AGA在太阳能电池阵寿命预测中的应用

针对在轨卫星的GaInP/GaAs/Ge三结太阳能电池阵输出电流衰减模型中参数辨识精度低和后期误差大且实时性不高的问题，提出采用遗传算法(GA)对太阳能电池阵电流衰减模型进行参数辨识。针对GA过早收敛及易陷入局部最优解的缺点，提出了基于自适应交叉变异算子和线性变换适应度函数的自适应遗传算法(AGA)进行参数辨识。然后把模型预测的输出电流与遥感测得的输出电流进行比较。实验结果表明，运用AGA和广义最小二乘法（GLS）得到的均方根误差(RMSE)分别为1.427×10^-4和1.337×10^-3，误差平方和(SSE)分别为2.0571×10^-6和1.92×10^-3，AGA辨识后的模型预测更加准确，对卫星运行后期电源的管理具有实际意义。

Application of Adaptive Genetic Algorithm in Life Prediction of Solar Cell

In order to solve the problems of low accuracy of parameter identification and large late error and low real time performance of GaInP/GaAs/Ge triple junction solar cell array in orbit satellite,a genetic algorithm (GA) is proposed to identify the parameters of the solar cell array current attenuation model.In order to overcome the shortcoming that GA prematurely converges and falls into local optimal solution,a adaptive genetic algorithm (AGA) based on adaptive crossover and mutation operator and linear transformation fitness function was proposed for parameter identification.Then,the predicted output current was compared to the sensed output current to verify the accuracy of the predicted output current.Experimental results show that the root mean square error (RMSE) of AGA and generalized least squares (GLS) is respectively 1.427×10^-4 and 1.337×10^-3,sum of error squares (SSE) is respectively 2.0571×10^-6 and 1.92×10^-3.The model prediction after AGA identification is more accurate.It is of practical significance to manage the power supply in the later stage of satellite operation.