| 摘要: |
| 对航空发动机的双变量解耦控制方法进行了研究,提出了一种基于遗传算法的PID神经网络解耦控制算法。该算法将遗传算法用于多层前向神经网络的连接权系数的学习,克服了BP算法易陷入局部权值的缺点,并具有PID神经网络控制器结构简单规范、动态和静态性能良好等优点。 |
| 关键词: 航空发动机 遗传算法 神经网络 解耦 多变量系统 |
| DOI： |
| 分类号:V233.7 |
| 基金项目: |
|
| A PID neural network decoupling control of aeroengine |
|
FU Qiang,FAN Ding
|
|
1.School of Flight Technique,Civil Aviation Flight Univ.of China,Guang’Han 618307,China;2.School of Power and Energy,Northwestern Polytechnical Univ.,Xi’an 710072,China
|
| Abstract: |
| A PID neural network decoupling method based on genetic algorithm is investigated for aeroengine binary control system.Genetic algorithm is applied in training weighting factor of multilayered forward neural network.Thus the disadvantage of BP algorithm which would easily fall into partial extreme value can be overcome.The advantage of PID neural network controller includes simple structure and good dynamic as well as static performance.This method can be applied to the decoupling control of the multivariable system such as aeroengine. |
| Key words: Aeroengine Genetic algorithm Neural network Decoupling Multivariable system |
