Uncertainty Of Wind Energy Research Articles

Wind power prediction based on NGO-LSTM

Abstract At present, wind energy is mainly utilized through wind power grid connections. However, because of the uncertainty of wind energy, the grid connection will inevitably have an impact on the system. In order to adjust the grid dispatching plan in real time and improve the grid’s capacity to absorb wind power, a short-term wind speed prediction method based on NGO-LSTM was proposed. Taking the measured data of wind farms as an example, the prediction accuracy of the proposed method is compared with that of existing algorithms. Experimental results show that the R2 of the NGO-LSTM coupling model is closest to 1, and the RMSE, MAE, and MAPE are all smaller than the other two models, verifying the effectiveness and advanced nature of the proposed method.

Short-term wind power prediction using a novel model based on butterfly optimization algorithm-variational mode decomposition-long short-term memory

The precise forecasting of wind power output is crucial for the integration of large-scale renewable energy generation into the power grid. It plays a vital role in ensuring safe and stable operation of power grid, diminishing fuel consumption, and minimizing environmental impacts. This research presents a machine learning prediction model that integrates intelligent optimization algorithms and data decomposition techniques for short-term wind power output forecasting. The initial phase of this research develops a prediction model utilizing variational mode decomposition and long short-term memory networks. To mitigate the randomness and uncertainty of wind energy, and improve prediction accuracy, the butterfly optimization algorithm is introduced to optimize the parameters of variational mode decomposition and long short-term memory networks. Several in-depth case studies are carried out on the actual wind power generation dataset in mainland China to confirm the feasibility and effectiveness of the proposed hybrid model. Experimental results demonstrate that the proposed model outperforms the compared model in terms of prediction accuracy across different seasons, showing good practicality and generalizability.

Multi-temporal Scale Wind Power Forecasting Based on Lasso-CNN-LSTM-LightGBM

Due to the increasingly severe climate problems, wind energy has received widespread attention as the most abundant energy on Earth. However, due to the uncertainty of wind energy, a large amount of wind energy is wasted, so accurate wind power prediction can greatly improve the utilization of wind energy. To increase the forecast for wind energy accuracy across a range of time scales, this paper presents a multi-time scale wind power prediction by constructing an ICEEMDAN-CNN-LSTM-LightGBM model. Initially, feature selection is performed using Lasso regression to identify the most significant variables affecting the forecast for wind energy across distinct time intervals. Subsequently, the ICEEMDAN is utilized to break down the wind power data into various scales to capture its nonlinear and non-stationary characteristics. Following this, a deep learning model based on CNN and LSTM networks is developed, with the CNN responsible for extracting spatial features from the time series data, and the LSTM designed to capture the temporal relationships. Finally, the outputs of the deep learning model are fed into the LightGBM model to leverage its superior learning capabilities for the ultimate prediction of wind power. Simulation experiments demonstrate that the proposed ICEEMDAN-CNN-LSTM-LightGBM model achieves higher accuracy in multi-time scale wind power prediction, providing more reliable decision assistance with the management and operation of wind farms.

Complementary operation based sizing and scheduling strategy for hybrid hydro-PV-wind generation systems connected to long-distance transmission lines

Wind and solar power are expected to play important roles in many countries to achieve carbon neutrality; however, their inherent instabilities pose significant threats to existing power grids. Because hydropower has been recognized as a viable compensatory resource for solar and wind energy uncertainties, many studies have sought to determine optimal scheduling strategies for hydro-PV, hydro-wind, and hydro-PV-wind systems. However, few studies have simultaneously considered the sizing and scheduling of large-scale hydro-PV-wind hybrid systems connected to long-distance transmission lines. Therefore, this paper develops a mathematical metric to measure the wind and solar output complementarity and incorporates it into a multi-objective sizing and scheduling model for a hybrid hydro-PV-wind system, in which resource complementarity, profit generation, system reliability, and power curtailment trade-offs are concurrently considered. The ɛ-constraint method is employed and decision makers’ attitude parameters are introduced to transform the proposed model into its equivalent single objective form. A case study in China reveals that the maximum wind and solar power output complementarity rate can be at least 0.19 for the studied hybrid hydro-PV-wind system. The incorporation of wind enables the hybrid hydro-PV-wind system to provide 4.11% more load than a hydro-PV system. In addition, if 2% of power curtailment is permitted, 1.53% more load can be served and 14.9% higher profits can be earned. To better explore the application of hybrid systems, management recommendations are provided for system operators, the power industry, and management administrations.

