相似误差订正方法在风电短期风速预报中的应用研究

STUDY ON APPLICATION OF ANALOG BIAS CORRECTION METHOD TO SHORT-TERM WIND SPEED PREDICTION IN WIND FARMS

关键词:  风电场  风速  短期预报  相似误差订正

Key words:  wind farms  wind speed  short-term prediction  analog bias correction

摘要:

为降低风电场短期预报风速误差，减少风电场短期风功率偏差积分电量，提高风电场发电功率预测准确率，分季节研究了相似误差订正方法对ECMWF单台风机预报风速的订正效果。结果表明：相似误差订正后不同风机预报风速的误差差距减小；预报风速的平均绝对偏差和均方根误差明显降低，其中夏季和秋季华能义岗风电场两个指标降低幅度均超过0.1 m/s、会宁丁家沟风电场均超过0.2 m/s；订正风速削减了原始预报的极值，可反映大部分时段实况风速3 h内的趋势变化，个别时段订正风速与实况趋势相反；订正后预报风速在风功率敏感区的平均绝对偏差明显降低，华能义岗风电场四季降低幅度在0.112~0.242 m/s之间、会宁丁家沟风电场四季降低幅度在0.131~0.430 m/s之间，有效降低了原始预报误差带来的短期风功率偏差积分电量扣分值；订正风速较原始预报更多分布在风功率敏感区。该方法实际应用灵活，对提高风电场短期预报风速准确率有可观的效果，并可有效减少短期风功率偏差积分电量考核。

Abstract:

To reduce the error of short-term wind speed prediction in wind farms, reduce the deviation integral quantity of short-term wind power, and improve the accuracy of wind farm power generation forecasting, the present research studies the effect of analog bias correction method on ECMWF wind speed forecasting of single wind turbine by season. The results show that the error difference between wind speed forecasts for different wind turbines is reduced after similar errors are corrected; the average absolute deviation and root mean square error of the forecasted wind speed are significantly reduced. The corrected wind speed reduces the extreme value of the original forecast, which can reflect the 3-hour trend of the actual wind speed change in most periods of time with exceptions in certain periods. After the correction, the average absolute deviation of the forecasted wind speed in the wind power sensitive area is significantly reduced. The seasonal reduction range of Huaneng Yigang Wind Farm is between 0.112~0.242 m?s-1 and that of Huining Dingjiagou Wind Farm is between 0.131~0.430 m?s-1, effectively reducing the deviation integral charge of short-term wind power caused by the original prediction error. Compared with that in the original forecast, the corrected wind speed happens more frequently in the sensitive area. The analog bias correction method is flexible in practical application. It can effectively decrease the deviation integral quantity of short-term wind power and thus has a considerable effect on improving the accuracy of short-term wind speed prediction on wind farms.