Research On Wind Power Prediction Method Based On Flow Field Analysis Of Wind Farm

In the whole country,the development of wind farms in complex terrain has increased and in which,the flow field distribution is affected by the coupling of atmospheric environment,turbines’operating characteristics and terrain factors.The spatial and temporal unevenness in the speed distribution is serious and uncertain,which poses a challenge to the wind power prediction.In response to this problem,this paper carried out the theoretical and experimental research on wind farms on key issues such as wake characteristics of complex terrain,and wind power variation of wind turbines under the effect of terrain coupling wake,and proposed a wind power prediction method for complex terrain based on wind farm flow.The main innovations are as follows:A three-dimensional wake model of wind turbines considering wind shear was proposed and the wake characteristics of complex terrain were analyzed.Based on the Jensen model,considering the effect of wind shear and combining the characteristics of Gaussian distribution,a 3D wake model that predicts the wind speed distribution of vertical height plane was proposed based on the idea of curve rotation.Two different types of high-precision Doppler LiDARs were used to conduct field experiments,and the measured data were compared with the predicted wind speed of wake model.At the same time,an experimental study on the wake characteristics of complex terrain wind farms was carried out.Results showed that due to the influence of the incoming wind shear effect,the wind speed distribution of vertical height plane in the wake area presents the characteristics of asymmetric distribution.The wake model considering the wind shear effect can better predict the distribution characteristics of the wind speed distribution of vertical height plane in the wake area.The new wake model had a certain reference value for the power prediction and micro location selection of complex terrain wind farms.The upward wind speed of the wake center increased gradually,and the downward wind speed increased first and then decreased due to the tower shadow effect and ground friction.The wake center sank with the downwind direction altitude drop,and the height of the wake center sank linearly with the altitude drop.The greater the turbulence intensity of the wake area,the faster the wake recovery rate,and the turbulence intensity of the wake area was determined synthetically by inflow turbulence intensity and axial thrust coefficient of wind turbine.The 3D Jensen_Gaussian wake model and the wind speed calculation method of the wind turbine under the mixed wake effect were derived theoretically.3D Jensen_Gaussian wake models without considering wind shear and considering wind shear were derived respectively,which lays a theoretical foundation for the analysis of three-dimensional distribution characteristics of wake.Considering the difference of hub height,turbines’ diameter and control performance of adjacent wind turbines,the calculation method of relative distance of wind turbines in a certain wind direction,the judgment standard of whether the downstream wind turbine is in the wake of the upstream one and the calculation method of wind speed in the hub center and average inflow speed of single wind turbine were put forward.On the basis of wind speed calculation under the single wake effect,the square sum model was used to analyze the wake superposition effect and calculate the wind speed of the wind turbine in the mixed wake area,which provides theoretical support for the analysis of mixed wake effect and the proposal of yaw control strategy.Based on LiDAR measurement and SCADA data,the wake interference law of wind turbine in flat terrain wind farm and its influence on wind turbine power were studied.Typical wind turbines in flat terrain were selected for research,the wake distribution characteristics of different impeller diameters in flat terrain were given,the wake interference law of the same and different impeller diameters were studied respectively,and the influence of wake interference on the power of wind turbines was analyzed.Results showed that the wake characteristics of the wind turbine were closely related to the diameter of the wind turbine,the larger the impeller diameter,the smal ler the wake width,the greater the wake depth and velocity deficit,the slower the wake dissipation,and the more obvious the wake effect.The wake of the upstream wind turbine exacerbated the wake of the downstream wind turbine,which was reflected in the wider wake width,greater speed deficit and depth.The wake of the upstream wind turbine close to the downstream wind turbine would also be affected,which was reflected in the sudden increase of wake width,depth and speed deficit when the wake was close to the downstream wind turbine.When the incoming wind speed changes,the influence of wake effect on the power generation was different,and the power loss increased sharply with the decrease of wind speed.The wake effect of large-scale wind turbine should be paid more attention to,because the wake effect of large-scale wind turbine would cause great power loss even in a long distance.The yaw control strategy of wind turbine based on flow field analysis was studied.The flow field characteristics around the wind turbine in different positions of the wind farm were given.According to the principle of wake offset,the change of the total power of two wind turbines under different yaw angles was analyzed.Based on the traditional restart function of aiming at incoming wind direction,the yaw control strategy and specific implementation process of the wind turbine considering the wake effect were given.Results show that when wind turbines were arranged in series,with the increase of the yaw angle of upstream wind turbine,wind energy utilization coefficient and output power of upstream wind turbine decrease continuously.Wake offset increased first and then decreases with yaw angle,output power of downstream wind turbine increases first and then decreases accordingly,so the yaw angle should not be too large,and the optimal yaw angle is about 25°-30°.The yaw control strategy based on flow field analysis can improve the overall power output of wind turbines to a certain extent,with an increase of 13.5%.A "customized" wind power forecasting method for complex terrain wind farm was proposed based on machine learning.Case study showed that under the influence of topography and wake effect,wind speed and power generation of wind turbines vary greatly under different topography conditions,and the wind power output of wind farm is uncertain,based on which,a "customized" wind power forecasting method based on deep learning was proposed.In spatial dimension,taking into account the measured values of wind conditions in the wind farm,the LS-SVR model was used to predict the distribution of wind conditions.A spatial data mining model between wind measuring points and specific location of the wind turbine was established.Then,the wind power prediction problem of whole wind farm was decomposed into the wind power prediction of specific wind turbine.In time dimension,the multi-step prediction of wind power was carried out by using P-SVR model,which avoids the iterative cumulative error and improves the accuracy of multi-step prediction.Experimental results show that the proposed model has greatly improved the prediction accuracy,prediction time and model efficiency.The research results of this paper were of great significance for improving the accuracy of wind power prediction and the efficiency of wind turbine operation in complex terrain wind farms,which could promote the development of the wind power industry to a certain extent.