the proposed input features is demonstrated with IEEE 39 bus test system. When the Energy-based kinetic energy terms are applied as input features it is shown that the model can be trained with fewer samples of data.

The paper reports on the activity of an EU funded research project named OMASES that concerns a preventive approach to dynamic security assessment (DSA) and control of large power systems. The aims of the project are presented. A short description of the hardware and software architecture for implementing on-line DSA are recalled. Some hints about the theoretical background for the specific applications within OMASES (TSA-transient stability assessment, VSA-voltage stability assessment, TS-training simulator and MS-market simulator) are described. Preliminary results of the prototyping phase are commented upon.

On-line dynamic security assessment offers several significant engineering and financial benefits. However, Current EMS systems can not perform online dynamic security assessment (DSA) to ensure the ability of the power system to withstand some contingencies. We have designed and implemented a reliable and effective system, called BCU-DSA, for online dynamic security assessments and energy margin calculations of practical power systems. BCU-DSA system achieves high level of reliability and accuracy by effective exploration of the merits of both the improved BCU classifiers and the detailed time-domain simulation program. The system has been evaluated on practical power system models with promising results. The BCU-DSA system gives exact stability/instability assessments and accurate energy margin for every contingency of practical power systems.

With the development of synchronised measurement technique, online dynamic security assessment (DSA) is of great significance to prevent power system blackout. Recently, based on the phasor measurement unit (PMU) data, intelligent data‐driven techniques have been rapidly developed for online DSA owing to their fast decision speed, less data requirement, and decision rule discovery ability. The interpretable decision rules provide useful information for preventive control and post‐event auditing. However, in case of incomplete measurement events, such as PMU failure, communication loss, and phasor data concentrator failure, the accuracy and the transparency of the intelligent models can be significantly impaired by such incomplete data. To overcome this problem, this paper proposes a robust white‐box model that can sustain DSA accuracy and survive the model interpretability and transparency against incomplete PMU data. The proposed model is tested on New England 39‐bus system and demonstrates higher robustness over existing methods.

A new dynamic security assessment framework based on semi-supervised learning and data editing

The decarbonization of energy systems has posed unprecedented challenges in system complexity and operational uncertainty that render it imperative to exploit cutting-edge artificial intelligence (AI) technologies to realize real-time, autonomous power system operation and control. In particular, deep reinforcement learning (DRL)-based approaches in power systems are extensively studied and implemented in several trials worldwide. Nevertheless, the vulnerability of DRL brings new security threats to power systems that have not been well identified and investigated in the literature. To this end, this paper proposes a physics-constrained vulnerability assessment methodological framework for the DRL-based power system operation and control, with a special focus on the problem of security-constrained optimal power flow (SCOPF). In particular, we develop a novel adversarial example generation method, defined as a false data injection attack against the DRL-based SCOPF (FDIAI), to realize a targeted adversarial attack considering the nonlinear physical constraints in power systems via two main stages of constructor function design and unconstrained optimization problem transformation. In this way, the proposed FDIAI can significantly influence the decision-making procedure of DRL while successfully evading the bad data detection mechanism in power systems. Case studies are conducted to explore the stealthiness and effectiveness of FDIAI and then show its severe impacts on system operation and control on the winners' models of the Learning to Run a Power Network (L2RPN) competitions, including L2RPN IJCNN 2019 (IJCNN), L2RPN WCCI 2020 (WCCI), and L2RPN NeurIPS 2020 (NeurIPS).

In recent years, as power systems have become larger, power oscillations have become multifaceted, and so online Dynamic Security Assessment (DSA) has become essential to estimate these low-frequency electromechanical modes. To use this information for stabilization control of angle stability, it is important to obtain information on not only the frequency and damping of those modes but also the relative phase relation between each generator. This paper introduces a technique to estimate the relative phase relation of generators from time series data. The proposed method is applied to a two-variable ARMA model to estimate the low-frequency modes, and is applied to the cross spectrum to estimate the eigenvector efficiently. The proposed technique is verified by numerical examples with the IEEE 39-Bus system and the West 30-machine system model (115-Bus) which simulates the power system of Japan, and is compared with the eigenvector estimated by eigenvalue analysis.

