Offshore Wind Speed Research Articles

Future projections of offshore wind energy resources in China using CMIP6 simulations and a deep learning-based downscaling method

Good knowledge of future wind energy resources is crucial for sitting and planning studies of wind farms. The simulation results from the Coupled Model Intercomparison Project Phase 6 (CMIP6) and a proposed new downscaling method based on the bidirectional gated recurrent unit (BiGRU) are both used in this paper to study future offshore wind energy resources in China. The proposed new downscaling method is validated and compared to two traditional methods. It is found that the spatial patterns of downscaled wind speed are highly consistent with the reference data, and biases are significantly reduced by the new method, especially in coastal and shallow water areas. Using the new method, we downscale the CMIP6 future projected simulation results and generate a new dataset of offshore wind speeds in China for the period of 2021–2100 with a resolution of 0.25°. Multi-model ensemble (MME) results project small decreases in the offshore wind speed and wind power density over the East China Sea and increases in those parameters over the South China Sea, for the middle and end of the 21st Century (2041–2060 and 2081–2100) under two representative scenarios (SSP2-4.5 and SSP5-8.5).

Quantitative assessment of offshore wind speed variability using fractal analysis

Proper understanding of offshore wind speed variability is of essential importance in practice, which provides useful information to a wide range of coastal and marine activities. In this paper, long-term wind speed data recorded at various offshore stations are analyzed in the framework of fractal dimension analysis. Fractal analysis is a well-established data analysis tool, which is particularly suitable to determine the complexity in time series from a quantitative point of view. The fractal dimension is estimated using the conventional box-counting method. The results suggest that the wind speed data are generally fractals, which are likely to exhibit a persistent nature. The mean fractal dimension varies from 1.31 at an offshore weather station to 1.43 at an urban station, which is mainly associated with surface roughness condition. Monthly variability of fractal dimension at offshore stations is well-defined, which often possess larger values during hotter months and lower values during winter. This is partly attributed to the effect of thermal instability. In addition, with an increase in measurement interval, the mean and minimum fractal dimension decrease, whereas the maximum and coefficient of variation increase in parallel.

Investigation of Marine Wind Veer Characteristics Using Wind Lidar Measurements

A proper understanding of marine wind characteristics is of essential importance across a wide range of engineering applications. While the offshore wind speed and turbulence characteristics have been examined extensively, the knowledge of wind veer (i.e., turning of wind with height) is much less understood and discussed. This paper presents an investigation of marine wind field with particular emphasis on wind veer characteristics. Extensive observations from a light detection and ranging (Lidar) system at an offshore platform in Hong Kong were examined to characterize the wind veer profiles up to a height of 180 m. The results underscored the occurrence of marine wind veer, with a well-defined two-fold vertical structure. The observed maximum wind veer angle exhibits a reverse correlation with mean wind speed, which decreases from 2.47° to 0.59° for open-sea terrain, and from 7.45° to 1.92° for hilly terrain. In addition, seasonal variability of wind veer is apparent, which is most pronounced during spring and winter due to the frequent occurrence of the low-level jet. The dependence of wind veer on atmospheric stability is evident, particularly during winter and spring. In general, neutral stratification reveals larger values of wind veer angle as compared to those in stable and unstable stratification conditions.

Predictive Analysis of Machine Learning Schemes in Forecasting Offshore Wind Speed

High variability of wind in the farm areas causes a drastic instability in the energy markets. Therefore, precise forecast of wind speed plays a key role in the optimal prediction of offshore wind power. In this study, we apply two deep learning models, i.e. Long Short-Term Memory (LSTM) and Nonlinear Autoregressive EXogenous input (NARX), for predicting wind speed over long-range of dependencies. We use a four-month-long wind speed/direction, air temperature, and atmospheric pressure time series (all recorded at 10 m height) from a meteorological mast (Vigra station) in the close vicinity of the Havsul-I offshore area near Ålesund, Norway. While both predictive methods could efficiently predict the wind speed, the LSTM with update generally outperforms the NARX. The NARX suffers from vanishing gradient issue and its performance declines by abrupt variability inherited in the input data during training phase. It is observed that this sensitivity will significantly decrease by integrating, for example, the wind direction at low frequencies in the learning process. Generally, the results showed that the predictive models are robust and accurate in short-term and somewhat long-term forecasting of wind.

