论文部分内容阅读
航空燃油泵驱动电机负责给燃油泵提供所需的动力。由于燃油驱动电机与燃油泵之间存在很强的非线性及耦合性,使得燃油泵无法正常运行在工况点,造成严重的能量损耗。针对该问题,提出了利用Legendre神经网络法来解决燃油泵与驱动电机转速匹配问题。首先,利用Legendre神经网络法拟合出燃油泵的特性曲线,然后再利用该方法建立以燃油泵的流量、扬程为输入,驱动电机转速为输出的神经网络模型。燃油泵与驱动电机转速匹配仿真结果表明,相比于传统的最小二乘法,Legendre神经网络法具有更高的精度。
The aviation fuel pump drive motor is responsible for providing the required power to the fuel pump. Due to the strong non-linearity and coupling between the fuel-driven motor and the fuel pump, the fuel pump can not operate normally at the operating point, resulting in serious energy loss. To solve this problem, this paper proposes to use Legendre neural network method to solve the problem of speed matching between fuel pump and drive motor. Firstly, the characteristic curve of the fuel pump is fitted by using the Legendre neural network method, and then the neural network model with the flow of the fuel pump and the lift as the input and the output of the drive motor is output. Fuel Pump and Drive Motor Speed Matching Simulation results show that the Legendre neural network method has higher accuracy than the traditional least squares method.