With the increasing integration of renewable energy resources and other forms of dispersed generation, more and more variances and uncertainties are brought to modern power systems. The dynamic security assessment (DSA) of modern power systems is facing challenges in ensuring its accuracy for unpredictable operating conditions (OC). This paper proposes a novel approach that uses random forest (RF) for online DSA. Hourly scenarios are generated for the database according to the forecast errors of renewable energy resources, which are calculated from historical data. Fed with online measurement data, it is able to not only predict the security states of current OC with high accuracy, but also indicate the confidence level of the security states one minute ahead of the real time by an outlier identification method. The results of RF together with outlier identification show high accuracy in the presence of variances and uncertainties due to wind power generation. The performance of this approach is verified on the operational model of western Danish power system with around 200 transmission lines and 400 buses.

The increasing complexity of power systems as well as the rapid expansion of renewable energy resources and deregulation of electricity industry lead to continual demand for operating power systems at higher efficiencies. The modern power systems are pushed to operate close to their secure limits. When subjected to a severe contingency, the stressed power systems may lose its dynamic security, resulting in catastrophic consequences such as cascading failure and even wide-spread blackout. Under this new reality, online and real-time power system stability assessments have great impacts on avoiding blackout events. The traditional stability methods are mainly based on time-domain simulation, which are not fast enough to satisfy the online and real-time application requirements. In recent years, with the development of advanced real-time system monitoring and communication techniques, the data-driven methods have been identified as powerful tools to help power systems counteract large disturbance and avoid instability. To solve above threats, this Ph.D. research aims to develop data-analytics methodologies from power grid measurement data to recognize the system stability, covering both pre-fault dynamic security assessment (DSA) and post-fault real-time short-term voltage stability (STVS) assessment areas under some extreme scenarios and practical issues. Specific objectives of this Ph.D. research are to develop an intelligent system for the future smart grid. Based on several kinds of promising advanced data-analytics technologies, including randomized learning, incremental broad learning (BL), transfer learning (TL), adversarial machine learning (ML), and generative adversarial network (GAN), etc, an intelligent system will be developed to interpret the power grid stability characteristics, and to support stability-constrained operation control actions. Hence, it can make power systems achieve seven goals, including extensibility, efficiency, robustness, security, personalization, reliability, and effectiveness. The outcomes of this Ph.D. research will enable a more secure, intelligent, and knowledge-transparent future of the next-generation smart grid. The main contributions through this Ph.D. research are as follows: Firstly, the temporal-adaptive data-driven methods are proposed based on optimized hybrid ensemble learning mechanism for online STVS assessment. Compared to the traditional data-driven stability methods, the proposed methods can balance the tradeoff between the stability accuracy and speed and select the most suitable Pareto set. Secondly, in the area of DSA, an online updating DSA method is developed by using incremental BL technique. The incremental BL method only needs to calculate the pseudoinverse of new state matrix without the computations of the whole output weights matrix. In this way, it does not need re-training the whole model to continuously maintain and improve the online DSA accuracy. Besides, the proposed method can achieve under three different increment scenarios, including increment of enhancement hidden nodes, features, and new training instances. Thirdly, in the area of missing phasor measurement unit (PMU) data, two different fully robust data-driven stability methods are developed, respectively for pre-fault DSA and post-fault STVS assessment. The former one is based on GAN without observability-constrained feature subsets to ensure its availability under any PMU missing scenario. The latter one is based on super-resolution perception and incremental BL, which can achieve both satisfactory stability accuracy and speed. Fourthly, in order to accurately quantify the robustness of the ML-based model under adversarial perturbations on safety-critical power system stability problem, an adversarial robustness verification method for data-driven stability models is proposed. Besides, a model-free and attack-independent robust index are defined to evaluate the ability of the ML-based stability models against adversarial examples, which can be used to select the candidate model and provide formal robustness guarantee for stability application in practice. Rigorous mathematic proof is provided for the robust index under both linear and non-linear binary scenarios. Moreover, the proposed robust index can be used to mitigate the adversarial examples via adversarial training. Finally, three different TL-based DSA methods are proposed to use one known fault to evaluate other related but different unknown faults. The first two methods aim to reduce the distribution differences between the trained data and unknown data with the complete unlearned faults. The third one considers the different faults with incomplete data scenario. Through applying the adversarial training process, the DSA model trained by one fault can work for a different but related fault by confusing domain discriminator. With the property of feature extractor network, the proposed integrated TL-based method can solve the missing data issue for the unknown faults. All the methodologies have been tested on benchmark power systems to verify their effectiveness, and comparative and comparative studies with existing methods in the literature are conducted where applicable.

