Search

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]

Due to the development of new technologies, change of generation mix and appearance of newly formed energy supply hubs, there is a large year‐on‐year change in the marginal loss factors in power systems. Since any change of marginal loss factors could have significant impacts on payment of loads and profitability of generators, it is necessary to carry out a comparative study on the loss factor‐based locational marginal pricing methods. Considering that a systematic comparison of various locational marginal pricing methods has not been reported in existing publications, this work presents a comparative study of the loss factor‐based locational marginal pricing methods that are widely adopted in electricity markets. Advantages and disadvantages of each locational marginal pricing method are explored in detail, and could serve as references in selecting appropriate locational marginal pricing methods in practice. The selected five locational marginal pricing models are tested in two standard power systems, that is, the IEEE 5‐bus and 39‐bus systems. Then, through numerical experiments and detailed analysis, key findings about the reference point dependency of loss factors, accuracy of loss estimation, load payment, generation income, and market settlement surplus are summarised and elaborated. It is found that marginal loss factors‐based locational marginal pricing methods tend to produce a higher market settlement surplus and can lead to a lower generation income than other locational marginal pricing methods. [ABSTRACT FROM AUTHOR]

With a large number of inverter‐interfaced renewable power generation, fault‐induced delayed voltage recovery (FIDVR) events have become a serious threat to power system stability assessment. This study proposes a novel data‐driven method based on probabilistic prediction, ensemble learning, and multi‐objective optimisation programming (MOP) to rapidly predict the FIDVR severity index for real‐time FIDVR assessment. Distinguished from the existing single machine learning (ML) algorithm data‐driven method, the proposed method combines different randomised learning algorithms to acquire a more diversified ML outcome. The probabilistic prediction models the uncertainties existing in the prediction process, which quantifies the prediction confidence over a progressive observation window. Besides, the FIDVR can be evaluated through the time‐adaptive framework to achieve the best FIDVR speed and accuracy with the MOP framework. The simulation results on the New England 10‐machine 39‐bus system display its preponderance over the single ML, and also demonstrate its better speed and accuracy performance in FIDVR assessment. [ABSTRACT FROM AUTHOR]

High‐level wind power penetration has significantly changed the power system's static and dynamic characteristics, which tends to decrease the overall system transient instability. This study proposes a multi‐objective preventive‐emergency coordinated control method against transient instability in the presence of uncertain wind power generation. This method optimally coordinates preventive control (PC) and emergency control (EC) to simultaneously minimise the total control cost and the expected stability margin. PC includes synchronous generator dispatch and wind farm power curtailment to reduce the overall percentage of renewable power to a safe level. EC adopts an emergency demand response scheme, in which the demand is immediately reduced for stabilising post‐contingency system and avoiding wide‐spread blackout. Wind power uncertainty is modelled through Taguchi's testing scenarios. The proposed method was verified on both New England 39‐bus system and Nordic‐32 system using industry‐grade dynamic models and simulation software. Simulation results verified that the proposed method can optimally balancing system control cost and system expected stability margin. [ABSTRACT FROM AUTHOR]

In this study, an advanced control strategy is proposed for hybrid energy storage systems (HESS) to smooth wind power generation fluctuations. Compared with the limited performance of solo energy storage system, the HESS, composing of lithium‐ion battery (LiB) and a flywheel energy storage system (FESS), can comparatively show improved flexibility and adaptivity. A novel sequence control scheme for the HESS is proposed in this study to improve the overall economic and smoothing performances. Specifically, based on variable‐interval reference outputs, the control horizon of the HESS is statistically determined. Afterwards, a sequence control scheme for the HESS is presented to improve the internal collaboration of HESS media. Especially when the LiB and FESS are synchronously charging/discharging to smooth wind power, the charge/discharge power are optimally distributed between the LiB and FESS via an optimiser to minimise the equivalent cost during each control cycle. Case studies are conducted to demonstrate the performance of the proposed method. [ABSTRACT FROM AUTHOR]

