Accurate Wind Forecasting Research Articles

A novel dynamic ensemble of numerical weather prediction for multi-step wind speed forecasting with deep reinforcement learning and error sequence modeling

Accurate wind forecasts for one day ahead or longer periods have significant impacts on the safe and efficient dispatch of power grids, where Numerical Weather Prediction (NWP) serves as the essential tool, such as ensemble NWP integrating multiple single simulations. Typically, ensembles include all single members with fixed weights; however, the relative accuracy of each member may change over time. This study introduces an attractive idea: improving ensemble performance by dynamically recognizing and avoiding low-performing members. It proposes a dynamic ensemble strategy based on NWP, reinforcement learning and error sequence correction. The process begins with Weather Research and Forecasting ensemble simulations. A dynamic framework is then constructed by mapping the multi-step ensemble problem into a Markov decision process, which is further solved using deep deterministic policy gradient. Subsequently, a hybrid deep learning model, comprising temporal convolutional network and bidirectional long short-term memory, is constructed for error sequence estimation of dynamic ensemble, using the high-frequency information of NWP as input. Conducting experiments at two wind farms, and focusing on the 24-h wind speed forecast with a 15-min time resolution, the proposed system demonstrates a reliable and stable ensemble throughout the entire forecasting horizon, significantly reducing the probability of large forecasting errors.

Artificial Intelligence in Wind Speed Forecasting: A Review

Wind energy production has had accelerated growth in recent years, reaching an annual increase of 17% in 2021. Wind speed plays a crucial role in the stability required for power grid operation. However, wind intermittency makes accurate forecasting a complicated process. Implementing new technologies has allowed the development of hybrid models and techniques, improving wind speed forecasting accuracy. Additionally, statistical and artificial intelligence methods, especially artificial neural networks, have been applied to enhance the results. However, there is a concern about identifying the main factors influencing the forecasting process and providing a basis for estimation with artificial neural network models. This paper reviews and classifies the forecasting models used in recent years according to the input model type, the pre-processing and post-processing technique, the artificial neural network model, the prediction horizon, the steps ahead number, and the evaluation metric. The research results indicate that artificial neural network (ANN)-based models can provide accurate wind forecasting and essential information about the specific location of potential wind use for a power plant by understanding the future wind speed values.

SHORT-TERM WIND SPEED FORECASTING USING DEEP VARIATIONAL LSTM

Wind speed and power at wind power stations affect the efficiency of a wind farm, so accurate wind forecasting, a nonlinear signal with high fluctuations, increases security and better efficiency than wind power. We are looking for wind speed for a wind farm in Iran. In this research, a combined neural network created from variational autoencoder (VAE), long-term, short-term memory (LSTM), and multilayer perceptron (MLP) for dimension Reduction and encoding is proposed for predicting short-term wind speeds. The data used in this research is related to the statistics of 10 minutes of wind speed in 10- meter, 30-meter, and 40-meter wind turbines, the standard deviation of wind speed, air temperature, and humidity. To compare the proposed model (V- LSTM-MLP), we implemented three deep neural network models, including Stacked Auto-Encoder (SAE), recurrent neural networks (Regular LSTM), and hybrid model Encoder-Decoder recurrent network (LSTM-Encoder-MLP) presented on this dataset. According to the RMSE statistical index, the proposed model is worth 0.1127 for a short time and performs better than other types on this dataset.

Scientific challenges to characterizing the wind resource in the marine atmospheric boundary layer

Abstract. With the increasing level of offshore wind energy investment, it is correspondingly important to be able to accurately characterize the wind resource in terms of energy potential as well as operating conditions affecting wind plant performance, maintenance, and lifespan. Accurate resource assessment at a particular site supports investment decisions. Following construction, accurate wind forecasts are needed to support efficient power markets and integration of wind power with the electrical grid. To optimize the design of wind turbines, it is necessary to accurately describe the environmental characteristics, such as precipitation and waves, that erode turbine surfaces and generate structural loads as a complicated response to the combined impact of shear, atmospheric turbulence, and wave stresses. Despite recent considerable progress both in improvements to numerical weather prediction models and in coupling these models to turbulent flows within wind plants, major challenges remain, especially in the offshore environment. Accurately simulating the interactions among winds, waves, wakes, and their structural interactions with offshore wind turbines requires accounting for spatial (and associated temporal) scales from O(1 m) to O(100 km). Computing capabilities for the foreseeable future will not be able to resolve all of these scales simultaneously, necessitating continuing improvement in subgrid-scale parameterizations within highly nonlinear models. In addition, observations to constrain and validate these models, especially in the rotor-swept area of turbines over the ocean, remains largely absent. Thus, gaining sufficient understanding of the physics of atmospheric flow within and around wind plants remains one of the grand challenges of wind energy, particularly in the offshore environment. This paper provides a review of prominent scientific challenges to characterizing the offshore wind resource using as examples phenomena that occur in the rapidly developing wind energy areas off the United States. Such phenomena include horizontal temperature gradients that lead to strong vertical stratification; consequent features such as low-level jets and internal boundary layers; highly nonstationary conditions, which occur with both extratropical storms (e.g., nor'easters) and tropical storms; air–sea interaction, including deformation of conventional wind profiles by the wave boundary layer; and precipitation with its contributions to leading-edge erosion of wind turbine blades. The paper also describes the current state of modeling and observations in the marine atmospheric boundary layer and provides specific recommendations for filling key current knowledge gaps.

