With the diversification of users’ energy demands, accurate load forecasting is an important prerequisite for optimal scheduling and economic operation of the system, but a single‐load forecasting ...method cannot effectively predict multi‐energy loads accurately. Therefore, this paper proposes a multi‐energy load forecasting method based on bidirectional long short‐term memory (BiLSTM) and parallel feature extraction networks. Firstly, residual network and convolutional block attention module were used to extract the spatial coupling features of multi‐energy load data. Secondly, BiLSTM is used to capture the temporal features and long‐term dependencies in the load data, and the spatial coupling features are fused to obtain non‐linear prediction results. Finally, the non‐linear prediction results and the linear prediction results obtained by using multi‐energy linear regression were linearly superimposed to obtain the final prediction results. In this paper, IES load data of Tempe Campus of Arizona State University was used to verify and compared with several existing methods, and the results showed that Weighted Mean Absolute Percentage Error decreased by more than 20%.
The method of residual network and convolutional block attention module to extract the spatial coupling features of multi‐energy load in parallel and combine with multiple linear regression for multi‐energy load prediction.
With the development of smart grid and active distribution network, the flexible load recourse would play a key role in the electricity market. In this paper, we proposed a framework that the ...distributed storage energy systems, electric vehicles, and temperature control loads are aggregated in the flexible load aggregator, trading in day-ahead energy and reserve markets. The framework is modeled as a bilevel optimization model. In the propose model, the operation problem of the FLA is modeled in upper-level problem, which is to maximize the profit of the aggregator. The biding and offering strategic of Gencos and flexible load aggregator in the independent system operator are presented in lower-level problem, which aim at improving the social benefits. Karush-Kuhn-Tucker and dual theory are used to transform the nonlinear bilevel problem to a mixed-integer linear programming of single-level model. Finally, the numerical studies based on modifying PJM-5bus power system, showing the effectiveness of the proposed framework and bilevel model.
Nowadays, the efficient and reliable protection and location schemes for MMC-MVDC (Modular Multilevel Converter-Medium Voltage Direct Current) grid are few. This paper is the first to propose a ...scheme to not only protect the feeders and the busbar, but locate the segments in MMC-MVDC grid. To improve the reliability, this paper analyzes the transient characteristics of the pole-to-pole fault and then obtains the characteristic frequency band. Based on S-transform, STCFB (S-transform characteristic frequency band) Phase of fault component is utilized to construct the identification criterion for faulty feeder and faulty segment. The whole scheme can be divided into three steps, namely, protection starting criterion, faulty feeder and busbar protection criterion, and faulty segment location criterion. Firstly, the current gradient method is utilized to quickly detect the fault and start the protection device. Secondly, the non-unit protection criterion on busbar and feeders is proposed according to STCFB Phase of the voltage and current fault component. Thirdly, according to the STCFB Phase on both sides of the feeder segment, the faulty segment can be located. A radial MMC-MVDC distribution network model was built in PSCAD/EMTDC software to evaluate the performance of the protection and location method. Simulation results for different cases demonstrate that the proposed scheme has high accuracy, good adaptability and reliability.
An active distribution network (ADN) differs from a traditional distribution network in many aspects, one of which is the integration of a large amount of distributed generation (DG), especially ...intermittent photovoltaics (PVs). The integration of intermittent PVs has both pros and cons for the distribution system. As the platform on which new techniques work and the main body of a greener future energy system, the development of an ADN has to be sustainable, need-oriented, and environmentally friendly, and the traditional technical–economic evaluation method cannot meet the requirements and provide advice in the decision-making process. Based on the concept of sustainable development, we used an ADN with the integration of a large number of distributed PVs (DGPVs) as an example and established a multi-dimensional index system to evaluate the sustainable development level (SDI) of the ADN. The analysis was based on a platform we built with consideration of the investment feasibility of the DGPVs’ investors, state and industrial policies, and their interactions with the distribution system. We first compared the development of DGPVs and the SDI of the ADN as the carrier of DGPVs under different state policies, and second, we compared the SDIs of three city ADNs with different solar resources and demand levels, but under the same state policy. The analysis results showed that different integration levels of DGPVs can be set for a city/area ADN with different solar resources and demand to achieve a comparable SDI, and a comprehensive incentive mechanism could be adopted for the development of DGPVs. In this way, the benefits of different parties can be considered at the same time and finally, the coordination of the sustainable development of multi-parties can be achieved.
Due to the development of intelligent electric devices and advanced metering infrastructures, demand response will be widely utilized in the trading of the electricity market and maintain energy ...balance of the power system. In this paper, a two-stage nested bilevel model for the optimal bidding strategy of a load agent (LA) with incentive-based demand response in day-ahead and balancing markets is proposed. In the upper-level model, the optimal trading strategy of the LA is formulated to maximize the operating profit of the LA in the day-ahead energy and balancing markets. On the other hand, the lower-level proposes the clearing market model of the independent system operator, which aim to maximize social welfare. The LA acts as a price-maker in the first-stage of the bilevel model, which is bilevel nonlinear programming (BNLP) problem. Karush-Kuhn-Tucker conditions and dual theory are used to transform the BNLP into single-level programming. The LA acts as a price-taker accepts the day-ahead energy clearing-price in the second-stage of the bilevel model, which is a bilevel mixed-integer linear stochastic programming problem. Finally, implementing the two-stage nested bilevel model on modifying the 8-bus power system demonstrates the applicability of the proposed model and analysis the sensitivity of the LA' profit to unit price and the committed reserve capacity demand of the day-ahead reserve market.