Simplified Approach for Wind Uncertainty Cost Functions using a Mixture of Uniform Probability Distribution

Wind energy generation is an individual source of energy considered in renewable energy resources. It is completely dependent on the natural process of air blown in the area or past forecasting of air blown data. It is considered an uncertain source of energy and less reliable in the context of real-time energy provisions. To make this wind energy more reliable and consistent, we are introducing a simplified approach for wind uncertainty cost functions based on uniform probability distributions and its evaluation for usage. We have developed analytical mathematics cost functions based on several assumptions and they are constructed using Weibull and Rayleigh probability distributions. The proposed method produces the required results to make this energy reliable and consistent with the distribution networks. Computing time and complexity are less as compared to the other methods to minimize the uncertainty of wind energy.

Robust generation expansion planning in power grids under renewable energy penetration via honey badger algorithm

Robust reliability Generation Expansion Planning (GEP) turns out to be a crucial step for an efficient energy management system in a modern power grid, especially under renewable energy employment. The integration of all such components in a GEP model makes it a large-scale, nonlinear, and mixed-variable mathematical modeling problem. In this paper, the presence of wind energy uncertainty is analyzed. Both long and short-term uncertainties are incorporated into the proposed GEP model. The first step concerns the impact of long-term wind uncertainties through the annual variations of the capacity credit of two real sites in Egypt at Zafaranh and Shark El-ouinate. The second step deals with the short-term uncertainties of each wind site. The wind speed uncertainty of each wind site is modeled by probability distribution function. Then, wind power is estimated from the wind power curve for each wind site and Monte-Carlo Simulation is performed. Fast Gas Turbine and/or Pump Hydro Storage are incorporated to cope with short-term uncertainties. Sensitivity analysis is implemented for 3, 6, and 12 stages as short and long planning horizons to minimize the total costs with wind energy penetration and emission reduction over planning horizons. Also, a novel Honey Badger Algorithm (HBA) with model modifications such as Virtual Mapping Procedure, Penalty Factor Approach, and the Modified of Intelligent Initial Population Generation is utilized for solving the proposed GEP problem. The obtained results are compared with other algorithms to ensure the superior performance of the proposed HBA. According to the results of the applicable test systems, the proposed HBA performs better than the others, with percentage reductions over CSA, AO, BES, and PSO ranging up to 4.2, 2.72, 2.7, and 3.4%, respectively.

A novel power control scheme for distributed DFIG based on cooperation of hybrid energy storage system and grid-side converter

Due to the uncertainty of wind energy, the wind power is difficult to be dispatched and may cause the voltage fluctuations for distributed network. Therefore, a novel power control scheme is proposed to manage the output power of the distributed doubly-fed induction generator (DFIG) by cooperating the hybrid energy storage system (HESS). In the proposed scheme, the grid-side converter of DFIG is controlled to manage overall output power of the DFIG/HESS hybrid system in accordance to the power dispatching command. The HESS is utilized to deal with the power difference between DFIG generation power and the power dispatching command. The HESS is embedded in the DC-link bus of DFIG and is composed of superconducting magnetic energy storage and batteries. Additionally, in order to avoid HESS from overcharging and over-discharging, the pitch angle control and power dispatching command are adjusted by considering the state of charge (SOC) of HESS. Finally, the simulation model of distributed DFIG and HESS is built on the real-time digital simulation (RTDS) platform, and the simulation results validate the effectiveness and reliability of the proposed power control scheme.