Shared electric vehicle (EV) is an emerging component across interdependent power and transportation networks that affect economic, environmental and societal benefits. Studies on how to collaboratively enable the optimal planning of these interacted networks is less explored. In this work, a joint distribution network expansion planning framework integrated with a shared EV charging station is proposed. One of the objectives is to minimise overall investment cost considering operational waiting cost of shared EV charging service and power loss, while the other aims to maximise the utilisation of charging infrastructures. To overcome the difficulties in solving the multi‐objective mixed‐integer non‐linear problem, Tchebycheff decomposition based evolutionary algorithm is modified to find the Pareto front showing the trade‐offs between goals above. Besides, stochastic traffic assignment model is used to obtain traffic distribution and travel demand. The final optimal solution is decided by the fuzzy satisficing method. The performance of the proposed approach is evaluated on a 54‐node test system. Sensitivity analysis is performed to assess the impact of key parameters from the perspectives of transportation and distribution network sectors. The numerical results demonstrate the capability and feasibility of the proposed method. [ABSTRACT FROM AUTHOR]

With development of two‐way communication technology, residential users are able to reshape their energy consumption patterns based on demand response signals. This study proposes an optimal residential energy resource scheduling model to minimise the home electricity cost while fully considering the user's life convenience, the user's thermal comfort, and renewable uncertainties. The proposed model accounts for the characteristics of shiftable appliance, air‐conditioning system, electric vehicle's charging pattern, and renewable generation of both wind and solar power. Wasserstein distance metric and K‐medoids‐based scenario generation and reduction techniques are used to address the renewable uncertainty. An adaptive thermal comfort model is employed to estimate the user's indoor thermal comfort degree. A waiting cost model is applied to measure the user's preference on the household appliance's operation. In addition, a recently proposed metaheuristic optimisation algorithm (the natural aggregation algorithm) is used to solve the proposed model. The simulation results show the proposed model is effective in minimising the household's daily electricity bill while preserving the user's comfort level. [ABSTRACT FROM AUTHOR]

Voltage/VAR regulation (VVR) implemented by capacitor banks (CBs), on‐load tap changers (OLTCs), and photovoltaic (PV)‐associated inverters is effective to enhance voltage stability and reduce power loss. On the other hand, conservation voltage reduction (CVR) has been widely adopted in distribution networks to reduce load demand. Existing works mainly focus on each of them. This paper proposes a VVR method with CVR, forming a multi‐objective optimisation problem which minimises (i) voltage collapse proximity indicator (VCPI), (ii) load demand, and (iii) power loss. Besides, PV power generation as uncertainty significantly impairs the VVR results. To deal with the uncertainty issue, this paper proposes a rolling‐horizon framework to determine VVR and applies Taguchi's orthogonal array testing (TOAT) to model the uncertain PV power generation, achieving rolling‐horizon‐based multi‐objective robust VVR. The proposed model is tested and demonstrated on the IEEE 33‐bus and 69‐bus systems, and simulation results verify that the proposed method can support robust solutions of the multi‐objective VVR problem for decision‐making. [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

Advances in metering and two‐way communication technologies foster the studies of Home Energy Management System (HEMS). This study proposes a new HEMS, which optimally schedules the distributed residential energy resources (DRERs) in a smart home environment with varying electricity tariff and high solar penetrations. The uncertainties of solar power output are captured by using Monte Carlo sampling technique to generate multiple solar output scenarios based on the probabilistic solar radiation model. The homeowner's rigid and elastic restrictions on the operations of the automatically controlled household appliances are modelled. Based on this, an optimal DRER scheduling model is proposed to minimise the home operation cost while taking into account the homeowner's requirements. A new heuristic optimisation algorithm recently proposed by the authors, i.e. natural aggregation algorithm, is used to solve the proposed model. Simulations based on real Australian solar data are conducted to validate the proposed method. [ABSTRACT FROM AUTHOR]

This study presents a two‐stage planning framework of the battery energy storage system (BESS) and micro‐turbine (MT) in a microgrid. In the first stage, the optimal allocation decisions are made before the actual realisation of the operational uncertainties. In the second stage, the optimal operation strategies are made for the microgrid by minimising the costs paid to the main grid, fuel cells, MTs, BESSs and controllable loads (CLs). The hot water system and interruptible load are considered as CLs. Their mathematical models are built to investigate their roles in smoothing renewable energy. In addition, efforts are made to keep the linearity of the formulated optimisation problem, and the backward scenario reduction method is adopted to further enhance the computational efficiency. The modified IEEE 33‐bus radial system is used as a microgrid to verify the effectiveness of the proposed approach for both islanded and grid‐connected microgrids. Sensitivity analysis has been conducted to compare the economics and robustness of the identified solutions. [ABSTRACT FROM AUTHOR]

