Aggregated Wind Power Research Articles

Enhancing Regional Wind Power Forecasting through Advanced Machine-Learning and Feature-Selection Techniques

In this study, an in-depth analysis is presented on forecasting aggregated wind power production at the regional level, using advanced Machine-Learning (ML) techniques and feature-selection methods. The main problem consists of selecting the wind speed measuring points within a large region, as the wind plant locations are assumed to be unknown. For this purpose, the main cities (province capitals) are considered as possible features and four feature-selection methods are explored: Pearson correlation, Spearman correlation, mutual information, and Chi-squared test with Fisher score. The results demonstrate that proper feature selection significantly improves prediction performance, particularly when dealing with high-dimensional data and regional forecasting challenges. Additionally, the performance of five prominent machine-learning models is analyzed: Long Short-Term Memory (LSTM) networks, Artificial Neural Networks (ANNs), Support Vector Machines (SVMs), Convolutional Neural Networks (CNNs), and Extreme-Learning Machines (ELMs). Through rigorous testing, LSTM is identified as the most effective model for the case study in northern Italy. This study offers valuable insights into optimizing wind power forecasting models and underscores the importance of feature selection in achieving reliable and accurate predictions.

Large scale integration of wind power – influence of geographical allocation

This paper investigates the influence of geographical allocation of wind power generation in Northern Europe, assuming large scale integration of wind power. The work applies a linear cost optimization model of the heat and power sector with a 1-hour time resolution. The model minimizes the sum of running costs to meet the heat and power demand and the wind power and transmission investment costs. Wind data are taken from modelled wind speed data from the Swedish Meteorological and Hydrological Institute. The Nordic countries and Germany were divided into regions and the 200 sites with the highest yearly output were chosen to represent the region. The model gives the most favourable distribution of wind power between the regions. In addition, the paper provides an assessment of the effect of geographical distribution of wind power with respect to influence on the aggregated wind power production (only considering the wind power generation itself). The modelling results show that the largest investments in wind power are made in the windy region of Southern Norway. However, depending on the cost of transmission allocating wind power near large load centers in Germany may also be favourable. As for the assessment of distribution of wind power, the wind data gives that if the 400 best sites in Europe were used, this would result in a capacity factor of 38.5% and a lowest output of 2.5 % of rated power (applying 2009 wind data).

Aggregated wind power characteristic curves and artificial intelligence for the regional wind power infeed estimation

The wind power generation is highly dependent on current weather conditions. In the course of the energy transition, the generation levels from volatile wind energy are constantly increasing. Accordingly, the prediction of regional wind power generation is a particularly important and challenging task due to the highly distributed installations. This paper presents a study on the role of regional wind power infeed estimation and proposes a multi-aggregated wind power characteristics model based on three scaled Gumbel distribution functions. Multi-levels of wind turbines and their allocation are investigated for the regional aggregated wind power. Relative peak power performance and full load hours are compared for the proposed model and the real measurement obtained from a local distribution system operator. Furthermore, artificial intelligence technologies using neural networks, such as Long Short-Term Memory (LSTM), stacked LSTM and CNN–LSTM, are investigated by using different historical measurement as input data. The results show that the suggested stacked LSTM performs stably and reliably in regional power prediction.

Chance‐constrained co‐expansion planning for power systems under decision‐dependent wind power uncertainty

AbstractVariability and uncertainty in wind resources pose significant challenges to the expansion planning of wind farms and associated flexible resources. In addition, the spatial smoothing effect, indicating the impact of wind farm scale on aggregated wind power prediction errors, further aggravates the challenge. This paper proposes a chance‐constrained co‐expansion planning method considering the spatial smoothing effect, where the expansion of wind farm capacity, batter energy storage capacity, and power transmission lines are co‐optimized. Specifically, a decision‐dependent uncertainty (DDU) model is established capturing the dependency of wind power uncertainties on wind farm expansion decisions under the spatial smoothing effect. Unlike traditional optimization diagram where decisions are made under only decision‐independent uncertainty (DIU) with fixe properties, properties of decision‐dependent uncertain parameters would be inversely altered by decisions. To effectively tackle the coupling relation between decisions and DDU, DDU‐based chance constraints are formulated in an analytical manner, where the decisions and decision‐dependent uncertain parameters are expressed in a closed form. Eventually, with piecewise linearization of the DDU model and the polynomial approximation of cumulative distribution function of uncertain parameters, the proposed chance‐constrained optimization model with DDU is converted into a mixed‐integer second‐order cone program (MISOCP). Case studies verify the effectiveness of the proposed method.

