Acquiring accurate molecular-level information about petroleum is crucial for refining and chemical enterprises to implement the“selection of the optimal processing route”strategy.With the development of data predic...Acquiring accurate molecular-level information about petroleum is crucial for refining and chemical enterprises to implement the“selection of the optimal processing route”strategy.With the development of data prediction systems represented by machine learning,it has become possible for real-time prediction systems of petroleum fraction molecular information to replace analyses such as gas chromatography and mass spectrometry.However,the biggest difficulty lies in acquiring the data required for training the neural network.To address these issues,this work proposes an innovative method that utilizes the Aspen HYSYS and full two-dimensional gas chromatography-time-of-flight mass spectrometry to establish a comprehensive training database.Subsequently,a deep neural network prediction model is developed for heavy distillate oil to predict its composition in terms of molecular structure.After training,the model accurately predicts the molecular composition of catalytically cracked raw oil in a refinery.The validation and test sets exhibit R2 values of 0.99769 and 0.99807,respectively,and the average relative error of molecular composition prediction for raw materials of the catalytic cracking unit is less than 7%.Finally,the SHAP(SHapley Additive ExPlanation)interpretation method is used to disclose the relationship among different variables by performing global and local weight comparisons and correlation analyses.展开更多
The coupled characteristics of the cooling powers and temperatures of the two cold heads of the two-stage pulse tube refrigerators(PTRs)increase the difficulties of modeling and application.This study proposes an arti...The coupled characteristics of the cooling powers and temperatures of the two cold heads of the two-stage pulse tube refrigerators(PTRs)increase the difficulties of modeling and application.This study proposes an artificial neural network(ANN)for predicting the power performance of a two-stage PTR.The experiment results show that the temperatures and cooling powers of the cold heads exhibit strong nonlinear characteristics so as to bring extreme difficulty for the performance modeling analytically.The ANN includes four-layer to predict the two cooling powers of the cold heads while the two temperatures of cold heads are the inputs.After the hyperparameter tuning,the number of the neural of two hidden layers is 8 and 6,respectively.Compared with the modelling methods of Decision Trees,Random Forests,multivariate polynomial fitting and Support Vector Regression,the ANN model can achieve the highest coefficient of determination R2 of 0.998 and the maximum relative errors of 1.6%as well as less than 5 milliseconds prediction time.Through SHapley Additive exPlanations(SHAP)analysis,the contributions of the 1st and 2nd temperatures to the cooling powers are quantified.By removing low-contribution data points identified through SHAP analysis selectively,the training dataset reduces by 34%and the maximum relative error of the model is still no more than 5%.It provides powerful guidance for selecting the appropriate training dataset and powerful tools for coupling with the other refrigerator models and systems.展开更多
The production of green hydrogen via alkaline water electrolysis necessitates porous composite membranes with high ionic conductivity and high bubble-point pressure.However,the mainstream preparation process of porous...The production of green hydrogen via alkaline water electrolysis necessitates porous composite membranes with high ionic conductivity and high bubble-point pressure.However,the mainstream preparation process of porous composite membranes involves many parameters,rendering this a complex high-dimensional optimization problem.Traditional trial-and-error experimentation is inefficient and often fails to explore the performance boundaries.In this study,an XGBoost-based machine learning model is developed and trained with laboratorycollected datasets,achieving satisfactory predictive performance.The model provides critical insights into the relationships between six manufacturing parameters and the two core performance parameters of the membrane.Subsequently,prediction based on coarse-grained grid and reverse search are performed on the model to identify optimal parameter regions,followed by manual refinement through feature analysis.This integrated approach ultimately identifies three high-performance composite membrane candidates,which are experimentally vali-dated.This work demonstrates a highly efficient and accurate machine learning-driven paradigm for the development of advanced porous composite membrane in alkaline water electrolysis.展开更多
基金the National Natural Science Foundation of China(22108307)the Natural Science Foundation of Shandong Province(ZR2020KB006)the Outstanding Youth Fund of Shandong Provincial Natural Science Foundation(ZR2020YQ17).