Wide‐area measurement system (WAMS) has been widely deployed in modern power systems to achieve data‐driven dynamic security assessment (DSA). Considering the unavailability of measurement data during unintentional events, a robust classification model for phasor measurement unit (PMU)‐based DSA is proposed. The proposed approach identifies the observability of different PMU combinations in the system and an ensemble of random vector functional link networks is trained with observability‐constrained feature subsets to ensure its viability under any PMU missing condition. Meanwhile, the available operating features are comprehensively utilised by the proposed classification model to support accurate DSA under both normal and abnormal data collection conditions. The proposed classification model is tested on New England 39‐bus system with high accuracy and strong robustness. [ABSTRACT FROM AUTHOR]

Deployment of battery energy storage (BES) in active distribution networks (ADNs) can provide many benefits in terms of energy management and voltage regulation. In this study, a stochastic optimal BES planning method considering conservation voltage reduction (CVR) is proposed for ADN with high‐level renewable energy resources. The proposed method aims to determine the optimal BES sizing and location to minimise the total investment and operation cost considering energy saving achieved by CVR, while satisfying system operational constraints in the presence of stochastic renewable power generation. The uncertainty of load composition is also modelled through scenario analysis. The proposed planning scheme is tested in a modified IEEE 15‐bus system and 43‐bus radial system, respectively. The numerical results validate that the combination of CVR and BES can achieve more energy savings. [ABSTRACT FROM AUTHOR]

The increasing penetration of wind power (WP) and demand response (DR) programs into modern power systems poses more challenges on transmission expansion planning (TEP). To ensure the economical, secure and reliable operations of power systems, this study presents a risk‐averse TEP framework. Instead of using the deterministic security criterion, an insecurity risk cost (RC) is proposed to provide network planners with the insight into the problem, options and future implications in decision making. Specifically, this RC can quantify the system security degree, considering the probability and the severity of contingencies. Meanwhile, the economic value of DR is modelled and incorporated into the optimal operation solutions. Moreover, to enhance the computational efficiency, an iterative solution algorithm based on the Benders decomposition is developed to solve the formulated TEP problem. The proposed approach is numerically verified on the Garver's 6‐bus, IEEE 24‐bus RTS, and 2383‐bus polish systems. Case study results demonstrate that the proposed approach can effectively investigate the impacts of large‐scale integration of WP and DR on system operations and planning. Moreover, the proposed risk‐averse approach is economically efficient and more robust to stochastic variations. [ABSTRACT FROM AUTHOR]

For cross‐regional transmission projects, the two‐part transmission pricing mechanism is suggested so as to promote the sustainable development of cross‐regional electricity trading. In the two‐part transmission pricing mechanism, appropriately determining the capacity charging ratio (CCR) is an important issue not well solved. Given this background, a Z‐number‐based risk‐minimised negotiation model is developed for a transmission company and a power purchaser to achieve an agreeable CCR under incomplete information. The uncertainty distribution of the future annual electricity transmission quantity is first estimated by the Z‐number‐based multiple Z‐valuations; and then, the benefit and risk loss measured by the well‐established conditional value at risk (CVaR) are analysed for the participating two parties. Subsequently, the negotiation model where each negotiator is to minimise its risk loss under a given lowest acceptable benefit constraint and the estimations of the opponent's risk tolerance and negotiation strategy is presented to determine the optimal offer. Finally, the ± 500 kV Xiluodu−Guangdong direct current (DC) transmission project in the southern region of China is employed to demonstrate the basic characteristics of the proposed model. [ABSTRACT FROM AUTHOR]

Demand side management is one of the key topics of smart grids. This study integrates the service computing paradigm in smart grid domain and proposes a demand side personalised recommendation system (PRS). The proposed PRS employs service recommendation techniques to infer residential users' potential interests and needs on energy saving appliances, and then it recommends energy saving appliances to users, therefore potentially creating opportunities to save energy for the grid. The proposed approach starts by applying a non‐intrusive appliance load monitoring (NILM) method based on generalised particle filtering to disaggregate the end users' household appliance utilisation profiles from the smart meter data. Based on the NILM results, several inference rules are applied to infer the preferences and energy consumption patterns, and to form the user profile. In parallel, information retrieval techniques are applied to extract keywords from the textual appliance advertisements (Ads), and to define the appliance profile. Finally, the similarity measurement method is applied to compare the user profile and appliance profile, to rank the appliance Ad, and to make the recommendations. Experiments are conducted to validate the proposed system. [ABSTRACT FROM AUTHOR]

