基于遗传编程的吸附床传热状态预测  

Heat Transfer Prediction of Absorption Tower Based on Genetic Programmir

作　　者：张元英 冯黛丽 冯妍卉[1,2] 张欣欣 ZHANG Yuanying;FENG Daili;FENG Yanhui;ZHANG Xinxin(School of energy and environmental engineering,University of Science and Technology Beijing,Beijing 100083,China;Beijing Key Laboratory of Energy Saving and Emission Reduction for Metallurgical Industry,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]北京科技大学能源与环境工程学院,北京100083 [2]北京科技大学冶金工业节能减排北京市重点实验室,北京100083

出　　处：《工程热物理学报》2023年第6期1720-1725,共6页Journal of Engineering Thermophysics

基　　金：国家自然科学基金(No.52176054,No.52236006,No.51876007);中国外国专家项目(No.G2022105009L)。

摘　　要：脱硝吸附塔是烟气处理系统中的重要环节。塔内的热交换受到多方因素的非线性影响,给准确快速的评估吸附剂换热状态带来巨大挑战。为解决此问题,使用数值模拟与均一介质方法,对一种插管式吸附塔换热进行了模拟与多元遗传编程回归,获得了白盒表达式并验证模型准确性。分析了因变量、遗传编程参数等设定对回归的影响。最终结果表明使用遗传编程的回归效果优异,在工业生产与探究变量物理关系中,具有强大的应用潜力。The adsorption tower of denitration is an important step in the flue treatment system.The heat exchange in the tower is affected by nonlinear factors,which brings great challenges to accurately and rapidly evaluating the heat transfer of adsorbents.To solve this change,the numerical simulation and homogeneous medium were adopted to simulate the heat transfer of a cannulated adsorption tower.Based on the data,regression of multi-genetic programming was carried out.The white-box expression was obtained and the accuracy of the model was verified.The infuence of the dependent variable and the parameter of regression was analyzed.The final results show that the regression of genetic programming is excellent,and has strong application potential in industrial production and exploring the physical relationship between variables.

关 键 词：吸附床换热 遗传编程 误差分析 机器学习 符号回归 

分 类 号：TK121[动力工程及工程热物理—工程热物理]