Based on analysis of construction and operation of micro integrated energy systems (MIES), this paper presents economic optimization for their configuration and sizing. After presenting typical ...models for MIES, a residential community MIES is developed by analyzing residential direct energy consumption within a general design procedure. Integrating with available current technologies and local resources, the systematic design considers a prime mover, fed by natural gas, with wind power, photovoltaic generation, and two storage devices serving thermal energy and power to satisfy cooling, heating and electricity demands. Control strategies for MIES also are presented in this study. Multi-objective formulas are obtained by analyzing annual cost and dumped renewable energy to achieve optimal coordination of energy supply and demand. According to historical load data and the probability distribution of distributed generation output, clustering methods based on
K
-means and discretization methods are employed to obtain typical scenarios representative of uncertainties. The modified non-dominated sorting genetic algorithm is applied to find the Pareto frontier of the constructed multi-objective formulas. In addition, aiming to explore the Pareto frontier, the dumped energy cost ratio is defined to check the energy balance in different MIES designs and provide decision support for the investors. Finally, simulations and comparision show the appropriateness of the developed model and the applicability of the adopted optimization algorithm.
The frequency of typhoons in China has gradually increased, resulting in serious damage to low-voltage power grid lines. Therefore, it is of great significance to study the influencing factors and ...predict the amount of damage, which contributes to enhancing wind resistance and improving the efficiency of repairs. In this paper, 18 influencing factors with a correlation degree higher than 0.75 are selected by grey correlation analysis, and then converted into six common factors by factor analysis. Additionally, an extreme learning machine optimized by an improved gravitational search algorithm, hereafter referred to as IGSA-ELM, is established to predict the damage caused to the low-voltage lines by typhoons and verify the effectiveness of the factor analysis. The results reveal that the six common factors generated by factor analysis can effectively improve the prediction accuracy and the fitting effect of IGSA-ELM is better than those of the extreme learning machine (ELM) and the extreme learning machine based on particle swarm optimization (PSO-ELM). Finally, this article proposes valid policy recommendations to improve the anti-typhoon capacity and repair efficiency of the low-voltage lines in Guangdong Province.
High-resistance clamping grounding is a common grounding mode of modular multilevel converter (MMC) based DC distribution network. Due to the current limiting function of grounding resistor, ...pole-to-ground fault can only cause limited steady-state fault current. Consequently, it is very difficult to locate fault section. Besides, there are a few research works on pole-to-ground fault recovery strategy of a DC distribution system. This study purposes a fault criterion based on DC pole-to-ground voltage amplitude and changing rate. Then, by extracting polarities of a DC feeder current travelling wave using the method of wavelet transform modulus maximum, a rapid fault section locating method is presented. Furthermore, a feasible pole-to-ground fault recovery strategy is introduced, which can reduce outage range and time on the premise of guaranteeing system stability. The performances of the proposed methodologies were validated by numerical simulation.
Due to the labyrinthine topology of the MMC-MVDC system, identification of the faulty feeder especially for a single-pole-to-ground (SPTG) fault is a challenging task. Based on the characteristics ...analysis of capacitance charge–discharge, this article obtains the typical characteristics of the faulty feeder. To improve the sensitivity, the characteristic frequency band of capacitance current is extracted by S-transform which has excellent time–frequency resolution. According to the analysis, S-transform characteristic frequency band (STCFB) of positive and negative differential current of faulty feeder is negatively correlated while the healthy feeders are positively correlated. To distinguish the faulty feeder, an identification method based on Pearson correlation coefficient (PCC) is constructed. In terms of the energy difference between the positive and negative differential current, the faulty pole can be verified. Finally, a typical radial flexible DC distribution grid model is constructed in PSCAD, and the simulation results confirm the reliability and sensitivity of the identification method.
For an innovative spiral spring energy storage system, the permanent magnet synchronous generator (PMSG) is utilized as the energy conversion device due to its simple structure, low weight and high ...torque. During power generation, the output torque and moment of inertia of the spiral spring are changing continuously and simultaneously and the parameters of the PMSG show uncertainties. Furthermore, the DC link voltage of the converter should be stable and the power injected into the grid needs to be controlled. First, the change features of the external power source and the uncertainties of the generator's internal parameters are expressed as the comprehensive disturbances, which are introduced into the dynamic model of the PMSG and also modify the dynamic model. Then, the high gain observers are utilized to estimate the comprehensive disturbances, and an improved robust backstepping control scheme integrating L2 gain and high gain observers is proposed. Secondly, the gridside inverter controller for the DC voltage loop and reactive power loop is designed based on the backstepping theory. Finally, hardware implementation is fulfilled to verify the presented algorithm. The results show that high gain observers are able to accurately estimate the external and internal interferences; the proposed control scheme can effectively suppress the external and internal interferences and guarantees output current, operating speed of the PMSG and output reactive power to correctly track respective references, and effectively stabilize the DC link voltage.