【摘要】 为了实现纳卫星轨道热环境的辨识,在纳卫星热系统动态特性模型以及BP神经网络的基础上,建立纳卫星轨道热环境辨识模型。首先,利用周期性空间外热流输入求解纳卫星热系统动态特性模型,然后,选取数据样本完成BP神经网络的训练。最后,将训练后的BP神经网络应用于实际的纳卫星飞行模式中的轨道热环境辨识,进行验证。通过辨识结果与实际轨道热环境的对比,证明了纳卫星轨道热环境辨识模型的泛化能力及准确性。更多还原

【Abstract】 In order to identify the thermal environment of Nano satellite on orbit,an identification system of thermal environment was established based on the thermal dynamic model of Nano satellite and BP neural networks.A periodic thermal flow as the input was used,and the thermal dynamic model of Nano satellite was solved.The data sample was chosen to finish the training of BP neural networks.After training,the identification system was used in an actual flight of Nano satellite.Compared with the identification result and actual thermal environment data,the identification system of thermal environment is proved accurately and can be used widely.更多还原