Quantitative assessment of offshore wind speed variability using fractal analysis

Proper understanding of offshore wind speed variability is of essential importance in practice, which provides useful information to a wide range of coastal and marine activities. In this paper, long-term wind speed data recorded at various offshore stations are analyzed in the framework of fractal dimension analysis. Fractal analysis is a well-established data analysis tool, which is particularly suitable to determine the complexity in time series from a quantitative point of view. The fractal dimension is estimated using the conventional box-counting method. The results suggest that the wind speed data are generally fractals, which are likely to exhibit a persistent nature. The mean fractal dimension varies from 1.31 at an offshore weather station to 1.43 at an urban station, which is mainly associated with surface roughness condition. Monthly variability of fractal dimension at offshore stations is well-defined, which often possess larger values during hotter months and lower values during winter. This is partly attributed to the effect of thermal instability. In addition, with an increase in measurement interval, the mean and minimum fractal dimension decrease, whereas the maximum and coefficient of variation increase in parallel.

German Bight storm activity, 1897–2018

AbstractThis study investigates the evolution of German Bight (southeastern North Sea) storminess from 1897 to 2018 through analysing upper quantiles of geostrophic wind speeds, which act as a proxy for past storm activity. Here, geostrophic wind speeds are calculated from triplets of mean sea level pressure observations that form triangles over the German Bight. The data used in the manuscript are provided by the International Surface Pressure Databank and the national meteorological services of Denmark, Germany, and the Netherlands. The derivation of storm activity is achieved by enhancing the established triangle proxy method via combining and merging storminess time series from numerous partially overlapping triangles in an ensemble‐like manner. The utilized approach allows for the construction of robust, long‐term and subdaily German Bight storminess time series. Further, the method provides insights into the underlying uncertainty of the time series. The results show that storm activity over the German Bight is subject to multidecadal variability. The latest decades are characterized by an increase in activity from the 1960s to the 1990s, followed by a decline lasting into the 2000s and below‐average activity up until present. The results are backed through a comparison with reanalysis products from four datasets, which provide high‐resolution wind and pressure data starting in 1979 and offshore wind speed measurements taken from the FINO‐WIND project. This study also finds that German Bight storminess positively correlates with storminess in the NE Atlantic in general. In certain years, however, notably different levels of storm activity in the two regions can be found, which likely result from shifted large‐scale circulation patterns.

Research on Offshore Short-term Wind Speed Prediction Based on the CSA Modeling Improved by Random Algorithm

Aiming at improving the prediction accuracy of offshore short-term wind speed, a model based on random cuckoo search algorithm (RCSA) and artificial neural network (ANN) was proposed. Firstly, RCSA was obtained by introducing a random factor to improve CSA, and then a RCSA-ANN model for predicting offshore short-term wind speed was established. Secondly, a wind tower was built in Luchao Port, Shanghai, offshore meteorological data were measured, and the training of the model was carried out. Finally, the precision of RCSA-ANN model was verified by comparison and analysis with BP-ANN and CSA-ANN models. The results show that the improved CSA method is simple, reliable and effective, which solves the problem that the algorithm is easy to fall into local optimum. The average error of RCSA-ANN model is not only lower than that of BP-ANN model, but also much lower than that of CSA-ANN model, and the prediction accuracy of these three models decreases in turn. RCSA-ANN model has high prediction accuracy and can precisely predict fluctuating wind speed sequences, and it also has good application potential.

Comparative study of onshore and offshore wind characteristics and wind energy potentials: A case study for southeast coastal region of China

Wind farms have been developed at both onshore and offshore sites to ease the global energy crisis in recent decades. Nevertheless, the differences between onshore and offshore wind characteristics and their wind energy potentials have rarely been studied. Based on two original sets of wind records respectively obtained from onshore and offshore wind measurement towers in southeast coastal region of China, this paper presents comparative evaluation for the onshore and offshore wind characteristics and their wind energy potentials in this region. The Weibull distribution function is firstly verified to be a reliable model for both onshore and offshore wind speed predictions. Furthermore, the variations of mean wind speed, Weibull parameters and wind power density are studied through monthly and hourly time scales. Finally, the economic feasibility for the wind turbine installations at the onshore and offshore site is examined in terms of the operating probability and the annual energy production (AEP) of a selected wind turbine. The results show that the offshore wind energy is more even and available than the onshore wind energy for the southeast coastal region. For the sake of economic feasibility and annual energy production, it is recommended to choose the hub heights of the selected wind turbine as 70 m and 40 m at the onshore site and offshore site, respectively.