Validation of the CFOSAT Scatterometer Data With Buoy Observations and Tests of Operational Application to Extreme Weather Forecasts in Taiwan Strait

AbstractThe China‐France Oceanography Satellite (CFOSAT) launched in 2018 is equipped with the Chinese Scatterometer (CSCAT), which is designed to measure high‐precision sea surface wind fields in the global ocean. This study aims to validate the CFOSAT wind fields with moored buoy‐measured ground truth data from August 2019 to July 2021. The test area was chosen as the Taiwan Strait, which is an important navigation channel and rich fishery grounds in the East Asia. It is also a high wind speed area particularly during typhoons and cold air passages. Thus, the accurate wind forecasts are crucially needed to avoid capsizing and casualties. The validation results give the correlation coefficient (R2) of 0.92, the root mean square error (RMSE) of 1.75 m s−1, and the absolute deviation (AD) of 1.10 m s−1 for the wind speeds, and R2 of 0.96, the RMSE of 21.87°, and the AD of 1.17° for the wind directions. During cold air passages, the AD of wind speeds decreases to 1.03 m s−1, which is smaller than 1.17 m s−1 during typhoons. Using the CFOSAT wind as historical observation data for refining of forecast work was tested by model output statistics revised method. The test results show that the refined forecast results of wind fields are improved up to 4%–17%. Thus, the validation and test results obtained in this study indicate that updating satellite winds would have great contributions to refined forecasting of local sea surface winds.

Evaluating modelled winds over an urban area using ground‐based Doppler lidar observations

AbstractWind information in urban areas is essential for many applications related to air pollution, urban climate and planning, safety of drone‐related operations, and assessment of urban wind energy potential. These applications require accurate wind forecasts, and obtaining this information in an urban environment is challenging as the morphology of a city varies from street to street, altering the wind flow. Remote sensing techniques such as Doppler lidars (light detection and ranging) provide a unique opportunity for wind forecast verification as they can provide both the vertical profile of the horizontal wind and the spatial variation in the horizontal domain at high resolution. In this study, the performance of numerical weather prediction (NWP) models, analysis systems, and large‐eddy simulation (LES) models have been analysed by comparing the modelled winds against Doppler lidar observations under various atmospheric conditions and from season to season, in the coastal environment of Helsinki, Finland. The long‐term mean vertical profile of the modelled horizontal wind shows good agreement with observations; the NWP model and the analysis systems selected here exhibit different strengths and weaknesses depending on the atmospheric conditions but no significant diurnal variation in performance. However, both the model and analysis systems show differences in their spatially‐averaged bias when investigating different wind directions. LES verification shows that these models can potentially provide winds down to street level, given pre‐computed scenarios of atmospheric conditions. For Helsinki, the observed winds are stronger during winter than summer, and, on average, higher wind speeds were observed at the urban site than the sub‐urban site.

Sensitivity analysis of the WRF model: Assessment of performance in high resolution simulations in complex terrain in the Canary Islands

Canary Islands and other regions have been greatly damaged by weather events during the last decades. For this reason, the main duty of National Meteorological Services is to minimize socio-economic losses by forecasting adverse weather episodes with enough time in advance. To achieve this goal, the use of numerical weather prediction models is highly relevant. And, even more crucial, is to comprehend the model accuracy.In this paper, an exhaustive sensitivity analysis of the Weather Research and Forecasting (WRF) model over the Canary Islands has been carried out. The complex terrain of the archipelago makes the islands a test bench of high interest. Four scores were used to assess the accuracy of the model configurations: Bias, mean absolute error (MAE), root of the Mean Squared Error (RMSE), and the correlation coefficient (r). Initially, twenty-five WRF model configurations were considered. However, a preliminary test discarded inadequate configurations, and reduced the number to six. The variables of interest were air temperature at 2 m (T2m), maximum 1-h wind gust at 10 m and 3-h rainfall accumulation. The results indicated a systematic wind speed underestimation. This underestimation is related to the influence of the location and the complex orography. The most accurate wind forecasts were obtained using the Mellor-Yamada-Janjic Planetary Boudary Layer (PBL) scheme with the WSM6 microphysics (MP) scheme. Another major conclusion is that, for precipitation, the PBL scheme has a greater impact than the MP scheme. Finally, the results show that the Boulac – Thompson combination is the most accurate regarding T2m forecast.

