Turbulent flame image classification using Convolutional Neural Networks

作     者：Rathziel Roncancio Aly El Gamal Jay P.Gore 

作者机构：School of Mechanical EngineeringPurdue UniversityWest LafayetteIN47907USA School of Electrical and Computer EngineeringPurdue UniversityWest LafayetteIN 47907USA 

出 版 物：《Energy and AI》 (能源与人工智能（英文）)

年 卷 期：2022年第10卷第4期

页      面：87-94页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：The authors of Ref. made the experimental data available for this study 

主　　题：CNN Flame Neural network Turbulent 

摘      要：Pockets of unburned material in turbulent premixed flames burning CHs,air,and CO_(2) were studied using OH Planar Laser-Induced Fuorescence(PLIF)images to improve current *** flames are ubiquitous in most natural gas air combustors running gas turbines with dry exhaust gas recirculation(EGR)for land-based power *** improvements continue in the charactenization and understanding of turbulent flames with EGR particularly for transient events like ignition and *** and/or islands of unburned material within bumed and unburned turbulent media are some of the features of these *** features reduce the heat release rates and increase the carbon monoxide and hydrocarbons *** present work involves Convolutional Neural Networks(CNN)based dassification of PIF images containing unburned pockets in three turbulent flames with 0%,5%,and 10%CO_(2).The CNN model was constructed using three convolutional layers and two fully connected layers using dropout and weight *** of 94.2%,92.3%and 89.2%were registered for the three flames,*** present approach represents significant computational time savings with respect to conventional image processing methods.