Europe's offshore winds assessed with synthetic aperture radar, ASCAT and WRF

Abstract. Europe's offshore wind resource mapping is part of the New European Wind Atlas (NEWA) international consortium effort. This study presents the results of analysis of synthetic aperture radar (SAR) ocean wind maps based on Envisat and Sentinel-1 with a brief description of the wind retrieval process and Advanced Scatterometer (ASCAT) ocean wind maps. The wind statistics at 10 and 100 m above mean sea level (a.m.s.l.) height using an extrapolation procedure involving simulated long-term stability over oceans are presented for both SAR and ASCAT. Furthermore, the Weather Research and Forecasting (WRF) offshore wind atlas of NEWA is presented. This has 3 km grid spacing with data every 30 min for 30 years from 1989 to 2018, while ASCAT has 12.5 km and SAR has 2 km grid spacing. Offshore mean wind speed maps at 100 m a.m.s.l. height from ASCAT, SAR, WRF and ERA5 at a European scale are compared. A case study on offshore winds near Crete compares SAR and WRF for flow from the north, west and all directions. The paper highlights the ability of the WRF model to simulate the overall European wind climatology and the near-coastal winds constrained by the resolution of the coastal topography in the WRF model simulations.

Deep Learning Applications with Practical Measured Results in Electronics Industries

This editorial introduces the Special Issue, entitled “Deep Learning Applications with Practical Measured Results in Electronics Industries”, of Electronics. Topics covered in this issue include four main parts: (I) environmental information analyses and predictions, (II) unmanned aerial vehicle (UAV) and object tracking applications, (III) measurement and denoising techniques, and (IV) recommendation systems and education systems. Four papers on environmental information analyses and predictions are as follows: (1) “A Data-Driven Short-Term Forecasting Model for Offshore Wind Speed Prediction Based on Computational Intelligence” by Panapakidis et al.; (2) “Multivariate Temporal Convolutional Network: A Deep Neural Networks Approach for Multivariate Time Series Forecasting” by Wan et al.; (3) “Modeling and Analysis of Adaptive Temperature Compensation for Humidity Sensors” by Xu et al.; (4) “An Image Compression Method for Video Surveillance System in Underground Mines Based on Residual Networks and Discrete Wavelet Transform” by Zhang et al. Three papers on UAV and object tracking applications are as follows: (1) “Trajectory Planning Algorithm of UAV Based on System Positioning Accuracy Constraints” by Zhou et al.; (2) “OTL-Classifier: Towards Imaging Processing for Future Unmanned Overhead Transmission Line Maintenance” by Zhang et al.; (3) “Model Update Strategies about Object Tracking: A State of the Art Review” by Wang et al. Five papers on measurement and denoising techniques are as follows: (1) “Characterization and Correction of the Geometric Errors in Using Confocal Microscope for Extended Topography Measurement. Part I: Models, Algorithms Development and Validation” by Wang et al.; (2) “Characterization and Correction of the Geometric Errors Using a Confocal Microscope for Extended Topography Measurement, Part II: Experimental Study and Uncertainty Evaluation” by Wang et al.; (3) “Deep Transfer HSI Classification Method Based on Information Measure and Optimal Neighborhood Noise Reduction” by Lin et al.; (4) “Quality Assessment of Tire Shearography Images via Ensemble Hybrid Faster Region-Based ConvNets” by Chang et al.; (5) “High-Resolution Image Inpainting Based on Multi-Scale Neural Network” by Sun et al. Two papers on recommendation systems and education systems are as follows: (1) “Deep Learning-Enhanced Framework for Performance Evaluation of a Recommending Interface with Varied Recommendation Position and Intensity Based on Eye-Tracking Equipment Data Processing” by Sulikowski et al. and (2) “Generative Adversarial Network Based Neural Audio Caption Model for Oral Evaluation” by Zhang et al.