The integration of large‐scale renewable energy and demand response (DR) resources in smart grids have brought in emerging challenges for transmission expansion planning (TEP), particularly in terms of system security. The conventional TEP models have not fully addressed the cost and the feasibility of corrective control (CC) actions such as generation rescheduling and load curtailment under contingencies. Moreover, the optimality of CC depends on the pre‐contingency state, the post‐contingency state, as well as the existence and viability of the involved CC actions. In this study, first the authors have given the explicit definition of CC risk index (CCRI), which evaluates the expected system performance under a set of contingencies (i.e. risk of incurring security issues). With the authors' improvement, the CCRI is now mathematically tractable and may have wide applications to TEP problems. Afterwards, the authors have proposed a multi‐objective TEP framework with tradeoffs between cost and risk. A relatively new yet superior multi‐objective evolutionary algorithm called the multi‐objective evolutionary algorithm (MOEA)/D is introduced and employed to find Pareto optimal solutions. The proposed model is numerically verified on the modified IEEE RTS 24‐bus and 118‐bus systems. According to the simulation results, the proposed model can provide information regarding variants of risks and coordinate the optimum planning and DR solutions. [ABSTRACT FROM AUTHOR]

Conventional approaches for measurement‐based load modelling uses measured voltage (V) as the input to calculate the load model output – P and Q powers. This assumes that the P and Q response is a function of V. In fact, there is an inherent interaction among the three variables (P, Q, and V) especially when motor load occupies a large proportion. With increased penetration of wind power, the fault‐induced dynamic voltage response is becoming a serious concern given the low‐voltage‐ride‐through (LVRT) requirement of wind turbines. This paper firstly shows that given different load model parameters, the V responses can vary significantly. Then, an improved method is proposed for more accurate measurement‐based load modelling. The proposed method incorporates the V response into the load model output, therefore is able to accurately reflect the dynamic voltage trajectories. Given the load model parameters, the system responses of P, Q, and V are all simulated via time‐domain simulations using industry‐grade software, and problem is to search the load model parameters to minimise the fitting error between the simulated and measured system responses. A trajectory sensitivity index is used to identify the well‐conditioned parameters, and a parallel‐differential evolutionary algorithm is designed to solve the model. [ABSTRACT FROM AUTHOR]

Static compensators (STATCOMs) are able to provide rapid and dynamic reactive power support within a power system for voltage stability enhancement. While most of previous research focuses on only an either static or dynamic (short‐term) voltage stability criterion, this study proposes a multi‐objective programming (MOP) model to simultaneously minimise (i) investment cost, (ii) unacceptable transient voltage performance, and (iii) proximity to steady‐state voltage collapse. The model aims to find Pareto optimal solutions for flexible and multi‐objective decision‐making. To account for multiple contingencies and their probabilities, corresponding risk‐based metrics are proposed based on respective voltage stability measures. Given the two different voltage stability criteria, a strategy based on Pareto frontier is designed to identify critical contingencies and candidate buses for STATCOM connection. Finally, to solve the MOP model, an improved decomposition‐based multi‐objective evolutionary algorithm is developed. The proposed model and algorithm are demonstrated on the New England 39‐bus test system, and compared with state‐of‐the‐art solution algorithms. [ABSTRACT FROM AUTHOR]

A new approach for power system transient stability preventive control is proposed by performing trajectory sensitivity analysis on the one‐machine‐infinite‐bus (OMIB) equivalence of multi‐machine systems. Exact instability time/angle are determined from the equivalent OMIB power–angle curve; the trajectory sensitivity is calculated at the instability time and the transient stability of the multi‐machine system is controlled by constraining the OMIB's angle excursion at the instability time to that of the critical OMIB which corresponds to the marginally stable condition of the system. The required preventive control action (generation rescheduling) can be efficiently solved via a linear programming model with the OMIB trajectory sensitivities‐based constraints. Simulation results on the New England 10‐machine 39‐bus system and a 285‐machine and 1648‐bus system validate its effectiveness and superiority over previous trajectory sensitivity applications to this problem. [ABSTRACT FROM AUTHOR]

