决策树算法在获取连续流反应器动力学参数中的应用  

Acquisition of Reaction Kinetic Data Using Decision Tree Algorithm in Designing Continuous Reactor

作　　者：蒋雄 Jiang Xiong(Jiujiang Tianci Hi-Tech Materials Co.,Ltd.,Jiujiang Jiangxi 332599,China)

机构地区：[1]九江天赐高新材料股份有限公司,江西九江332599

出　　处：《江西化工》2024年第1期109-113,共5页Jiangxi Chemical Industry

摘　　要：为帮助企业工艺人员在设计连续反应器前获取合理的动力学数据并快速进行动力学测试方案设计,研究者选择动力学测试方法和对应测试试验仪器,设计了一套动力学分析测试系统。用户可通过输入反应方程,勾选预设14个具体影响反应动力学测试的问题,要求系统一方面完成反应物网络及主反应可能的能级图的绘制,另一方面通过决策树机器学习算法处理预设的99万条数据,从中找到最佳的动力学测试建议并呈现。整个机器学习过程不超过4s,整个分析测试系统运行时间不超过1min,大大提高了研发人员进行动力学测试的效率。In order to help the technical personnel of the enterprise to obtain reasonable kinetic data and quickly design the kinetic test scheme before designing the continuous reactor,select the kinetic test method and the corresponding test instrument,the researchers designed a set of kinetic analysis test system,By inputting the reaction equation,the user selects 14 preset problems that specifically affect the reaction kinetics test.The analysis and suggestion system completes the drawing of the reactant network and the possible energy level diagram of the main reaction on the one hand,and on the other hand,through the decision tree machine learning algorithm,process the preset 990,000 pieces of data,find the best kinetic test suggestion and present it.The entire machine learning process does not exceed 4s,and the running time of the entire analysis and testing system does not exceed 1 minute,which improves the efficiency of dynamic testing for R&D personnel.

关 键 词：决策树算法 反应动力学 连续反应器 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] O643.1[自动化与计算机技术—控制科学与工程]