Investigating the potential impact of wind farms on Lake Erie

Power generation is a leading cause of air pollution and major source of global warming emissions. Renewable energy resources, like wind and solar power, generate electricity with little to no global warming emissions and are reliable, affordable and beneficial for health, economy and climate. There are two sectors in wind energy: offshore and onshore. Offshore wind speeds are faster and steadier and coastal areas often exhibit a high energy demand. Offshore wind farms are coming to the Great Lakes. We therefore investigate the potential impact of wind farms on Lake Erie’s dynamic and thermal structure using the COHERENS (a Coupled Hydrodynamical-Ecological model for Regional and Shelf Seas) and simulate a large wind farm with 432 offshore turbines located in the shallow southern waters of the central basin. The simulation is run twice to compare physical parameters such as temperature and circulation pattern and velocity results in the absence and presence of a large wind farm. In the case of no wind farm, model results are validated with data from buoys located in Lake Erie, while with the wind farm, the results show that the central basin is impacted by the wind turbines. This occurs because the reduced wind speed and stress leads to less mixing, lower current speeds and higher surface water temperature. There is no significant impact, however, in the eastern and western basins. This research examines Lake Erie, since this lake has a high potential for offshore wind turbine installation due to its proximity to population centers and its shallow depth.

Spatio-Seasonal Variations in Long-Term Trends of Offshore Wind Speeds Over the Black Sea; an Inter-Comparison of Two Reanalysis Data

Spatio-seasonal variability of long-term trends in mean and 95th percentile wind speeds for the term between 1979 and 2016, over the Black Sea is presented. Our aim is to contribute the existing literature by presenting the inhomogeneous spatial distribution of the long-term trends in both moderate and severe wind speeds on a monthly basis. The analysis is conducted by using two different data; European Centre for Medium-Range Weather Forecasts-ERA-Interim and U.S. National Centers for Environmental Prediction-Climate Forecast System Reanalysis (CFSR) to perform a comparative analysis. The non-parametric Mann–Kendall and Sen’s Slope methods are used to determine the trends and their significance over the Black Sea. CFSR winds presented higher interannual variability than the ERA-Interim. ERA-Interim indicates that annual mean and 95th percentile wind speeds have decreasing trends down to − 0.17%/year and − 0.20%/year in the Sea of Azov, while they have an increasing trend up to 0.35%/year and 0.38%/year in the eastern part, respectively. Results indicate that wind speeds are increasing over 28% ~ 36% of the Black Sea surface area while the wind speeds are decreasing over 2% ~ 4% of the surface area. Pacific North American Oscillation presented an influence almost all over the Black Sea with statistically significant correlation coefficients over 0.5. North Atlantic Oscillation dominates over the southwestern, western and northern Black Sea with inverse correlation coefficients over 0.6. ERA-Interim and CFSR data illustrated a similar distribution pattern over the Black Sea in means of the relation of variations in wind speeds to the teleconnection indices.

Statistical and synoptic analyses of offshore wind variations

In this study, variations of offshore wind speed in the northern and central coasts of British Columbia are examined using wind observations from nine buoys distributed in the region. Wind speeds from all buoys are extrapolated to the standard (10 m) wind measurement height as well as to the wind turbine hub height (100 m) using Monin–Obukhov similarity theory. Sustained winds above several thresholds are analysed and values of 50 and 100‐year return extreme wind speed levels are calculated for all locations. The percentage of sustained winds between the wind turbine cut‐in and cut‐out wind speed thresholds at 100 m indicates short periods without power generation. The highest 50 and 100‐year return level values are found at locations west of Haida Gwaii, while the lowest values are calculated at the most southern and eastern locations. Spatial coherency analysis of high wind speeds between all locations shows high same‐day coherency ratios between all buoys located offshore of the mainland coast, which emphasizes the importance of the atmospheric circulation at the synoptic‐scale as the main driver of intense wind events; consequently, a synoptic‐scale circulation analysis is carried out by applying principal components analysis and k‐means clustering, and then relating the calculated synoptic patterns to some of the previously classified wind categories. The synoptic‐scale circulation analysis reveals that high wind speed events are mainly associated with a system of low pressure located west or northwest of the study region, which induces intense southerly to southeasterly winds. Low wind speeds are often associated with a high‐pressure system (Pacific High) originating southwest of the region that mainly predominates during the summer.

