论文部分内容阅读
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。
In order to solve the problem of on-orbit heat flow identification and real-time temperature prediction, a sequential Monte Carlo SMC (Sequential Monte Carlo) algorithm is proposed to identify the changes of orbital heat flow in real time. Combining with the two-layer lumped parameter model, temperature change. The SMC algorithm uses the posterior probability density function (PPDF) method of structured temperature to filter and predict, so that the on-line identification of external heat flow and temperature prediction performance are greatly improved. The simulation results show that the sequential Monte Carlo filtering method can realize the dynamic characteristic identification of satellite thermal system, guarantee the reliability of adaptive identification of low-dimensional model parameters, and realize the on-line prediction technology of temperature variation of controlled objects.