基于ANFIS模型的风功率超短期预测研究

摘要

提出一种基于自适应神经模糊推理系统(ANFIS)模型的风功率超短期预测方法。该方法以测风塔在层高70m测得的风向、风速和在层高20m测得的气温为输入，以风场的实际有功出力为输出，选取山西某风场的历史数据进行训练，建立了风功率超短期ANFIS预测模型。最后将ANFIS预测结果与基于BP神经网络的预测结果进行了对比分析，结果表明本文提出的预测方法预测精度较高。%An ultra-short-term forecasting method is proposed based on ANFIS for wind power. The wind direction, wind speed which are measured in height of 70m from anemometer tower and temperature which is measured in height of 20m from anemometer tower, take these data as input and actual active power of the wind farm as output. The ultra-short-term forecasting model is established on the basis of ANFIS for wind power by training the historical data of a wind farm in Shanxi. The experimental results are compared with that of BP neural networks. And result from the study shows that ANFIS may give a high precision.