|
标题：基于神经网络的航空三级式起动发电机简化一体化模型 |
|
作者：焦宁飞，韩旭，刘卫国 |
|
| 2019年第1期 访问次数：386次 |
|
摘要：航空三级式无刷同步起动发电机（简称三级式起发电机）因其可靠性高、发电技术成熟等优势而在多电飞机起动发电一体化技术研究中受到青睐。针对该电机的详细分析与起动控制研究需要准确而又高效的仿真模型。针对三级式起发电机多级结构特点，本文将其励磁系统视为具有非线性传递特性的特殊“电刷滑环系统”，并通过BP-神经网络拟合获取其传递特性。以励磁系统神经网络简化模型为基础，将包含多级结构的三级式起发电机简化为传统单级式电励磁同步电机，并建立三级式起发电机简化一体化模型。计算机仿真对比分析结果表明，本文所建立的简化一体化模型具有模型简单、仿真精度高、仿真速度快的优势。
关键词：航空起动发电一体化；三级式无刷同步电机；神经网络；电机建模
Abstract: Threestage brushless synchronous starter/generator (TBSSG) is favored in the research about integrated starter/generator technologies in more Electric aircraft for its reliable structure and mature power generation technologies. Detailed analysis and start control for TBSSG need an accurate and efficient simulation model. Considering the multistage structure characteristics of TBSSG, the excitation system was regarded as a special “brush slipring system” with nonlinear transmission characteristics in this paper, and the nonlinear transmission characteristics was obtained through BPneural network fitting method. Based on the simplified neural network model of the excitation system, TBSSG with multistage structure can be simplified into a traditional electrically excited synchronous machine, and then a simplified integrated model of TBSSG was built. Simulation results verified that the proposed simplified integrated model has the advantages of simple structure, high simulation precision and fast simulation speed.
Key words: aircraft starter/generator； threestage brushless synchronous machine； neural network； electric machine modeling
|
|
| |
|
