电子元件与材料2025，Vol.44Issue(9)：1073-1078,6.DOI:10.14106/j.cnki.1001-2028.2025.0252
基于多传感器数据融合技术的电子元件温度检测研究
Research on temperature detection of electronic components based on multi-sensor data fusion technology
摘要
Abstract
Due to the significant increase in local heat flux density caused by the high integration and miniaturization of electronic components,traditional single-sensor temperature detection methods suffer from measurement baseline drift induced by sensor nonlinear responses and environmental noise.These issues,combined with thermal radiation coupling and multi-physics field interference,lead to dynamic detection errors.Additionally,a single sensor modality is unable to cover a wide temperature range or capture transient thermal characteristics,resulting in significant temperature detection errors.This study proposed a temperature detection method for electronic components based on multi-sensor data fusion technology.Multiple types of temperature sensors were used to acquire multi-source sensor data,which was applied to compensate for baseline drift errors.By integrating the Backpropagation(BP)neural network with the Genetic Algorithm(GA),a multi-sensor data fusion model(GA-BP)was constructed.The GA-BP model utilized the Genetic Algorithm to optimize the weights and thresholds of the BP neural network;it performed subsequent collaborative processing via the input layer,hidden layer,and output layer to ultimately yield the temperature detection results of electronic components.Experimental results showed that after error compensation,the standard deviation between sensor data and actual values was significantly reduced to near the theoretical limit,with the detection error ranging from 0.2％to 0.8％.During the transient thermal load phase,this method controls the dynamic detection error within 0.3％through dynamic weight optimization and thermal radiation compensation,thus verifying its robustness across wide temperature ranges and complex operating conditions.关键词
数据融合技术/电子元件/误差补偿/权重与阈值/多传感器/温度检测Key words
data fusion technology/electronic components/error compensation/weight and threshold/multi-sensor/temperature detection分类
计算机与自动化引用本文复制引用
段新娥..基于多传感器数据融合技术的电子元件温度检测研究[J].电子元件与材料,2025,44(9):1073-1078,6.基金项目
江苏省高等学校基础科学(自然科学)研究重大项目(22KJA520003) （自然科学）
