CAJ | 학술논문

提高短期风电功率的预测精度对保证电力系统安全、稳定地运行具有重大意义.针对风速信号的强奇异性,采用脊波神经网络建立短期风电功率的预测模型;同时利用混沌 DNA 遗传算法确定脊波神经网络的隐层结构,采用粒子群算法优化网络的连接权值及方向向量.对新疆某风电场的输出功率进行了预测实验,并比较了优化前后脊波网络模型的预测性能.研究结果表明采用粒子群与混沌 DNA 遗传算法组合优化后的脊波神经网络均方根误差降至12.3%,预测精度得到显著提高.
제고단기풍전공솔적예측정도대보증전력계통안전、은정지운행구유중대의의.침대풍속신호적강기이성,채용척파신경망락건립단기풍전공솔적예측모형;동시이용혼돈 DNA 유전산법학정척파신경망락적은층결구,채용입자군산법우화망락적련접권치급방향향량.대신강모풍전장적수출공솔진행료예측실험,병비교료우화전후척파망락모형적예측성능.연구결과표명채용입자군여혼돈 DNA 유전산법조합우화후적척파신경망락균방근오차강지12.3%,예측정도득도현저제고.
Increasing the prediction accuracy of short-term wind power plays a key role in improving the security and stability of electric grid system. Aiming at the characteristics of wind speed signal, this paper uses the Ridgelet Neural Network for building the short-term wind power forecasting model. At the same time we use the chaos DNA genetic to identify the hidden layer of Ridgelet Neural Network, and use the Particle Swarm Optimization algorithm to adjust the weight value and direction vector. Based on the measured historical data of wind farm in Xinjiang, we did a experiment on wind power forecasting and compared the performance of the optimized model with that of the original one. Experimental results indicate that the RMSE(root mean squared error)of Ridgelet Neural Network can reduce to 12. 3% by using the combination of chaos DNA genetic and particle swarm optimization algorithm, and the prediction accuracy is improved significantly.