CAJ | 학술논문

风速的随机性和波动性较强，单一算法预测模型的精度不高。为此，提出基于蚁群优化算法的灰色模型和遗传神经网络算法相结合的预测方法；给出改进灰色模型，并利用蚁群算法的全局寻优能力，以残差平方和最小为原则，对改进灰色模型的权值进行优化，实现了对风速的预测。为进一步提高精度，把蚁群优化灰色模型的结果作为遗传神经网络算法的输入，实测风速数据作为遗传神经网络算法的输出，对网络进行训练，进一步减小了风速预测的偏差。预测结果与风电场实测数据的对比分析验证了多算法结合的预测方法的准确性和有效性。
풍속적수궤성화파동성교강，단일산법예측모형적정도불고。위차，제출기우의군우화산법적회색모형화유전신경망락산법상결합적예측방법；급출개진회색모형，병이용의군산법적전국심우능력，이잔차평방화최소위원칙，대개진회색모형적권치진행우화，실현료대풍속적예측。위진일보제고정도，파의군우화회색모형적결과작위유전신경망락산법적수입，실측풍속수거작위유전신경망락산법적수출，대망락진행훈련，진일보감소료풍속예측적편차。예측결과여풍전장실측수거적대비분석험증료다산법결합적예측방법적준학성화유효성。
Wind speed is with strong randomness and volatility,and the single algorithm model of wind forecasting is with low accuracy.To improve the forecasting precision of wind speed,a new wind speed forecasting method was proposed in this paper,it combined the gray model,the ant colony optimization algorithm and the genetic neural network.The improved grey model was giv-en.Using the global optimization ability of ant colony algorithm,the weights of improved grey model was optimized with the least squares criterion,and the forecast of wind speed could be thus realized.In order to further improve the forecasting precision,genetic neural network was trained to further reduce errors,the results of gray model with ant colony optimization was taken as the inputs of the genetic neural network,and the measured wind speed data was taken as the output. The comparative analysis of forecasting results and field testing results in a certain wind farm showed that the forecasting method was with high accuracy and effectiveness.