Short-term wind farm cluster power prediction based on dual feature extraction and quadratic decomposition aggregation

The intermittency and uncertainty of wind energy affect the accuracy of wind power prediction, which is not conducive to the safe and stable operation of the power system. Aiming at this problem, a short-term power prediction method for multi-Stacking wind power clusters based on dual feature extraction and quadratic decomposition aggregation is proposed. Firstly, a graph network matrix is built based on the spatial relationship of wind farm clusters to extract spatial features, and then a subset of features is filtered using a feature selection method combining filter and wrapper. The original signal is then decomposed using a quadratic decomposition method combining variational modal decomposition and swarm decomposition to reduce the volatility of the signal; the wind power signals are reconstructed using multiscale permutation entropy, and finally the reconstructed signals are fed into multi-Stacking models for information modelling depending on the frequency to predict rolling multi-step power of wind farm clusters. A real wind farm cluster arithmetic example is used for validation. The experimental results show that NRMSE, NMAE, and MAPE are improved by 15.8 %, 8.99 %, and 6.79 % compared with the traditional model. The method has important reference value for both grid scheduling and stable operation.

Calculation and optimization of inter-regional available transfer capability considering wind power uncertainty and system improvement measures

The intermittency and uncertainty of wind energy make transmission congestion management increasingly complex. In this paper, we investigate the impact of wind energy on available transmission capacity (ATC) and then explore ways to mitigate transmission congestion from the grid and load side, respectively. The impact of a stochastic variable is considered by applying Latin Hypercube Sampling (LHS) and backward curtailment techniques to generate typical scenarios. ATC is improved by the optimal allocation of Thyristor Controlled Series Compensation (TCSC) on the grid side and the use of Demand Response (DR) on the load side. The source-grid-load cooperative optimisation model for ATC is solved by an improved particle swarm optimisation (PSO) algorithm. Based on the IEEE-30 bus system, an experimental scheme is designed and analytical calculations are performed. The results show that the joint application of TCSC and DR can help to improve ATC in a comprehensive way. The work done in this paper can achieve the purpose of promoting renewable energy consumption by improving the utilisation efficiency of the transmission network without changing the network structure.

Multi-stage risk-based assessment for wind energy accommodation capability: A robust and non-anticipative method

The volatility and uncertainty of wind energy has brought great challenges to its accommodation in electric power systems. To determine a wind power trading strategy in the real-time power market, the accurate assessment of wind energy accommodation capability (WAC) is of great significance. However, most current studies investigate the wind energy accommodation potential from the perspective of dispatch, but rarely provide quantitative assessment results. Additionally, the non-anticipativity of decisions has not been considered in the assessment yet. This paper proposes a novel quantitative assessment method for WAC based on the multi-stage robust optimization (RO), which addresses the anticipativity issues in the traditional two-stage method while maintaining the decision-making process' robustness. In the assessment method, the operational risk is integrated via a tractable scheme to obtain dynamic admissible WAC boundaries. To obtain the optimal solution of the proposed multi-stage RO problem, a fast robust dual dynamic programming (FRDDP) algorithm is employed, where the adaptive techniques are developed to accelerate the computation. Numerical studies on the modified IEEE 14-Bus system and a real-world system in Zhejiang Province of China validate the effectiveness of the proposed assessment model and adaptive acceleration techniques. The simulation results demonstrate the presented method brings a 18.87% reduction in operational cost, and reduces 29.34% curtailment of wind energy. Compared with the original FRDDP, the adaptive technique significantly reduces the computational consumption by 73.34% on the real-world test system.