Dispatch Planning of a Wide-Area Wind Power-Energy Storage Scheme Based on Ensemble Empirical Mode Decomposition Technique

This article addresses the challenging task of developing a procedure for the day-ahead dispatch planning of wind power which emanates from a wide geographical area. Using the complete ensemble empirical mode decomposition technique, it is shown that the low-frequency components of the area aggregated wind power account for the largest proportions of the perturbing energy harnessed from the wind. By taking advantage of the slow-varying characteristics of the low-frequency components, accurate forecast of these components is readily obtained and incorporated into the developed dispatch planning procedure. The dispatchability of the wide-area wind generation is facilitated by the buffering actions offered by a centralized power dispatch energy storage system, operating under a proposed power flows control strategy. The efficacy of the developed procedure is illustrated using a pumped hydroelectric system as the dispatch energy storage medium.

Feature-Driven Improvement of Renewable Energy Forecasting and Trading

Inspired from recent insights into the common ground of machine learning, optimization and decision-making, this paper proposes an easy-to-implement, but effective procedure to enhance both the quality of renewable energy forecasts and the competitive edge of renewable energy producers in electricity markets with a dual-price settlement of imbalances. The quality and economic gains brought by the proposed procedure essentially stem from the utilization of valuable predictors (also known as features ) in a data-driven newsvendor model that renders a computationally inexpensive linear program. We illustrate the proposed procedure and numerically assess its benefits on a realistic case study that considers the aggregate wind power production in the Danish DK1 bidding zone as the variable to be predicted and traded. Within this context, our procedure leverages, among others, spatial information in the form of wind power forecasts issued by transmission system operators (TSO) in surrounding bidding zones and publicly available in online platforms. We show that our method is able to improve the quality of the wind power forecast issued by the Danish TSO by several percentage points (when measured in terms of the mean absolute or the root mean square error) and to significantly reduce the balancing costs incurred by the wind power producer.

High Frequency Evidence on the Effects of Wind Power Generation on the Electricity Price

Exploiting the hourly data from the New York Independent System Operator (NYISO) database, we model the impact of wind generation on electricity spot prices and on their dynamic volatility behavior while taking into account the variations in zonal wind output. We find that the growing share of wind generation not only leads to a decline in electricity prices, but also results in an increased price volatility. The results indicate that a 1% rise in the aggregate wind power generation leads to a 0.012% decline in the average NYISO prices while the mentioned price effect in the wind producing zones is estimated to be a 0.055% to 0.085% downward adjustment of the zonal prices for a 1% growth in wind generation within the zone. Our findings have two implications: First, the increasing downward pressure on electricity prices due to the growing share of wind generation can challenge the financial sustainability of the electricity markets in the long-run, which highlights the need to proceed towards an effective market structure to ensure the recovery of fixed generation costs. Second, the induced price volatility implies that effective use of price risk hedging strategies should be considered by all actors in the power markets.

Reduction of aggregate wind power variability using Empirical Orthogonal Teleconnections: An application in the Iberian Peninsula

Abstract In this work, the ability of the Empirical Orthogonal Teleconnections technique to provide useful information for decisions on wind capacity allocation aimed at enhanced output stability is explored. Using data from a high-resolution simulation with the regional climate model REMO over the Iberian Peninsula, the performance of Empirical Orthogonal Teleconnections is assessed against the outcome of random wind farm siting, and also against an alternative procedure employing Principal Component Analysis. Results show that siting informed by the Empirical Orthogonal Teleconnections methodology leads to increased probabilities of achieving higher firm capacities and lower output variance, improving not only the random allocation, but that given by the Principal Component Analysis procedure as well. The benefits stem from the smoothing effect of aggregating the output from wind farms located within specific areas identified by our technique. Our appraisal also considers the spatial extent of the areas made available to the siting choice, as well as the effects on the capacity factor. Being a comparatively inexpensive technique, Empirical Orthogonal Teleconnections offers good prospects for further diagnosis of the wind intermittency problem and its mitigation through spatial aggregation, and also as a preliminary, dimension reductive assessment for the application of more computationally demanding methods.

