序列蒙特卡洛滤波在卫星传热反问题中的应用

为了解决卫星在轨外热流辨识与实时温度预测问题,提出采用序列蒙特卡洛SMC(Sequential Monte Carlo)算法实时辨识卫星轨道外热流变化,结合双层集总参数模型,快速准确地预测星载仪器温度变化.SMC算法采用构造温度的后验概率密度函数PPDF的方法进行滤波和预测,使得外热流的在线辨识和温度预测性能得到较大提高.仿真结果表明序列蒙特卡洛滤波方法可以实现卫星热系统动态特性辨识,保证了低维模型参数的自适应辨识的可靠性,实现受控对象温度变化的在线预测技术.

Sequential Monte Carlo Approach to Inverse Problems in Heat Transfer of Satellite

The inverse problem in heat transfer of satellite is to reconstruct the orbit out heat flux depending on the given temperature observations in this paper. To identify dynamically the out heat flux and predict the temperature of satellite more easily and correctly,a Sequential Monte Carlo(SMC) approach was proposed,together with a double layer lumped parameter model. In the SMC approach,one attempts to construct the Posterior Probability Density Function(PPDF) of the state,this improves the perfoemace of out heat flux identification and temperature prediction. A simulation example of inverse problems in heat transfer of Satellite was presented. Simulation results demonstrate the performance of dynamical identification and precision of estimation were improved with the SMC,and the great potential of applying the SMC to identification problem of satellite.