A D-stacking dual-fusion, spatio-temporal graph deep neural network based on a multi-integrated overlay for short-term wind-farm cluster power multi-step prediction

The uncertainty of wind energy due to its non-stationary and random nature poses a major challenge to engineers responsible for power system scheduling. In the present research, a spatio-temporal graph deep neural network is used to learn the spatial characteristics and explore the strong correlation characteristics of wind-farm clusters. The Bagging, long short-term memory (LSTM), and random forest (RF) are superimposed based on the stacking integration algorithm. By analyzing the decomposed wind electron sequence characteristics, two Stacking models are used to predict the wind power of target wind-farms respectively and then a dual stacking model fusion strategy is formed. Finally, a multi-step prediction model with the spatio-temporal characteristics of D-stacking multi-integration fusion is designed for rolling multi-step prediction of wind-farm cluster power to obtain high-precision target wind-farm power. Through the actual wind power generation measured in northwest China to conduct case studies and comparative tests, it concludes that：(1) Spatio-temporal method in this paper can effectively extract the deep spatial features of wind farm clusters. (2) Dual fusion strategy improves Stacking effectively. (3) The proposed model can obtain accurate wind power prediction results, which is superior to 14 comparative algorithms proposed by other researchers.

Optimal scheduling in demand-side management based grid-connected microgrid system by hybrid optimization approach considering diverse wind profiles

Demand side management (DSM) is one of the trending economic strategies which shifts the elastic demand to the off-peak hours from the peak hours so as to reduce the overall generation cost of the system. The work done in this paper can be categorized in three phases. In the first phase, various wind speed to power conversion mathematical models available in literature are analysed to find out the one with maximum level of wind penetration. For second phase, an economic DSM strategy is implemented to restructure the forecasted load demand model for various participation levels. In the final phase the cost-effective optimization of two microgrid distribution systems are percolated. As an optimization tool, novel hybrid CSAJAYA has been used to carry on the study. Different types of grid participating and pricing strategies along with valve point loading effect and wind energy uncertainty are considered to amplify the complexity and practicality of the study. The generation costs reduced from 3 to 5% when the forecasted demand was reformed with 20% DSM participation for both the test systems. A detailed comparison with the results from various optimization tools studied confirms the effectiveness of the proposed hybrid approach. The hybrid optimization tool presented in this paper performs better in terms of central tendencies, nonparametric statistical analysis, and algorithm execution time.

Optimal wind energy generation considering climatic variables by Deep Belief network (DBN) model based on modified coot optimization algorithm (MCOA)

According to the increase of energy consumption, wind energy penetration in power generation systems, the wind energy uncertainty and also, due to complex and nonlinear relationships of climatic parameters such as wind speed, wind direction, temperature providing a model to the more accurate forecast of energy production is essential. In this study, an optimal Deep Belief Network (DBN) model has been represented to accurately predict wind energy. The DBN network model is optimized by Modified Coot Optimization Algorithm (MCOA). This procedure uses self-adaptive weighting technique and turbulent technique to prevent trapped local optimization. These two techniques solve the optimization problem, so that it will have a more accurate prediction than the original Coot Optimization Algorithm(COA). The analysis of the modified COA metaheuristic model shows the MCOA model has a minimum of a standard deviation compared to other metaheuristic algorithms. Therefore, the MCOA metaheuristic algorithm has maximum precision and maximum reliability. Also, the simulation of wind energy generation showed, the optimal DBN technique has the best simulation compared to other models. because this model utilizes two techniques self-adaptive weight and chaotic method to refine the weakness optimization process.

Compressed Air Energy Storage Capacity Configuration and Economic Evaluation Considering the Uncertainty of Wind Energy

The random nature of wind energy is an important reason for the low energy utilization rate of wind farms. The use of a compressed air energy storage system (CAES) can help reduce the random characteristics of wind power generation while also increasing the utilization rate of wind energy. However, the unreasonable capacity allocation of the CAES system results in high capital investment and a long payback period. In order to improve the economic benefits of energy storage, this paper studies the capacity configuration of compressed air energy storage systems under the condition of wind energy uncertainty. First, the typical hourly power distribution of wind power generation was obtained using historical data. Factors such as user load demand, time-of-use price of the power grid, system investment cost, power shortage cost, and power sales revenue were considered. Then, a model was built with the charging and discharging power and gas storage capacity of the CAES system as constraints, and the maximum return on investment and the minimum volume of the gas storage tank as targets. NSGA-II and TOPSIS optimal selection methods were used to solve the problem. Finally, the model was used to optimize a power operation case. The results show that in the case of an hourly load power demand of a factory using 3.2 MW, a wind farm would need to keep four wind turbines running every day, and a compressed air energy storage system with a rated power of 1 MW and a rated capacity of 7 MW would ensure the best project benefit. In this mode, 1.24 × 103 MWh of wind abandoning power could be reduced annually, 2.6 × 104 kg of carbon emissions could be reduced by increasing energy storage within the operation cycle, and the payback period of investment would only be 4.8 years.

