基于盲数和神经网络的电价预测模型

Model of price forecasting based on blind number and artificial neural networks

利用盲数理论结合经典电价预测方法,在有限的历史电价、负荷及其他相关数据的基础上对电价进行准确预测。首先提出了基于盲数和神经网络市场出清电价预测模型,利用BP神经网络对历史数据进行训练学习,在得到网络学习权重后,结合盲数理论,引用盲数代替实数进行价格预测。算例结果表明,模型消除不确定性因素对预测结果的影响,实际历史电价都很好地落在了预测结果最大置信度的置信区间中,较好地完成了预测任务,证实了设计的可行性和模型的可靠性。

The paper utilizes blind number theory with classical price forecasting ways to forecast electrical price accurately which is based on limited historical electrical prices, loads and other related data. It proposes model of market clearing price forecasting which is based on blind number and artificial neural networks. This model uses BP neural networks to train and learn from historical data. After getting weights of networks, networks use blind number instead of real number to forecast price. The results of examples show that the model can eliminate the influence of uncertain factors to forecasting result, as well as historical prices are all in the believable inter zone which has the biggest belief of forecasting results. The model completes the forecasting mission well which proves the feasibility of design and reliability of model.