论文部分内容阅读
为了解决在电力系统中,因为系统规模巨大而导致的电力系统负荷的预测难度增大问题,利用较为先进的递归型人工神经网络进行预测模型建立,同时最大可能地缩减储备池的规模,以保证神经网络在该问题上体现出优异的时间序列预测与非线性建模性能,本文构建了一种基于非线性读出器的新型液体状态机模型。面对基于概率神经网络输出器的液体状态机模型在分类和预测问题上所具备的优秀性能,本文设计了基于概率神经网络输出器的液体状态机电力负荷预测模型,同时与普通的递归型神经网络进行了对比实验。实验结果表明,利用这种非线性读出的液体状态机模型可以较为有效地对电力负荷进行预测,所得到的预测效果对于电力的调度和合理利用有一定的参考价值。
In order to solve the problem of increasing the power system load forecasting difficulty due to the huge system size in the power system, the advanced recursive artificial neural network is used to establish the prediction model and reduce the reserve pool size as much as possible to ensure Neural network shows excellent performance of time series prediction and nonlinear modeling on this problem. In this paper, a new model of liquid state machine based on nonlinear reader is constructed. Faced with the excellent performance of the liquid state machine model based on probabilistic neural network output device in classification and forecasting problems, this paper designs a liquid state machine power load forecasting model based on probabilistic neural network output device, and simultaneously with the ordinary recursive nerve Network conducted a comparative experiment. The experimental results show that the proposed model can be used to forecast the power load more effectively by using the non-linear readout model of liquid state machine. The predictive effect obtained is of certain reference value for the power scheduling and rational utilization.