谱共轭梯度算法反演航天器在轨瞬态外热流

Retrieval of transient external heat flux of spacecraft in orbit based on spectral conjugate gradient algorithm

航天器在轨运行期间,外表面热流对其热保护系统影响很大。为了防止航天器因外部热流的影响导致热控制系统故障,要对航天器外热流进行估计。但对热流直接测量是不易的,因此通过反演方法对航天器外热流进行辨识。首先,利用航天器外表面温度的测量值,建立了反演航天器在轨瞬态外热流的数学模型;其次,为了克服不适定性和加快算法收敛速度提出了基于偏差原则的谱共轭梯度算法,偏差原则可以消除反问题的不适定性,谱共轭梯度算法可以加快算法的收敛速度;最后,通过数值仿真试验对基于偏差原则的谱共轭梯度法的反演效果进行了检验。当测量值中加入2%的测量误差后,反演值与真实值的最大相对偏差为3. 68%,反演方法克服了反问题的不适定性。与共轭梯度算法相比,谱共轭梯度算法明显减少迭代次数。

During the orbital operation of spacecraft,the heat flux on the outer surface demonstrates a great impact on the thermal protection system of spacecraft. Owing to the influence of external heat flux,the fault of thermal control system of spacecraft can occur frequently. However,it is not easy to measure the heat flux directly. Hence,this paper applies the inverse method to identify the heat flux outside the spacecraft. First,the mathematical model of the transient heat flux of spacecraft is modelled,on the basis of the measured values of the outer surface temperature. Second,to overcome the ill-posedness of inverse problem and accelerate convergence rate,a spectral conjugate gradient algorithm with the deviation principle is proposed. The deviation principle can eliminate the ill-posedness of the inverse problem,and the spectral conjugate gradient algorithm can accelerate the convergence rate. Finally,the performance of the proposed algorithm is verified by the numerical simulation. When we add the measurement error level by 2%,the maximum relative deviation between the inversion value and the real value is 3. 68%. The inversion method overcomes the ill-posedness of the inversion problem. Compared with the conjugate gradient algorithm,the proposed spectral conjugate gradient algorithm can reduce the number of iterations significantly.