电化学传感器阵列结合机器学习对液相糖成分识别系统的设计  

Sugar Component Recognition System for Liquid Phase Based on the Combination of Electrochemical Sensor Array and Machine Learning

作　　者：张靖渝 连俊博 雍程翔 赖永志 李响陈 朱润昊 彭锐 赵睿智 余天玺 钟蕾 陈张昊 易晓梅[1,3] ZHANG Jingyu;LIAN Junbo;YONG Chengxiang;LAI Yongzhi;LI Xiangchen;ZHU Runhao;PENG Rui;ZHAO Ruizhi;YU Tianxi;ZHONG Lei;CHEN Zhanghao;YI Xiaomei(Key Laboratory of Forestry Sensing Technology and Intelligent Equipment of National Forestry and Grassland Administration,Forestry,Key Laboratory of Forestry Intelligent Monitoring and Information Technology of Zhejiang Province,School of Mathematics and Computer Science,Zhejiang A&F University,Hangzhou Zhejiang 311300,China;School of Humannities and Law,Zhejiang A&F University,Hangzhou Zhejiang 311300,China;School of Computer Science and Technology,Zhejiang Sci-Tech University,Hangzhou Zhejiang 310052,China)

机构地区：[1]浙江农林大学数学与计算机科学学院林业感知技术与智能装备国家林业局重点实验室,浙江省林业智能监测重点实验室,浙江杭州311300 [2]浙江农林大学人文与法律学院,浙江杭州311300 [3]浙江理工大学计算机科学与技术学院,浙江杭州310052

出　　处：《传感技术学报》2025年第3期555-561,共7页Chinese Journal of Sensors and Actuators

基　　金：浙江省新苗项目(2024R412A025);国家级大学创新项目(202410341077X)。

摘　　要：为了在生产的过程中可以精准检测到食品中的糖成分,设计了一种三电极电化学检测体系。对电极为铂电极,参比电极为饱和KCl甘汞溶液电极,工作电极分别为铜膜、泡沫铜和铜片。采用循环伏安法(CV)和计时电流法(i-t)检测含有不同浓度赤藓糖醇、阿洛酮糖、三氯蔗糖的混合溶液电化学响应。使用Savitzky-Golay(S-G)滤波对数据进行平滑除噪。通过客观权重算法(熵权法、CRTIC)和机器学习算法(极端梯度提升(XGBoost)、极端随机树(ExtraTrees))筛选特征值;根据数据特征建立Alexnet模型,使用Adam优化器优化模型。i-t能较好地判别溶液中的糖含量。基于S-G滤波和权重的优化Alexnet模型的预测准确率(90.9%)高于其他数据处理模型。所提方法具有响应速度快,准确率高的优点。In order to accurately detect the sugar composition of the food during the production process,a three-electrode electrochemical detection system is designed.Platinum electrode is used as counter electrode,saturated KCl calomel solution electrode is used as refer-ence electrode,and copper film,copper foam and copper sheet are used as working electrodes.Cyclic voltammetry(CV)and chrono-amperometry(i-t)are used to detect the electrochemical responses of mixed solutions containing erythritol,psicose and sucralose with different concentrations.The Savitzky-Golay(S-G)filter is used to smooth and remove noise from the data.Objective weight algorithm(entropy weight method,CRTIC)and machine learning algorithm(eXtreme gradient boosting(XGBoost),extremely randomized trees(ExtraTrees))are used to screen feature values.Alexnet model is built according to data characteristics,and the Adam optimizer is used to optimize the model.i-t can well distinguish the sugar composition in the solution.The optimized Alexnet model based on S-G filter and weights has a higher prediction accuracy(90.9%)than other data processing models.The proposed method has the advantages of fast response speed and high accuracy.

关 键 词：电化学 糖混合溶液 定量检测 多工作电极 机器学习算法 

分 类 号：TP212.6[自动化与计算机技术—检测技术与自动化装置]