The Potential of Advanced Scatterometer (ASCAT) 12.5 km Coastal Observations for Offshore Wind Farm Site Selection in Irish Waters

The offshore wind industry has seen unprecedented growth over the last few years. In line with this growth, there has been a push towards more exposed sites, farther from shore, in deeper water with consequent increased investor risk. There is therefore a growing need for accurate, reliable, met-ocean data to identify suitable sites, and from which to base preliminary design and investment decisions. This study investigates the potential of hyper-temporal satellite remote sensing Advanced Scatterometer (ASCAT) data in generating information necessary for the optimal site selection of offshore renewable energy infrastructure, and hence providing a cost-effective alternative to traditional techniques, such as in situ data from public or private entities and modelled data. Five years of the ASCAT 12.5 km wind product were validated against in situ weather buoys and showed a strong correlation with a Pearson coefficient of 0.95, when the in situ measurements were extrapolated with the log law. Temporal variations depicted by the ASCAT wind data followed the same inter-seasonal and intra-annual variations as the in situ measurements. A small diurnal bias of 0.12 m s−1 was observed between the descending swath (10:00 to 12:00) and the ascending swath (20:30 to 22:30), indicating that Ireland’s offshore wind speeds are slightly stronger in the daytime, especially in the nearshore areas. Seasonal maps showed that the highest spatial variability in offshore wind speeds are exhibited in winter and summer. The mean wind speed extrapolated at 80 m above sea level showed that Ireland’s mean offshore wind speeds at hub height ranged between 9.6 m s−1 and 12.3 m s−1. To best represent the offshore wind resource and its spatial distribution, an operational frequency map and a maximum yield frequency map were produced based on the ASCAT wind product in an offshore zone between 20 km and 200 km from the coast. The operational frequency indicates the percentage of time during which the observed local wind speed is between cut-in (3 m/s) and cut-out (25 m/s) for a standard turbine. The operational frequency map shows that the frequency of the wind speed within the cut-in and cut-off range of wind turbines was between 92.4% and 97.2%, while the maximum yield frequency map showed that between 40.6% and 59.5% of the wind speed frequency was included in the wind turbine rated power range. The results showed that the hyper-temporal ASCAT 12.5 km wind speed product (five consecutive years, two observations daily per satellite, two satellites) is representative of wind speeds measured by in situ measurements in Irish waters, and that its ability to depict temporal and spatial variability can assist in the decision-making process for offshore wind farm site selection in Ireland.

Assessing the climate change impact on the North African offshore surface wind and coastal low-level jet using coupled and uncoupled regional climate simulations

The North African Coastal Low-Level Jet (NACLLJ) is a semi-permanent feature offshore the north western African coast, linked to the cold nearshore upwelling of the Canary Eastern Boundary Current system. Its main synoptic drivers are the Azores Anticyclone over the ocean and the inland Sahara thermal low. The coastal jet events occur in one of the world’s most productive fisheries region, thus the evaluation of the effects of global warming in its properties is imperative. This study proposes an analysis of the annual and intra-annual attributes of the NACLLJ for two time periods 1976–2005 (historical) and 2070–2199 (future), resorting to coupled and uncoupled atmosphere–ocean simulations with the ROM model, as well as near surface offshore wind speed from the CORDEX-Africa ensemble. The future simulations follow the RCP8.5 greenhouse gas emissions scenario. Overall, the ROM coupled simulation presents the best performance in reproducing the present-climate near surface wind speed, offshore northwest Africa, compared to the remaining RCM simulations. The higher SST resolution in the coupled simulations favours much localised colder upwelling strips near the coast and consequently stronger jets. In future climate, a small increase in the surface wind speed is projected, mainly linked to the regions of coastal jet presence. The NACLLJ is projected to be more frequent and intense, encompassing larger areas. An increase of the jet seasonal frequencies of occurrence is projected for all seasons, which is larger from spring to autumn (up to 15, 16 and 22% more frequent, respectively). However, in some offshore areas the winter NACLLJ persistency is likely to double, relatively to present-climate. Higher inter-annual variability is also projected for the future NACLLJ seasonal frequencies. The strengthening of the coastal jet speeds is also significant, between 5 and 12% in all seasons. Additionally, the jet’s diurnal cycle shows an increase in jet occurrence across the day, particularly in the mid and late afternoon.

Evaluation of ERA-Interim and NCEP-CFSR Reanalysis Datasets against in-situ Measured Wind Speed Data for Keti Bandar Port, Pakistan

The evaluation and comparison of reanalysis datasets (NCEP-CFSR and ERA-Interim) against mast measured wind data for Keti-bander port of Pakistan have been performed. In the last decade, wind farms have been developed on the land, but there is still need for the development of offshore wind farms on the 1100 km coast of Pakistan. The wind data measured at the Keti Bandar Port of Pakistan at different heights for the duration of 20 months has been used in this study. The evaluations of satellite datasets against measured data were done based on statistical analysis, monthly mean time series and Weibull probability distribution function. The wind data at 30 m, 60 m and 85 m was extrapolated using the Logarithmic law from 10 m height. At 30 m height both satellite datasets show very good agreement with measured data, the predictions of CFSR data are much better than ERA-Interim. The percentage difference for mean wind speed, wind power density and Weibull parameters at 30 m are less than 5%. Overall NCEP-CFSR gave better results compared to ERA-Interim. These results may be used for the offshore wind resource assessment for Pakistan due to unavailability of offshore wind speed data.