Data-Driven Stochastic Robust Optimization for Industrial Energy System Considering Renewable Energy Penetration

Achieving carbon neutrality has been one of the main tasks these decades. In this study, renewable energies were introduced to reduce the greenhouse gas emissions of industrial energy systems. Considering solar heat and wind energy uncertainties, a data-driven stochastic robust optimization framework was proposed. Machine learning methods were applied to derive data information: a data mining method to classify the high-volume uncertain data; a kernel-based method to construct the uncertainty sets for each data class. The stochastic robust optimization model of the industrial energy system was developed as a bilevel optimization procedure: the outer level is a two-stage stochastic programming problem to optimize the expected objective value of different data clusters, and the robust optimization is nested internally to ensure robustness. A case study on the practical industrial energy system was performed, and the results are that the total annual cost is reduced by 1 507 730 $/a and 7.62% GHG emissions are decreased by introducing renewable energies; the proposed method is superior to the traditional ones in terms of PoR (2.91%) and robustness (99.79%). The results of multiobjective optimization considering economic and environmental revenues can provide multipreference schemes for decision-makers.

Assessing the CO2-emission risk due to wind-energy uncertainty

Wind power is a significant source of renewable energy, supporting reductions in emissions. However, the intermittency of wind power will cause uncertain carbon emissions and risks greenhouse-gas emission reduction. To address this issue, we analyzed the mechanism that converts wind-energy intermittency to carbon-emission uncertainty and developed metrics to assess the consequential risks of excessive carbon emissions. According to our empirical analyses on two American electricity markets, the Electric Reliability Council of Texas (ERCOT) and Pennsylvania-New Jersey-Maryland Interconnection (PJM), we found their carbon-emission risk is significant over the coming decade. The ten-year aggregated CO2 emissions between 2020 and 2030 have a 10% chance of exceeding their expected levels by 100 megatons. Our further analysis also shows that the CO2-emission risk is induced by the uncertainty of annual generation uncertainties and volatilities in the hourly wind power supply. Particularly, the two driving forces contribute heterogeneously to the carbon-emission risk in the two studied markets. Although the annual-generation randomness causes similar degrees of carbon-emission uncertainty in the two markets, the volatility-driven risk is only significant in the PJM market. According to the risk analyses, we suggested two types of investigations that are foundations for risk management: assessing the necessary installed capacity of wind energy that limits CO2-emission risk at desired confidence levels, and exploring hours when carbon emission is sensitive to wind-power volatility. The former helps set renewable-energy targets, while the latter is critical for designing policies hiring energy storage and zero-carbon generation technologies.

Risk-averse decision under worst-case continuous and discrete uncertainties in transmission system with the support of active distribution systems

Nowadays, risk-averse management is a principal concern for transmission system (TS) operator that involve different types of uncertainty including continuous uncertainties (e.g., wind energy uncertainty) and discrete uncertainties (e.g., generator/line outages). In this condition, risk-averse decision making for managing these uncertainties are extremely complex, and the complexity is more amplified by the worst-case uncertainties. Accordingly, in this study a novel contingency-constrained information gap decision theory (CC-IGDT) approach has been proposed to cope with worst-case continuous and discrete uncertainties. Also, active distribution systems (ADSs) with distributed energy resources are important components in a TS, and can play an important role in addressing the issue of risk-averse management for TS operator. Therefore, in this study a coupled operation model for the TS & ADSs with the CC-IGDT approach has been proposed. But, solve proposed coupled operation model is problematic, thus, to solve this problem a new four-level hierarchical optimization technique has been proposed. Finally, the IEEE 30-bus transmission and IEEE 33-bus distribution systems have been analyzed to show the effectiveness of the proposed CC-IGDT approach and the co-operation of TS & ADSs.

