计量学报2025，Vol.46Issue(4)：565-573,9.DOI:10.3969/j.issn.1000-1158.2025.04.14
基于DA-SVM-SHAP杂化策略的直接接触式蒸发器换热系数可解释性预测
Interpretable Prediction of Heat Transfer Coefficients for Direct Contact Evaporator Based on DA-SVM-SHAP Hybridization Strategy
摘要
Abstract
Aiming at the difficulty in describing the complex nonlinear relationship of the heat transfer coefficient of the direct contact evaporator in the low-temperature metallurgical waste heat recovery system,a dragonfly optimization algorithm(DA)is proposed to optimize the support vector machine(SVM)algorithm to achieve intelligent prediction of the heat transfer coefficient.The results show that the DA-SVM algorithm can take the inlet temperature of the heat transfer oil,the outlet temperature of the heat transfer oil,the outlet temperature of the organic working fluid,the steam flow rate,and the heat absorbed by the organic working fluid as input variables,and the heat transfer coefficient of the direct contact evaporator as output variables.Compared with four prediction models including BP,ELM,KELM,and XGBoost,the RMSE(root mean square error)of the DA-SVM model decreased by 10.5％and the R2 increased by 2.6％.By successfully explaining the contribution of input variables to prediction performance using SHAP values,the DA-SVM-SHAP model can achieve high-precision prediction of heat transfer coefficients and visualize the correlation between each input variable and the predicted heat transfer coefficient results.It has shown significant advantages in modeling nonlinear relationships in the heat transfer process and can provide a basis for further optimizing the operating parameters of low-temperature waste heat recovery systems and improving the heat transfer performance of evaporators.关键词
热学计量/直接接触式蒸发器/换热系数/蜻蜓优化算法/支持向量机Key words
thermal metrology/direct contact evaporator/heat transfer coefficient/DA/SVM引用本文复制引用
谭银珍,杨凯,张晓雪,姚钦文,王华,肖清泰..基于DA-SVM-SHAP杂化策略的直接接触式蒸发器换热系数可解释性预测[J].计量学报,2025,46(4):565-573,9.基金项目
云南省重大科技专项-西联研院专项(202302AQ370001) （202302AQ370001）
中央引导地方科技发展资金(202407AB110022) （202407AB110022）
云南省科技人才和平台计划(202405AF140068) （202405AF140068）