Grid-Style Modular Network Self-Organization Map for Detail Projection of Unknown Off-Shore Wind Speed

Grid-Style Modular Network Self-Organization Map for Detail Projection of Unknown Off-Shore Wind Speed

A New Proposal for Vertical Extrapolation of Offshore Wind Speed and an Assessment of Offshore Wind Energy Potential for the Hibikinada Area, Kitakyushu, Japan

The author’s research group has been conducting research on applications of various meteorological Grid Point Value (GPV) data offered by the Japan Meteorological Agency (JMA) to the field of wind power generation. In particular, the group’s research has been focusing on the following areas: 1) the use of GPV data from the JMA Meso-Scale Model (MSM-S; horizontal resolution: 5 km) and the JMA Local Forecast Model (LFM-S; horizontal resolution: 2 km), and 2) examinations of the prediction accuracy of local wind assessment with the use of these data. In both the MSM-S and the LFM-S, grid points are fixed at 10 m above the sea (ground) surface. The purpose of the present study is to establish a method in which the values of the MSM-S and LFM-S wind speed data from the 10 m height are used as the reference wind speed and are, using a power law, vertically extrapolated to 80 to 90 m heights, typical hub-heights of offshore wind turbines. For this purpose, the present study examined time-averaged vertical profiles of wind speed over the ocean based on the MSM-S data and in-situ data in the Hibikinada area, Kitakyushu City, Fukuoka Prefecture, Japan. As a result, it was revealed that a strong wind shear existed close to the sea surface, between the 10 and 30 m heights. In order to address the above-mentioned wind shear, a two-step vertical extrapolation method was proposed in the present study. In this method, two values of N, specifically for low and high altitudes (below and above approximately 30 m, respectively), were calculated and used. The data were created for the five years between 2012 and 2016. Similarly to previous analyses, the analysis of the created data set indicated that the potential of offshore wind power generation in the Hibikinada area, Kitakyushu City is quite high.

Offshore wind speed estimates from a high‐resolution rapidly updating numerical weather prediction model forecast dataset

AbstractIn association with the Department of Energy–funded Position of Offshore Wind Energy Resources (POWER) project, we present results from compositing a 3‐year dataset of 80‐m (above ground level) wind forecasts from the 3‐km High‐Resolution Rapid Refresh (HRRR) model over offshore regions for the contiguous United States. The HRRR numerical weather prediction system runs once an hour and features hourly data assimilation, providing a key advantage over previous model‐based offshore wind datasets. On the basis of 1‐hour forecasts from the HRRR model, we highlight the different climatological regimes of the nearshore environment, characterizing the mean 80‐m wind speed as well as the frequency of exceeding 4, 12, and 25 m s−1 for east and west coast, Gulf of Mexico, and Great Lake locations. Preliminary verification against buoy measurements demonstrates good agreement with observations. This dataset can inform the placement of targeted measurement systems in support of improving resource assessments and wind forecasts to advance offshore wind energy goals both in New England and other coastal regions of the United States.

Uncertainty modeling for chaotic time series based on optimal multi-input multi-output architecture: Application to offshore wind speed

Wind energy is attracting more attention with the growing demand for energy. However, the efficient development and utilization of wind energy are restricted due to the intermittency and randomness of wind speed. Although abundant investigations concerning wind speed forecasting have been conducted by numerous researchers, most of the studies merely attach importance to point forecasts, which cannot quantitatively characterize the uncertainties as developing intervals. In this study, a novel interval prediction architecture has been designed, aiming at constructing effective prediction intervals for a wind speed series, composed of a preprocessing module, a feature selection module, an optimization module, a forecast module and an evaluation module. The feature selection module, in cooperation with the preprocessing module, is developed to determine the optimal model input. Furthermore, the forecast module optimized by the optimization module is considered a predictor for giving prediction intervals. The experimental results shed light on the architecture that not only outperforms the benchmark models considered, but also has great potential for application to wind power systems.