利索-全球专利检索

  1. 登录
  2. 注册
发明公开
CN109782606A 基于改良型引力搜索方法的递回式小波艾尔曼神经网络 无效 - 撤回

基于改良型引力搜索方法的递回式小波艾尔曼神经网络
摘要：
本发明公开基于改良型引力搜索方法的递回式小波艾尔曼神经网络,递回式小波艾尔曼神经网络包括五层神经网络配置，分别为输入层、隐藏层、上下文层、输出反馈层和输出层；改良型引力搜索方法包括以下步骤：初始化、适应性函数、小波突变、族群更新、计算不同方向的总力、加速度计算、更新粒子的速度与位置、重复步骤二至七，直到达到停止标准。本发明递回式小波艾尔曼神经网络基于改良型引力搜索算法的递回式小波艾尔曼神经网络应用于韦尔斯涡轮机叶片角度及发电机之转速来达成波浪发电，可让具备高占比与高渗透率的再生能源并网发电的电力系统维持良好的动态特性与瞬时稳定度。
摘要（英）：

The invention discloses a recursive wavelet Elman neural network based on an improved gravity search method. The recursive wavelet Elman neural network comprises five-layer neural network configuration including an input layer, a hidden layer, a context layer and an output feedback layer. The improved gravity search method includes steps: carrying out initialization; carrying out adaptive functionprocessing; carrying out wavelet mutation; carrying out cluster group updating; calculating total forces in different directions; calculating an acceleration speed, updating a particle speed and position; and repeating steps from the step two to step seven till reaching a stopping standard. According to the invention, the recursive wavelet Elman neural network is applied to the Wells turbine blade angle and generator rotating speed to realize wave power generation; and the high-proportion high-permeability power system based on renewable energy for grid-connected power generation keeps the good dynamic characteristics and high transient stability.

0/0