基于ADQPSO-KELM风电功率短期预测模型的研究

为了改善传统风电功率预测方法中误差较大且稳定性较差的问题,引入量子粒子群(QPSO)优化算法、自适应早熟判定准则及混合扰动算子,构建了自适应扰动量子粒子群(ADQPSO)优化算法,通过ADQPSO算法对核极限学习机(KELM)模型进行优化,建立了自适应扰动量子粒子群优化的核极限学习机(ADQPSOKELM)风电功率短期预测模型,并利用内蒙古高尔真风电场采集的风电功率时间序列数据为试验样本进行48h预测分析。结果表明,ADQPSO-KELM风电功率短期预测模型与其他基于KELM优化的风电预测模型及传统风电预测模型相比,其预测的误差更小、准确度更高,且预测稳定性显著增强。

Research on Short-term Wind Power Prediction Model Based on ADQPSO-KELM

In order to solve the problem of large error and poor stability in traditional wind power prediction methods, the quantum particle swarm optimization (QPSO) algorithm, adaptive premature judgment criterion and hybrid perturbation operator were introduced to establish an adaptive perturbation quantum particle swarm optimization (ADQPSO) algorithm. The kernel extremum learning machine (KELM) model was optimized by ADQPSO algorithm. The short-term wind power prediction model based ADQPSO-KELM was established. The wind power time series data collected by Inner Mongolia Gaoerzhen wind farm was used to analyze the 48-hour prediction of the experimental samples. The results show that the short-term wind power prediction ADQPSO-KELM model is more accurate than other prediction models based KELM optimization and traditional prediction models, and the prediction error is small. The prediction stability is significantly enhanced.