鄱阳湖区风电场风电功率预报研究

Wind Power Forecasting Research of Wind Farm Around Poyang Lake

风力发电作为一种无污染可再生的能源,已逐渐成为许多国家能源战略可持续发展的重要组成部分。风电场风能预报是风力发电开发中的关键技术问题。为研究鄱阳湖区风力发电预报技术,采用中尺度模式WRF和微尺度模块CALMET对鄱阳湖区长岭风电场进行了200 m水平分辨率风能预报,并根据长岭机组理论功率曲线表和实测数据拟合出理论和实际发电机组功率曲线模型及平均有功功率与发电量模型。根据WRF+CALMET模式预报风速及建立的发电机组功率曲线模型和平均有功功率与发电量模型,预报了长岭风电场发电量。结果表明:长岭风电场23座风机逐小时风速预报值与观测值相关系数为0.42~0.61,均方根误差为2.59~3.68,相对误差为-13.7%~17.4%;对整个风场,预报风速与观测风速的相关系数为0.55,均方根误差为2.8,相对误差为-4.79%。实测发电量值高于预报值,平均偏大39.7 kW,相对误差为-12.6%,预报值与实测值相关性较好,相关系数达到0.52。总体来说,根据中尺度数值模式预报的风速结合风功率、发电量模型预测出的发电量与实测值较为接近,但各月差异性较大。

As a non-polluting and renewable energy source,wind power has gradually become an important component of sustainable development of national energy strategy.Wind power forecast is a key technical problem in the development of wind power generation.In order to research the forecast technology of wind power generation in Poyang Lake area,the mesoscale weather numerical model WRF and the high resolution model CALMET were used to predict the wind energy parameters in Changling wind farm in Poyang Lake area,the predicted horizontal resolution was 200 meters.In addition,according to theoretical power curve and measured data of wind turbine generator system in Changling wind farm,the theoretical and actual generator power curves,the average active power and the generation model was fitted,the model fitting error was tested and analyzed.We used the forecasting wind speed from WRF+CALMET model,and according to actual generator power curves,the average active power and the generation mode,we predicted the wind power generation of Changling wind farm.The results show as follows.The correlation between wind speed forecast values and observation values of 23 fans in Changling wind field are 0.420.61,the root mean square errors are 2.59 to 3.68.The relative errors are-13.7%to 17.4%.To the whole wind field,the correlation coefficient between forecast value and observed value is 0.55,the root mean square error is 2.8,the relative error is-4.79%.The measured electricity generation is 39.7 kW higher than the prediction,the relative error is-12.6%,the correlation between the predicted value and the measured value is acceptable,while the correlation coefficient is 0.52.In the whole,the prediction result of power generation from wind speed mesoscale model combined with wind power and generation mode is relatively close to the measured value,although the results vary greatly from month to month.