基于PSO-BP神经网络的雷达电源组件温度预测新方法  

A new temperature prediction method of radar power supply module based on PSO-BP neural network

作　　者：唐敏[1] 黄大荣[1] 唐环 赵宁 张宇 TANG Min;HUANG Da-rong;TANG Huan;ZHAO Ning;ZHANG Yu(Chongqing Jiaotong University,Chongqing 400074;Beijing Kewei Zhuoli Technology Co.,Ltd.,Beijing 100191;The Eighth Research Academy of CSSC,Nanjing 211153)

机构地区：[1]重庆交通大学,重庆400074 [2]北京可维卓立科技有限公司,北京100191 [3]中国船舶集团有限公司第八研究院,南京211153

出　　处：《雷达与对抗》2023年第1期54-58,68,共6页Radar & ECM

基　　金：国防科工局技术基础科研“十三五”第五批项目(JSZL2019207B008)。

摘　　要：针对雷达电源组件在长期运行中元器件温度过高,导致可靠性能退化,进而影响雷达其他组件及整体系统可靠性的问题,提出基于PSO-BP神经网络的雷达电源组件温度预测新方法。首先,根据雷达电源组件的物理特性详细分析其流动-传热;然后,以调整其大范围误差为目标,基于PSO-BP理论构建雷达电源组件温度预测模型,并设计相应的算法流程图;最后,以某型装备相控阵雷达的电源组件温度预测为工程背景,对所提的算法进行实验验证,实验结果满足工程要求。To overcome the problem that the temperature of the radar power module is too high in long-term operation,which leads to the degradation of reliability and affects the reliability of other radar components and the overall system,a new method of predicting the temperature of the radar power supply module is proposed based on the PSO-BP neural network.First,the flow-heat transfer of the radar power supply module is analyzed in detail according to its physical characteristics.On this basis,the temperature prediction model of the radar power supply module is built based on the PSO-BP theory to adjust the large range error,and the corresponding algorithm flow chart is designed.Finally,taking the temperature prediction of the power supply module of a phased array radar as the engineering background,the proposed algorithm is verified experimentally,and the results meet the engineering requirements.

关 键 词：雷达电源组件 温度预测 PSO-BP神经网络 温度场分析 

分 类 号：TN95[电子电信—信号与信息处理]