Less Information, Similar Performance: Comparing Machine Learning-Based Time Series of Wind Power Generation to Renewables.ninja

Driven by climatic processes, wind power generation is inherently variable. Long-term simulated wind power time series are therefore an essential component for understanding the temporal availability of wind power and its integration into future renewable energy systems. In the recent past, mainly power curve-based models such as Renewables.ninja (RN) have been used for deriving synthetic time series for wind power generation, despite their need for accurate location information and bias correction, as well as their insufficient replication of extreme events and short-term power ramps. In this paper, we assessed how time series generated by machine learning models (MLMs) compare to RN in terms of their ability to replicate the characteristics of observed nationally aggregated wind power generation for Germany. Hence, we applied neural networks to one wind speed input dataset derived from MERRA2 reanalysis with no location information and two with additional location information. The resulting time series and RN time series were compared with actual generation. All MLM time series feature an equal or even better time series quality than RN, depending on the characteristics considered. We conclude that MLM models show a similar performance to RN, even when information on turbine locations and turbine types is unavailable.

Conditional aggregated probabilistic wind power forecasting based on spatio-temporal correlation

Aggregated probabilistic wind power forecasting is important for power system operations. In this paper, an improved aggregated probabilistic wind power forecasting framework based on spatio-temporal correlation is developed. A Q-learning enhanced deterministic wind power forecasting method is used to generate deterministic wind power forecasts for individual wind farms. The spatio-temporal correlation between the member wind farms and the aggregated wind power is modeled by a joint distribution model based on the copula theory. The marginal distributions of actual aggregated wind power and forecasted power of member wind farms are built with Gaussian mixture models. Then, a conditional distribution of the aggregated wind power is deduced through the Bayesian theory, which is used for aggregated probabilistic forecasts. The effectiveness of the proposed aggregated probabilistic wind power forecasting framework is validated by using the Wind Integration National Dataset Toolkit. Numerical results of case studies at nine locations show that the developed aggregated probabilistic forecasting methodology has improved the pinball loss metric score by up to 54% compared to three benchmark models.

Simulating wind power forecast error distributions for spatially aggregated wind power plants

AbstractDispersion and aggregation of wind power plants lower the uncertainty of wind power by reducing wind power forecasting errors. Using quantitative methods, this paper studies the dispersion's impact on the uncertainty of the aggregated wind power production. A method to simulate day‐ahead forecast error distributions at different dispersion and forecasting skill scenarios is presented. The proposed method models the uncertainty of wind power forecasting on an annual basis and at different levels of production. As a result, the uncertainty in the forecasting of spatially dispersed wind power plants is modelled using two continuous distributions: Laplace and beta distribution. The analysis shows that even the production level uncertainty can be modelled in various dispersion and forecasting skill scenarios. The model for aggregated forecast error distributions requires only four variables: capacity‐weighted distance of the total wind power plant fleet, mean of the site‐specific mean absolute errors (MAEs), number of aggregated wind power plants, and the mean variability of the elevations from the proximities of the aggregated wind power plants. The results are especially promising when the number of aggregated wind power plants exceeds five, the terrain complexity is low or moderate, and the aggregation region is large. This is demonstrated through a case study for Texas, United States.

A distributionally robust bidding strategy for a wind power plant

This paper presents a two-stage distributionally robust model to derive optimal bidding strategies for an aggregated wind power plant (WPP), that participates as a price-maker in the day-ahead market, and a deviator in the balancing market. The market power is realized by using a bi-level model, which is then transformed into a mixed-integer linear programming model using the Karush-Kuhn-Tucker (KKT) optimality conditions and strong duality theory. The uncertainty in wind generation output is characterized by an ambiguity set that defines a family of distributions. The optimal decision is robust to the expectation over the worst-case distribution. With a case study based on a modified Swiss system, we verify the effectiveness of the proposed distributionally robust optimization model and compare its performance to that of robust and stochastic optimization models.

Co-simulation of a Doubly Fed Induction Generator Connected to a Power Network: The use of DSOGI for Phasor Extraction

This work presents a hybrid simulation tool that combines fast analysis of transient stability programs with accurate results of electromagnetic transients ones to evaluate the dynamic behavior of transmission power systems with high penetration of power electronic-based devices. The interaction between the two programs is performed by controllable current and voltage sources, which are used to interface external and detailed systems. The double second order generalized integrator is used to extract the positive-sequence phasor to transfer information between the environments. A local network server controls the data communication between the two simulation environments by means of the TCP/IP protocol. In the present paper, the proposed tool is used to simulate the integration of a aggregate wind power plant, based on a doubly-fed induction generator, into a 29-bus electrical network. Results and computational timings are then compared with the ones obtained with an electromagnetic transients program, demonstrating the accuracy and speed of the proposed strategy.

