翅片管换热器仿真软件精度自适应校正方法  

Accuracy Self-Adaptive Method of Fin and Tube Heat Exchanger Simulation Software

作　　者：芮银波[1] 丁国良[1] 吴志刚[1] 

机构地区：[1]上海交通大学制冷与低温工程研究所,上海200030

出　　处：《上海交通大学学报》2006年第8期1360-1364,共5页Journal of Shanghai Jiaotong University

基　　金：国家重点基础研究发展规划(973)项目(G2000026309)

摘　　要：提出了一种自动修正翅片管换热器仿真精度的方法.根据用户所能提供实验数据的数量,采用不同的策略,实验数据较少时采用简单的多项式回归的形式来建立管外空气侧换热系数修正因子的修正多项式;实验数据较多时采用人工神经网络的方式来得到管外空气侧换热系数修正因子的神经网络关联模型.用于实际的仿真计算时,修正前的换热量误差最大28.96%,平均为11.12%;采用多项式修正后误差最大15.33%,平均5.38%;采用神经网络修正后误差最大4.9%,平均1.0%.A method to correct the simulation accuracy of the fin-and-tube heat exchanger model was presented, which can choose correction strategies automatically according to the number of the experiment data points the users can offer. When the number of data points is small, the air side heat transfer coefficient is corrected with polynomial regress. When the number of data points is large, the air side heat transfer coefficient is corrected with artificial neural network. The practical simulation with this method shows：① the max and mean deviation of heat flow rate are 28. 96% and 11. 12% respectively, when no correction is made; ② the max and mean deviation of heat flow rate are 15.33 % and 5.38% respectively, with polynomial correction method; ③ the max and mean deviation of heat flow rate are 4.9% and 1.0%respectively, when artilicial neural network correction is used.

关 键 词：换热器 仿真 人工神经网络 

分 类 号：TB65[一般工业技术—制冷工程]