遗传神经网络在风电场风速估计中的应用

Wind Speed Forecasting Using Genetic Neural Model

风速的波动会影响风电场发电量的随机变化,对风速比较准确的估计可用于提高风力发电控制系统的性能。但由于风速的随机性和非线性,用常规的方法难以预测。将时间序列分析法、神经网络及遗传算法相结合,提出一种预测风速的建模方法。采用时序分析法确定网络输入变量数,用遗传算法动态调整BP网络的连接权值和阈值的方法来逼近和学习风速的非线性。实验结果表明,该模型性能良好,能有效跟踪风速和风力机发电功率的变化趋势,具有良好的估计精度。

Considering the stochastic behaviour of wind speed, a predication model is established which based on time series analytic approach, genetic algorithm and neural network, The number of input variable in this network is decided by the time series analytic approach. The predicted values agree well with the data which measured in a wind farm. The experiment results showed that the proposed method is a valid, and is useful tool for wind speed and wind turbine power predication.