基于QES-SVM的雷达装备剩余寿命预测

QES-SVM-based prediction of remaining life of radar equipment

针对雷达装备剩余寿命的单一模型预测精度不高的问题,使用二次指数平滑(QES)法和支持向量机(SVM)组合模型进行剩余寿命预测.首先以均方根作为指标参数,输入两种单一预测模型,分别得出各自剩余寿命计算结果;然后结合误差平方和最小法求出权重,计算组合模型预测结果,并与两种单一模型结果进行对比.仿真结果表明,QES-SVM组合模型预测精度与两种单一模型相比,误差更小,预测效果更佳,验证了该组合模型的有效性.

In order to address the problem that the accuracy of single model predicting radar equipment remaining life is not high, this paper uses the combined model of quadratic exponential smoothing(QES) and support vector machine(SVM) to make residual life prediction. First, root-mean-square is taken as the index parameter, and two single prediction models are input to obtain the calculation results of their remaining life respectively.Then, the weight is solved by combining the error-squared sum-of-squares minimization method, and the prediction results of the combined model are calculated and compared with the results of two single models. The simulation results show that the prediction accuracy of the QES-SVM combined model is smaller than that of the two single models, and that the prediction effect is better, which verifies the effectiveness of the combined model.