This paper presents a methodology based on Artificial Neural Network (ANN) structures for the dynamic security assessment (DSA) of power systems. Proposed methodology involves, ANN approach for fast and accurate estimation of critical clearing time (CCT) values of credible faults occurring in the system, considering changes in the loading conditions and system topology. CCT is an important indicator that measures the transient stability of the system against critical contingencies. Offline trained ANNs can monitor online DSA without suffering from excessive computational burden of time domain simulations (TDS). Decision Trees are used as a feature selection tool to reduce the training time and ANN complexity, increasing the CCT estimation performance of the ANN applications studied in this work, Multi-Layer Perceptron, Radial Basis Neural Network, Generalized Regression Neural Network and Adaptive Neuro-Fuzzy Inference Systems. The proposed approach is applied to 16 generator-68 bus test system operating at various loading conditions and system topologies.

The load margin due to voltage instability and small-signal instability can be a valuable measure for the operator of the power system to ensure a continuous and safe supply of electricity. However, if this load margin was calculated without considering system operating requirements, then this margin may not be adequate. This article proposes an algorithm capable of providing the power system load margin considering the requirements of voltage stability, small-signal stability, and operational requirements, as limits of reactive power generation of synchronous generators in dynamic security assessment. Case studies were conducted in the 107-bus reduced order Brazilian system considering a list of contingencies and directions of load growth.

This chapter discusses the issues associated with the successful design and implementation of an online dynamic security assessment system. The need for, the benefits of, and the current application status of such systems are presented. The discussions are primarily focused on the practical issues for implementing an online Dynamic Security Assessment (DSA) system, based on the experiences gained from recent applications. Finally, an overview of some future directions on the online DSA technologies is provided.

The intermittence and fluctuation of renewable energy bring significant uncertainty to the power system, which enormously increases the operational risks of the power system. The development of efficient interval prediction models can provide data support for decision making and help improve the economy and reliability of energy interconnection operation. The performance of Bayesian deep learning models and Bayesian shallow neural networks in short-term interval prediction of photovoltaic power is compared in this study. Specifically, an LSTM Approximate Bayesian Neural Network model (ABNN-I) is built on the basis of the deep learning and Monte Carlo Dropout method. Meanwhile, a Feedforward Bayesian Neural Network (ABNN-II) model is introduced by Feedforward Neural Network and the Markov Chain Monte Carlo method. To better compare and verify the interval prediction capability of the ABNN models, a novel clustering method with three-dimensional features which include the number of peaks and valleys, the average power value, and the non-stationary measurement coefficient is proposed for generating sunny and non-sunny clustering sets, respectively. Results show that the ABNN-I model has an excellent performance in the field of photovoltaic short-term interval forecasting. At a 95% confidence level, the interval coverage from ABNN-I to ABNN-II can be increased by up to 3.1% and the average width of the interval can be reduced by 56%. Therefore, with the help of the high computational capacity of deep learning and the inherent ability to quantify uncertainty of the interval forecast from Bayesian methods, this research provides high-quality interval prediction results for photovoltaic power prediction and solves the problem of difficult modeling for over-fitting that exists in the training process, especially on the non-sunny clustering sets.

Artificial neural networks (ANNs) have found many potential applications in power systems operation and control recently. This paper presents a categorization of the main significant applications of neural networks, which includes power system controller design, power system security assessment and load forecasting. It is desired that they are helpful to the construction of more efficient, robust ANNs to solve a broader range of problems in power systems.