Timescale classification in wind forecasting: A review of the state‐of‐the‐art

AbstractThe intermittency of the wind has been reported to present significant challenges to power and grid systems, which intensifies with increasing penetration levels. Accurate wind forecasting can mitigate these challenges and help in integrating more wind power into the grid. A range of studies have presented algorithms to forecast the wind in terms of wind speeds and wind power generation across different timescales. However, the classification of timescales varies significantly across the different studies (2010–2014). The timescale is important in specifying which methodology to use when, as well in uniting future research, data requirements, etc. This study proposes a generic statement on how to classify the timescales, and further presents different applications of these forecasts across the entire wind power value chain.

Ensemble Machine Learning-Based Wind Forecasting to Combine NWP Output With Data From Weather Station

Wind generation resources are the fastest growing energy resources throughout the world. An accurate wind forecasting is critical to the integration of a large amount of wind generation units into the grid operations. This paper presents an ensemble machine learning-based method to forecast wind power production, which uses both the wind generation forecasted by a numerical weather prediction (NWP) model and the meteorological observation data from weather stations. In this way, it takes advantage of the atmosphere models while capturing spatial and temporal correlation between meteorological observations at different locations. Three machine learning algorithms (artificial neural network, support vector regression, Gaussian process) are proposed to synthesize the meteorological data and the prediction from NWP. An ensemble forecast is then created by blending the results derived from three algorithms through a Bayesian model average. The performance of this ensemble forecast has been validated by the 2-year operational data collected at Electricity Reliability Council of Texas.

Intra-Interval Security Based Dispatch for Power Systems With High Wind Penetration

Security-constrained economic dispatch is executed based on consecutive dispatch intervals in real time. The short-term variations of wind power can cause the system to enter an unsecured operating region inside a dispatch interval, which is referred to as the intra-interval security (IIS) problem in this paper. IIS may not be addressed by existing dispatch methods as the traditional dispatch is to ensure the security at the interval end. This can lead to IIS risks in an interval when a high-wind-penetrated system encounters large intra-interval variations (IIVs) of wind power. To address this issue, an IIS preventive dispatch approach is proposed in this study. This approach models the wind power IIVs and develops a preventive dispatch to ensure the IIS over the interval in terms of line overloads while taking into account the tradeoff between security and economics. The proposed approach is implemented in a linear-programming model whose global optimum can be efficiently obtained by adding Benders-like cuts. Case study on the IEEE 118-bus system proves that a high level of nonlinear IIVs can cause the IIS issue (inside of an interval) even with accurate wind forecasts at both interval ends. The simulations show that some lines can have severe overloads inside the dispatch interval in spite of having flows below their capacity limits at both interval ends, and demonstrate that the proposed approach can mitigate such intra-interval overloads even under the worst case IIV scenario.

The challenge of integrating offshore wind power in the U.S. electric grid. Part II: Simulation of electricity market operations

The purpose of this two-part study is to analyze large penetrations of offshore wind power into a large electric grid, using the case of the grid operated by PJM Interconnection in the northeastern U.S. Part I of the study introduces the wind forecast error model and Part II, this paper, describes Smart-ISO, a simulator of PJM's planning process for generator scheduling, including day-ahead and intermediate-term commitments to energy generators and real-time economic dispatch. Results show that, except in summer, an unconstrained transmission grid can meet the load at five build-out levels spanning 7–70 GW of capacity, with the addition of at most 1–8 GW of reserves.In the summer, the combination of high load and variable winds is challenging. The simulated grid can handle up through build-out level 3 (36 GW of offshore wind capacity), with 8 GW of reserves and without any generation shortage. For comparison, when Smart-ISO is run with perfect forecasts, all five build-out levels, up to 70 GW of wind, can be integrated in all seasons with at most 3 GW of reserves. This reinforces the importance of accurate wind forecasts. At build-out level 3, energy from wind would satisfy between 11 and 20% of the demand for electricity and settlement prices could be reduced by up to 24%, though in the summer peak they could actually increase by up to 6%. CO2 emissions are reduced by 19–40%, SO2 emissions by 21–43%, and NOx emissions by 13–37%.This study finds that integrating up to 36 GW of offshore wind is feasible in the PJM grid with today's generation fleet and planning policies, with the addition of 8 GW of reserves. Above that, PJM would require additional investments in fast-ramping gas turbines, storage for smoothing fast-ramping events, and/or other strategies such as demand response.