文摘Acquiring accurate molecular-level information about petroleum is crucial for refining and chemical enterprises to implement the“selection of the optimal processing route”strategy.With the development of data prediction systems represented by machine learning,it has become possible for real-time prediction systems of petroleum fraction molecular information to replace analyses such as gas chromatography and mass spectrometry.However,the biggest difficulty lies in acquiring the data required for training the neural network.To address these issues,this work proposes an innovative method that utilizes the Aspen HYSYS and full two-dimensional gas chromatography-time-of-flight mass spectrometry to establish a comprehensive training database.Subsequently,a deep neural network prediction model is developed for heavy distillate oil to predict its composition in terms of molecular structure.After training,the model accurately predicts the molecular composition of catalytically cracked raw oil in a refinery.The validation and test sets exhibit R2 values of 0.99769 and 0.99807,respectively,and the average relative error of molecular composition prediction for raw materials of the catalytic cracking unit is less than 7%.Finally,the SHAP(SHapley Additive ExPlanation)interpretation method is used to disclose the relationship among different variables by performing global and local weight comparisons and correlation analyses.
基金supported by the Taishan Scholar Project(Grand No.tsqn202103142).
文摘The coupled characteristics of the cooling powers and temperatures of the two cold heads of the two-stage pulse tube refrigerators(PTRs)increase the difficulties of modeling and application.This study proposes an artificial neural network(ANN)for predicting the power performance of a two-stage PTR.The experiment results show that the temperatures and cooling powers of the cold heads exhibit strong nonlinear characteristics so as to bring extreme difficulty for the performance modeling analytically.The ANN includes four-layer to predict the two cooling powers of the cold heads while the two temperatures of cold heads are the inputs.After the hyperparameter tuning,the number of the neural of two hidden layers is 8 and 6,respectively.Compared with the modelling methods of Decision Trees,Random Forests,multivariate polynomial fitting and Support Vector Regression,the ANN model can achieve the highest coefficient of determination R2 of 0.998 and the maximum relative errors of 1.6%as well as less than 5 milliseconds prediction time.Through SHapley Additive exPlanations(SHAP)analysis,the contributions of the 1st and 2nd temperatures to the cooling powers are quantified.By removing low-contribution data points identified through SHAP analysis selectively,the training dataset reduces by 34%and the maximum relative error of the model is still no more than 5%.It provides powerful guidance for selecting the appropriate training dataset and powerful tools for coupling with the other refrigerator models and systems.
基金supported by the National Key Research and Develop-ment Program(2022YFB4202200)National Natural Science Founda-tion of China(22479114)+2 种基金Science and Technology Innovation Action Plan of Shanghai(No.25DZ3001302)the Fundamental Research Funds for the Central Universitiessupport with Research Institute of State Grid Zhejiang Electric Power CO.,LTD.
文摘The production of green hydrogen via alkaline water electrolysis necessitates porous composite membranes with high ionic conductivity and high bubble-point pressure.However,the mainstream preparation process of porous composite membranes involves many parameters,rendering this a complex high-dimensional optimization problem.Traditional trial-and-error experimentation is inefficient and often fails to explore the performance boundaries.In this study,an XGBoost-based machine learning model is developed and trained with laboratorycollected datasets,achieving satisfactory predictive performance.The model provides critical insights into the relationships between six manufacturing parameters and the two core performance parameters of the membrane.Subsequently,prediction based on coarse-grained grid and reverse search are performed on the model to identify optimal parameter regions,followed by manual refinement through feature analysis.This integrated approach ultimately identifies three high-performance composite membrane candidates,which are experimentally vali-dated.This work demonstrates a highly efficient and accurate machine learning-driven paradigm for the development of advanced porous composite membrane in alkaline water electrolysis.