Simulating subhourly variability of wind power output

AbstractAs installed wind power capacity grows, subhourly variability in wind power output becomes increasingly important for determining the system flexibility needs, operating reserve requirements, and cost associated with wind integration. This paper presents a new methodology for simulating subhourly wind power output based on hourly average time series, which are often produced for system planning analyses, for both existing wind plants and expanded, hypothetical portfolios of wind plants. The subhourly model has an AR(p)‐ARCH(q) structure with exogenous input in the heteroskedasticity term. Model coefficients may be fit directly to high‐pass filtered historical data if it exists; for sets of wind plants containing hypothetical plants for which there are no historical data, this paper presents a method to determine model coefficients based on wind plant capacities, capacity factors, and pairwise distances. Unlike predecessors, the model presented in this paper is independent of wind speed data, captures explicitly the high variability associated with intermediate levels of power output, and captures distance‐dependent correlation between the power output of wind plants across subhourly frequencies. The model is parameterized with 1‐minute 2014 plant‐level wind power data from Electric Reliability Council of Texas (ERCOT) and validated out‐of‐sample against analogous 2015 data. The expanded‐capacity model, fit to 2014 data, produces accurate subhourly time series for the 2015 wind fleet (a 49% capacity expansion) based only on the 2015 system's wind plant capacities, capacity factors, and pairwise distances. This supports its use in simulating subhourly fleet aggregate wind power variability for future high‐wind scenarios.

Estimation of the Daily Variability of Aggregate Wind Power Generation in Alberta, Canada

This paper describes a hierarchy of increasingly complex statistical models for wind power generation in Alberta applied to wind power production data that are publicly available. The models are based on combining spatial and temporal correlations. We apply the method of Gaussian random fields to analyze the wind power time series of the 19 existing wind farms in Alberta. Following the work of Gneiting et al., three space-time models are used: Stationary, Separability, and Full Symmetry. We build several spatio-temporal covariance function estimates with increasing complexity: separable, non-separable and symmetric, and non-separable and non-symmetric. We compare the performance of the models using kriging predictions and prediction intervals for both the existing wind farms and a new farm in Alberta. It is shown that the spatial correlation in the models captures the predominantly westerly prevailing wind direction. We use the selected model to forecast the mean and the standard deviation of the future aggregate wind power generation of Alberta and investigate new wind farm siting on the basis of reducing aggregate variability.

Reactive Power Capability Model of Wind Power Plant Using Aggregated Wind Power Collection System

This article presents the development of a reactive power capability model for a wind power plant (WPP) based on an aggregated wind power collection system. The voltage and active power dependent reactive power capability are thus calculated by using aggregated WPP collection system parameters and considering losses in the WPP collection system. The strength of this proposed reactive power capability model is that it not only requires less parameters and substantially less computational time compared to typical detailed models of WPPs, but it also provides an accurate estimation of the available reactive power. The proposed model is based on a set of analytical equations which represent converter voltage and current limitations. Aggregated impedance and susceptance of the WPP collection system are also included in the analytical equations, thereby incorporating losses in the collection system in the WPP reactive power capability calculation. The proposed WPP reactive power capability model is compared to available methodologies from literature and for different WPP topologies, namely, Horns Rev 2 WPP and Burbo Bank WPP. Performance of the proposed model is assessed and discussed by means of simulations of various case studies demonstrating that the error between the calculated reactive power using the proposed model and the detailed model is below 4% as compared to an 11% error in the available method from literature. The efficacy of the proposed method is further exemplified through an application of the proposed method in power system integration studies. The article provides new insights and better understanding of the WPPs’ limits to deliver reactive power support that can be used for power system stability assessment, particularly long-term voltage stability.

A Distributed Probabilistic Modeling Algorithm for the Aggregated Power Forecast Error of Multiple Newly Built Wind Farms

The extensive penetration of wind farms (WFs) presents challenges to the operation of distribution networks (DNs). Building a probability distribution of the aggregated wind power forecast error is of great value for decision making. However, as a result of recent government incentives, many WFs are being newly built with little historical data for training distribution models. Moreover, WFs with different stakeholders may refuse to submit the raw data to a data center for model training. To address these problems, a Gaussian mixture model (GMM) is applied to build the distribution of the aggregated wind power forecast error; then, the maximum a posteriori (MAP) estimation method is adopted to overcome the limited training data problem in GMM parameter estimation. Next, a distributed MAP estimation method is developed based on the average consensus filter algorithm to address the data privacy issue. The distribution control center is introduced into the distributed estimation process to acquire more precise estimation results and better adapt to the DN control architecture. The effectiveness of the proposed algorithm is empirically verified using historical data.