RETRACTED: Robust-based risk assessment of buildings in resiliency and self-scheduling operations for coordinated vehicle-to-home and renewable resources

• Investigating the impact of resources in grid-connected and resiliency modes • Analysis of uncertainty conditions in grid-connected and resiliency modes • The effect of DR programming in grid-connected and resiliency modes • The impact of RO procedure in grid-connected and resiliency modes One of the crucial tools for organizing energy consumption in buildings is a home energy management system that can minimize the energy cost by using mathematical algorithms. In this paper two practical and theoretical works are carried out. In the practical contribution, the smart home operation is modeled in the grid-connected and resiliency modes, by considering the unpredictable conditions in the upstream grid. In the terms of theoretical works, a comprehensive model that includes demand response (DR) and robust optimization (RO) techniques are considered in the grid-connected and resiliency operation modes, accurately. Uncertainty of energy price and wind energy production are items modeled separately by using the RO technique in the grid-connected and resiliency modes, respectively. The uncertainty of wind energy and electricity price is considered separately for four seasons, and the simulation results show that the system's performance has reduced daily operation costs. In the grid-connected mode, the system's operation cost in without-DR mode is about 16% higher than in the with-DR mode, but in resilience mode, the value is about 3% that represents the effectiveness of proposed DR program in grid-connected operation mode.

Short-Term Forecasting of Wind Energy: A Comparison of Deep Learning Frameworks

Wind energy has been recognized as the most promising and economical renewable energy source, attracting increasing attention in recent years. However, considering the variability and uncertainty of wind energy, accurate forecasting is crucial to propel high levels of wind energy penetration within electricity markets. In this paper, a comparative framework is proposed where a suite of long short-term memory (LSTM) recurrent neural networks (RNN) models, inclusive of standard, bidirectional, stacked, convolutional, and autoencoder architectures, are implemented to address the existing gaps and limitations of reported wind power forecasting methodologies. These integrated networks are implemented through an iterative process of varying hyperparameters to better assess their effect, and the overall performance of each architecture, when tackling one-hour to three-hours ahead wind power forecasting. The corresponding validation is carried out through hourly wind power data from the Spanish electricity market, collected between 2014 and 2020. The proposed comparative error analysis shows that, overall, the models tend to showcase low error variability and better performance when the networks are able to learn in weekly sequences. The model with the best performance in forecasting one-hour ahead wind power is the stacked LSTM, implemented with weekly learning input sequences, with an average MAPE improvement of roughly 6, 7, and 49%, when compared to standard, bidirectional, and convolutional LSTM models, respectively. In the case of two to three-hours ahead forecasting, the model with the best overall performance is the bidirectional LSTM implemented with weekly learning input sequences, showcasing an average improved MAPE performance from 2 to 23% when compared to the other LSTM architectures implemented.

Data-driven stochastic optimization for power grids scheduling under high wind penetration

To address the environmental concern and improve the economic efficiency, the wind power is rapidly integrated into smart grids. However, the inherent uncertainty of wind energy raises operational challenges. To ensure the cost-efficient, reliable and robust operation, it is critically important to find the optimal decision that can correctly and rigorously hedge against all sources of uncertainty. In this paper, we propose data-driven stochastic unit commitment (SUC) to guide the power grids scheduling. Specifically, given the finite historical data, the posterior predictive distribution is developed to quantify the wind power prediction uncertainty accounting for both inherent stochastic uncertainty of wind power generation and input model estimation error. For complex power grid systems, a finite number of scenarios is used to estimate the expected cost in the planning horizon. To further control the impact of finite sampling error induced by using the sample average approximation (SAA), we propose a parallel computing based optimization solution methodology, which can quickly find the reliable optimal unit commitment decision hedging against various sources of uncertainty. The empirical study over six-bus and 118-bus systems demonstrates that our approach can provide more cost-efficient and robust performance than the existing deterministic and stochastic unit commitment approaches.