以卷积神经网络数据采集方法拓宽真空教学实验  

Data collection by convolutional neural network to extend the vacuum teaching experiment

作　　者：彭博 杜尚耕 周云帆 郑远 王业伍[2] PENG Bo;DU Shanggeng;ZHOU Yunfan;ZHENG Yuan;WANG Yewu(Chu Kochen Honors CollegeZhejiang University,Hangzhou 310058,China;School of Physics,Zhejiang University,Hangzhou 310058,China)

机构地区：[1]浙江大学竺可桢学院,浙江杭州310058 [2]浙江大学物理学院,浙江杭州310058

出　　处：《物理实验》2023年第3期13-21,共9页Physics Experimentation

基　　金：浙江省高校实验室工作研究项目(No.YB202135);浙江大学物理学院教学研究项目(No.ZJUPHYS2022ZD)。

摘　　要：设计制作了基于玻璃扩散泵的两级泵真空实验装置,该实验装置具有良好的透视性,适合本科初级阶段的实验教学;实验中融合计算机识别算法,采用卷积神经网络高效采集数据.开展了真空的获得与测量实验教学方法的探索,包括:低真空的抽气与漏气实验,测量抽速与漏率;高真空的获得与测量,观测扩散泵运转过程;腔体真空度与位置之间的关系实验,探究流导对真空的影响.较为全面地将真空基本概念融入实验,并与自动化采集实验相结合,提高了实验效率.A two-stage pump vacuum experimental device based on glass diffusion pump was designed and made.The experimental device had good perspective and was suitable for experimental teaching in the primary stage of undergraduate.In the experiment,computer recognition algorithm was integrated and convolutional neural network was used to collect data efficiently.The teaching methods of vacuum acquisition and measurement experiment were explored including low vacuum extraction and leakage experiment to measure the extraction speed and leakage rate;obtaining and measuring high vacuum to observe the operation process of diffusion pump;the experiment on the relationship between cavity vacuum degree and position to explore the influence of flow conductance on vacuum.The basic concept of vacuum was integrated into the experiment comprehensively and combined with the automatic collection experiment,which improved the experiment efficiency.

关 键 词：真空实验 扩散泵 卷积神经网络 

分 类 号：O4-39[理学—物理] TB751[一般工业技术—真空技术]