A CVaR-Robust Risk Aversion Scheduling Model for Virtual Power Plants Connected with Wind-Photovoltaic-Hydropower-Energy Storage Systems, Conventional Gas Turbines and Incentive-Based Demand Responses

To make full use of distributed energy resources to meet load demand, this study aggregated wind power plants (WPPs), photovoltaic power generation (PV), small hydropower stations (SHSs), energy storage systems (ESSs), conventional gas turbines (CGTs) and incentive-based demand responses (IBDRs) into a virtual power plant (VPP) with price-based demand response (PBDR). Firstly, a basic scheduling model for the VPP was proposed in this study with the objective of the maximum operation revenue. Secondly, a risk aversion model for the VPP was constructed based on the conditional value at risk (CVaR) method and robust optimization theory considering the operating risk from WPP and PV. Thirdly, a solution methodology was constructed and three cases were considered for comparative analyses. Finally, an independent micro-grid on an industrial park in East China was utilized for an example analysis. The results show the following: (1) the proposed risk aversion scheduling model could cope with the uncertainty risk via a reasonable confidence degree β and robust coefficient Γ. When Γ ≤ 0.85 or Γ ≥ 0.95, a small uncertainty brought great risk, indicating that the risk attitude of the decision maker will affect the scheduling scheme of the VPP, and the decision maker belongs to the risk extreme aversion type. When Γ ∈ (0.85, 0.95), the decision-making scheme was in a stable state, the growth of β lead to the increase of CVaR, but the magnitude was not large. When the prediction error e was higher, the value of CVaR increased more when Γ increased by the same magnitude, which indicates that a lower prediction accuracy will amplify the uncertainty risk. (2) when the capacity ratio of (WPP, PV): ESS was higher than 1.5:1 and the peak-to-valley price gap was higher than 3:1, the values of revenue, VaR, and CVaR changed slower, indicating that both ESS and PBDR can improve the operating revenue, but the capacity scale of ESS and the peak-valley price gap need to be set properly, considering both economic benefits and operating risks. Therefore, the proposed risk aversion model could maximize the utilization of clean energy to obtain higher economic benefits while rationally controlling risks and provide reliable decision support for developing optimal operation plans for the VPP.

Assessing Smoothing Effects of Wind Power around Trondheim via Koopman Mode Decomposition

To cope with a large-scale introduction of renewables into power systems, it is important to understand the reduction of variability in the aggregated generation or the so-called smoothing effect. Knowledge of the degree of smoothing is used for assessing the potential impact of intermittent generation on the power system operation. Here, smoothing effects of aggregated wind power are assessed around Trondheim, Norway, by applying a recently proposed smoothing index based on the so-called Koopman Mode Decomposition (KMD). The method is shown to effectively decompose complex time-series of wind power outputs into a finite number of modes, each of which oscillates with a single frequency for all locations (or hypothetical wind farms). It is shown that the method is able to reconstruct the original power outputs well by only a small number of modes that retain the variability of the original time-series, and is able to provide a relevant quantification of the smoothing effects for each individual frequency (or mode).

A Statistical Modeling Methodology for Long-Term Wind Generation and Power Ramp Simulations in New Generation Locations

In future power systems, a large share of the energy will be generated with wind power plants (WPPs) and other renewable energy sources. With the increasing wind power penetration, the variability of the net generation in the system increases. Consequently, it is imperative to be able to assess and model the behavior of the WPP generation in detail. This paper presents an improved methodology for the detailed statistical modeling of wind power generation from multiple new WPPs without measurement data. A vector autoregressive based methodology, which can be applied to long-term Monte Carlo simulations of existing and new WPPs, is proposed. The proposed model improves the performance of the existing methodology and can more accurately analyze the temporal correlation structure of aggregated wind generation at the system level. This enables the model to assess the impact of new WPPs on the wind power ramp rates in a power system. To evaluate the performance of the proposed methodology, it is verified against hourly wind speed measurements from six locations in Finland and the aggregated wind power generation from Finland in 2015. Furthermore, a case study analyzing the impact of the geographical distribution of WPPs on wind power ramps is included.