Intelligent system (IS) using synchronous phasor measurements for transient stability assessment (TSA) has received continuous interests recently. For post‐disturbance TSA, one pivotal concern is the response time, which was reported in the literature as a fixed value ranging from 4 cycles to 3 s after fault clearance. Since transient instability can develop very fast, there is a pressing need for faster response speed. This paper develops a novel IS to balance the response speed and accuracy requirements. A set of classifiers are sequentially organised, each is an ensemble of extreme learning machines (ELMs), whose inputs are post‐disturbance generator voltage trajectories and outputs are the classification on the stable/unstable status of the post‐disturbance system and an evaluation of the credibility of the classification. A self‐adaptive TSA decision‐making mechanism is designed to progressively adjust the response time, such that the IS can do the classification faster, thereby allowing more time for emergency controls. The ELM ensemble classifiers can also be updated by on‐line pre‐disturbance TSA results due to its very fast learning speed. Case studies on the New England system and IEEE 50‐machine system have validated the high efficiency and accuracy of the IS. [ABSTRACT FROM AUTHOR]

Corrective security‐constrained optimal power flow (CSCOPF) considers the use of corrective control to remove system security violations in the post‐contingency state. Its optimality not only depends on the pre‐contingency state, but also the post‐contingency state as well as the involved corrective control actions. This study first gives a comprehensive review on the relevant OPF models and then proposes a hybrid method to solve the CSCOPF problem. It makes use of the evolutionary algorithms to randomly search the maximum feasible region and state‐of‐the‐art OPF solution technique (interior‐point method) to provide deterministic solutions in the found region. The two interact iteratively to progressively approach the final solution. The proposed method is verified on the IEEE 14‐bus and 118‐bus systems. Comparison studies show that (i) CSCOPF can better balance the security and economy and (ii) the hybrid method is overall superior (in solution quality, robustness and convergence characteristic) over the single evolutionary algorithm. Parallel processing is applied to speed‐up the computations. [ABSTRACT FROM AUTHOR]

Event‐driven load shedding is an effective countermeasure against voltage collapse in power systems. Conventionally, its optimisation relies on sensitivity‐based linear methods, which, however, could suffer from unrealistic assumptions and sub‐optimality. In this study, an alternative approach based on parallel‐differential evolution (P‐DE) is proposed for efficiently and globally optimising the event‐driven load shedding against voltage collapse. Working in a parallel structure, the approach consists of candidate buses selection, voltage stability assessment (VSA) and DE optimisation. Compared with conventional methods, it fully considers the non‐linearity of the problem and is able to effectively escape from local optima and not limited to system modelling and unrealistic assumptions. Besides, any type of objective functions and VSA techniques can be used. The proposed approach has been tested on the IEEE 118‐bus test system considering two cases for preventive control and corrective control, respectively, and compared with the two existing methods. Simulation results have verified its effectiveness and superiority over the compared methods. [ABSTRACT FROM AUTHOR]

Artificial Neural Network (ANN) has been recognized as a powerful method for short‐term load forecasting (STLF) of power systems. However, traditional ANNs are mostly trained by gradient‐based learning algorithms which usually suffer from excessive training and tuning burden as well as unsatisfactory generalization performance. Based on the ensemble learning strategy, this paper develops an ensemble model of a promising novel learning technology called extreme learning machine (ELM) for high‐quality STLF of Australian National Electricity Market (NEM). The model consists of a series of single ELMs. During the training, the ensemble model generalizes the randomness of single ELMs by selecting not only random input parameters but also random hidden nodes within a pre‐defined range. The forecast result is taken as the median value the single ELM outputs. Owing to the very fast training/tuning speed of ELM, the model can be efficiently updated to on‐line track the variation trend of the electricity load and maintain the accuracy. The developed model is tested with the NEM historical load data and its performance is compared with some state‐of‐the‐art learning algorithms. The results show that the training efficiency and the forecasting accuracy of the developed model are superior over the competitive algorithms. [ABSTRACT FROM AUTHOR]