基于IGWO组合模型的炮控系统健康预测方法  

Health Prediction Method of Gun Control System Based on an IGWO Combined Model

作　　者：胡修宇 李英顺 王德彪 刘海洋 HU Xiuyu;LI Yingshun;WANG Debiao;LIU Haiyang(School of Automation,Guangxi University of Science and Technology,Liuzhou 545000,Guangxi,China;School of Control Science and Engineering,Dalian University of Technology,Dalian 116200,Liaoning,China;Shenyang Shunyi Technology Company Limited,Shenyang 110000,Liaoning,China)

机构地区：[1]广西科技大学自动化学院,广西柳州545000 [2]大连理工大学控制科学与工程学院,辽宁大连116200 [3]沈阳顺义科技有限公司,辽宁沈阳110000

出　　处：《火炮发射与控制学报》2023年第3期28-34,共7页Journal of Gun Launch & Control

基　　金：辽宁省自然科学基金(XLYC1903015)。

摘　　要：针对单一模型对炮控系统组成部件的健康预测精度低、适用性差等问题,提出一种改进灰狼优化算法(IGWO)的组合预测方法。通过改进灰狼优化算法(GWO)的收敛因子,提高全局搜索能力,避免陷入局部最优值;选择在健康预测中效果较好的ARMA、RBFNN、LSSVM模型,搜索3种模型的最优权重系数并建立组合预测模型;以陀螺仪组为例,将其温度传感器阻值与漂移度数据分别带入单一模型、IGWO组合模型和其他组合模型中进行健康预测。通过对预测结果的误差评价分析与RUL误差比较,可以得出,IGWO在搜索组合模型最优权重系数方面优于GWO,预测精度也优于其他组合模型,且预测效果稳定。Aiming at the problems of low health prediction accuracy and poor applicability to the components of the gun control system by a single model,a combined prediction method based on an improved Grey Wolf Optimization Algorithm(IGWO)is proposed.By improving the convergence factor of the gray wolf optimization algorithm(GWO),the global search ability is improved and the local optimum value is avoided;ARMA,RBFNN,and LSSVM models with better effects in health prediction are selected,and the optimal weight coefficients of the three models are searched before a model is established.Taking the gyroscope assembly as an example,the resistance and drift data of its temperature sensor are brought into a single model,an IGWO combined model,and other combined models respectively for health prediction.Through an error evaluation analysis of the prediction results and the comparison with the RUL error,it can be concluded that the IGWO is better than the GWO in sear-ching for the optimal weight coefficient of the combined model,so is its prediction accuracy.The predicted result is also stable.

关 键 词：健康预测 改进灰狼优化算法 组合模型 最优权重 

分 类 号：TJ811.2[兵器科学与技术—武器系统与运用工程]