论文部分内容阅读
直升机旋翼系统的工作方式及其所承受的载荷形式使飞行实测载荷数据的有效性不高。研究如何利用有限的实测载荷及飞行参数数据建立直升机旋翼系统飞行参数识别模型,对于推进飞行载荷测试任务有重要意义。基于Matlab编程建立遗传算法优化的BP神经网络直升机旋翼系统飞行参数识别模型,实现通过现有载荷数据及飞参数据对旋翼系统飞行载荷预测仿真。预测的最大相对误差为10%、平均相对误差为3.7%,满足工程要求,并且较未使用遗传算法优化的BP神经网络预测结果好,表明所建立的飞行参数识别模型具有很好的学习能力和泛化能力。
The working mode of helicopter rotor system and its load form make the measured load data of flight less effective. Studying how to use the limited measured load and flight parameter data to establish the helicopter rotor flight system flight parameter identification model is of great significance to the advance flight load test. Based on Matlab programming, a genetic algorithm optimized BP neural network helicopter rotor flight system flight parameters identification model is established to realize the flight load forecasting simulation of the rotor system through the existing load data and flight parameters. The predicted maximum relative error is 10% and the average relative error is 3.7%, which meets the engineering requirements and is better than the BP neural network without genetic algorithm optimization. It shows that the established flight parameter identification model has good learning ability and Generalization.