提出了一种改进的动态模糊神经网络DFNN(Dynam ic Fuzzy Neural Network)的短期电价预测方法。首先对采集到的信息进行特征提取,然后利用模糊粗糙集理论中的信息熵进行属性简化、去掉冗余信息,最后用得到的属性作为动态模糊神经网络(DF...提出了一种改进的动态模糊神经网络DFNN(Dynam ic Fuzzy Neural Network)的短期电价预测方法。首先对采集到的信息进行特征提取,然后利用模糊粗糙集理论中的信息熵进行属性简化、去掉冗余信息,最后用得到的属性作为动态模糊神经网络(DFNN)的输入进行训练预测。在模糊神经网络内部引入递归环节,构成了动态模糊神经网络,并采用具有全局寻优能力的遗传算法来训练网络,克服了单纯BP算法易陷入局部最优解的困境。最后以美国加州电力市场公布的2000年数据进行了模型训练和预测,结果表明该方法所建立的预测模型具有较高的预测精度。展开更多
A new method of switched reluctance wind power generation position sensorless based on DFNN by FEA was proposed, Through current and magnetic linkage to get the angle of SRG rotor position, the nonlinear mapping of cu...A new method of switched reluctance wind power generation position sensorless based on DFNN by FEA was proposed, Through current and magnetic linkage to get the angle of SRG rotor position, the nonlinear mapping of cur- rent-magnetic linkage-angle was built, By training these sample data from FEA, the angle of SRG rotor position was replaced by the output of DFNN to achieve SRG position sensorless. Simulation results show that the error between actual rotor position and estimate rotor position is small; SRG can commutate with great accuracy; and the output voltage of SRG wind power system under variable wind speed is essentially constant.展开更多
文摘提出了一种改进的动态模糊神经网络DFNN(Dynam ic Fuzzy Neural Network)的短期电价预测方法。首先对采集到的信息进行特征提取,然后利用模糊粗糙集理论中的信息熵进行属性简化、去掉冗余信息,最后用得到的属性作为动态模糊神经网络(DFNN)的输入进行训练预测。在模糊神经网络内部引入递归环节,构成了动态模糊神经网络,并采用具有全局寻优能力的遗传算法来训练网络,克服了单纯BP算法易陷入局部最优解的困境。最后以美国加州电力市场公布的2000年数据进行了模型训练和预测,结果表明该方法所建立的预测模型具有较高的预测精度。
基金Supported by the National Natural Science Foundation of China （50977080） the Science ＆ Technology Department Project of Hunan Province （2010F J3116） the Education Department Project of Hunan Province （ 10A 114）
文摘A new method of switched reluctance wind power generation position sensorless based on DFNN by FEA was proposed, Through current and magnetic linkage to get the angle of SRG rotor position, the nonlinear mapping of cur- rent-magnetic linkage-angle was built, By training these sample data from FEA, the angle of SRG rotor position was replaced by the output of DFNN to achieve SRG position sensorless. Simulation results show that the error between actual rotor position and estimate rotor position is small; SRG can commutate with great accuracy; and the output voltage of SRG wind power system under variable wind speed is essentially constant.