Exploiting sparsity of interconnections in spatio-temporal wind speed forecasting using Wavelet Transform

Integration of renewable energy resources into the power grid is essential in achieving the envisioned sustainable energy future. Stochasticity and intermittency characteristics of renewable energies, however, present challenges for integrating these resources into the existing grid in a large scale. Reliable renewable energy integration is facilitated by accurate wind forecasts. In this paper, we propose a novel wind speed forecasting method which first utilizes Wavelet Transform (WT) for decomposition of the wind speed data into more stationary components and then uses a spatio-temporal model on each sub-series for incorporating both temporal and spatial information. The proposed spatio-temporal forecasting approach on each sub-series is based on the assumption that there usually exists an intrinsic low-dimensional structure between time series data in a collection of meteorological stations. Our approach is inspired by Compressive Sensing (CS) and structured-sparse recovery algorithms. Based on detailed case studies, we show that the proposed approach based on exploiting the sparsity of correlations between a large set of meteorological stations and decomposing time series for higher-accuracy forecasts considerably improve the short-term forecasts compared to the temporal and spatio-temporal benchmark methods.

Atmospheric Dispersion Study of TRS Compounds Emitted from a Pulp Mill Plant in Coastal Regions of the Uruguay River, South America

ABSTRACTThe atmospheric dispersion of total reduced sulfur (TRS) emissions from the pulp mill plant of Fray Bentos, Uruguay is simulated. The local authorities of the Environmental Monitoring Program (EMP) of Gualeguaychu, Argentina, received social complaints of malodor presence in different places of the region. An atmospheric dispersion model coupled to a boundary layer forecast model is used to simulate 11 events in which the EMP officials attended the scene in order to verify the situation. The validation of modeled winds with the observations from a meteorological tower indicates reasonably accurate wind forecasts. The spatial layout of the modeled TRS plumes is compared with the geographical distribution of points in the area where the social complaints were recorded. Nine of the 11 studied events are successful modeling cases since a positive (negative) in situ verification matches with a plume position over (far from) the site. In one of the two unsuccessful modeling cases, although the plume is marginally distant from the site, the average wind direction error is the largest one of all the events. In the other case the modeled plume is in fact over the site, but the situation was negatively verified. The reason for the disagreement could be the wind direction changes during the event. This was the longest modeled case that lasted for 7 hours and the plume was meandering during that time; first from SSW to the S, then back the SSW, and finally to the S and SSE. The conclusion of the study is that, despite the inherent uncertainty of numerical simulations, the implemented modeling system shows versatility and proves to be a useful tool not only for diagnostic studies but also for preventing conflictive situations since it can produce reasonably accurate forecast of plume position and its potential impact.

Analytic Assessments of the Economics, Reliability, and Operating Impacts of Wind Power Plants

This paper summarizes a number of projects undertaken at the National Renewable Energy Laboratory (NREL) that focus on the analysis and modeling of large-scale wind generation. The findings indicate that wind power plants can provide significant economic benefits to the generating system. Potential problems caused by the intermittency of a wind plant can be significantly reduced by accurate wind forecasting, geographic dispersion of turbines within a plant, as well as disperse wind-generating facilities. Wind plants do not require significant backup facilities because risk-analysis techniques assess the risk of system failure on a system-wide basis.

Modelling utility‐scale wind power plants. Part 2: Capacity credit

AbstractAs the worldwide use of wind turbine generators in utility‐scale applications continues to increase, it will become increasingly important to assess the economic and reliability impact of these intermittent resources. Although the utility industry appears to be moving towards a restructured environment, basic economic and reliability issues will continue to be relevant to companies involved with electricity generation. This article is the second in a two‐part series that addresses modelling approaches and results that were obtained in several case studies and research projects at the National Renewable Energy Laboratory (NREL). This second article focuses on wind plant capacity credit as measured with power system reliability indices. Reliability‐based methods of measuring capacity credit are compared with wind plant capacity factor. The relationship between capacity credit and accurate wind forecasting is also explored. Published in 2000 by John Wiley & Sons, Ltd.