Abstract
According to the requirements of aeroengine performance control, a neural network called auto-tuning neurons and gradient descent learning method are presented. A multivariable decoupling control algorithm based on the auto-tuning neurons is used for aeroengine multivariable control systems. The main difference between an auto-tuning neuron and a general neuron is that there are adjustable parameters of the activation function used in an auto-tuning neuron. Unlike traditional fully connected neural network, there are no synaptic connections among the independent neurons. The emphasis is focused on the research of the algorithm and the properties of the controller, as well as their application to the aeroengine control by means of computer simulation. Finally the aeroengine multivariable control system is designed. Simulation shows that the system has perfect performance of decoupling and adaptive capabilities.
| Original language | English |
|---|---|
| Pages (from-to) | 490-494 |
| Number of pages | 5 |
| Journal | Hangkong Dongli Xuebao/Journal of Aerospace Power |
| Volume | 22 |
| Issue number | 3 |
| State | Published - Mar 2007 |
Keywords
- Adaptive control
- Aerospace propulsion system
- Aroengine
- Auto-tuning neurons
- Decoupling control
- Multivariable control