The prediction accuracy of existing models of the rolling force of a thick plate is always very low.To address this problem,a high-precision genetic algorithm-backpropagation network(GA-BP)model of deformation resista...The prediction accuracy of existing models of the rolling force of a thick plate is always very low.To address this problem,a high-precision genetic algorithm-backpropagation network(GA-BP)model of deformation resistance was built,and its integration with the traditional fitted model was further established.Then,a novel rolling force model was obtained by embedding the integration model of deformation resistance in the original model of rolling force.According to this research idea,the industrial data are normalized at first.On this basis,the interactions among the process parameters were disclosed through the variance analysis,and then described by various virtual factors.These factors are set as part of input parameters.Then,the optimal structure of the GA-BP model of deformation resistance was determined and an integration model of deformation resistance was built.Finally,a novel rolling force model is obtained by substituting the traditional fitted deformation resistance into the Sims model with the integration model of the deformation resistance.The results proves that the introduction of virtual factors can increase the hit rate of±5%from 75.8%to 78%and can reduce the root mean square error from 4.72%to 4.48%.Besides,it is found that the mean relative error of the traditional fitted deformation resistance is 0.142,while that of the modified deformation resistance is only 0.03.In addition,the mean relative error in the original rolling force model is 0.145,while that of the present model is only 0.03.展开更多
Machine learning has started to be used in engine research to optimize combustion and predict fuel spray characteristics.This paper presents the development of a machine learning model using a Genetic Algorithm-Backpr...Machine learning has started to be used in engine research to optimize combustion and predict fuel spray characteristics.This paper presents the development of a machine learning model using a Genetic Algorithm-Backpropagation(GA-BP)neural network to predict spray penetration.The GA-BP neural network was selected for its ability to optimize neural network weights and thresholds,thereby improving model convergence and avoiding local minima,which are common challenges in complex,non-linear problems such as spray prediction.The model was trained using experimental data from diesel injector spray tests,and its accuracy was evaluated through parametric sensitivity analysis,examining the influence of various input factors.A comparison between the machine learning model and the traditional empirical formulas of spray penetration revealed that the machine learning model achieved greater accuracy.In terms of the sensitivity to inputs,it is interesting to find that the cognition of machines is different from that of humans.When an input parameter does not have any functional relationship with other input parameters,the absence of this input parameter will lead to a significant decrease in the accuracy of the output result.The results demonstrate that the machine learning approach offers higher accuracy and better generalizability compared to traditional empirical methods.This study recommends the ways to get better results of penetration prediction with BP neural networks,which is efficient in training and utilizing Artificial Neural Networks(ANNs).展开更多
基金funded by the National Natural Science Foundation of China(Grant Nos.52274388,U1960105 and 52074187)the authors express gratitude to reviewers for precious suggestions.
文摘The prediction accuracy of existing models of the rolling force of a thick plate is always very low.To address this problem,a high-precision genetic algorithm-backpropagation network(GA-BP)model of deformation resistance was built,and its integration with the traditional fitted model was further established.Then,a novel rolling force model was obtained by embedding the integration model of deformation resistance in the original model of rolling force.According to this research idea,the industrial data are normalized at first.On this basis,the interactions among the process parameters were disclosed through the variance analysis,and then described by various virtual factors.These factors are set as part of input parameters.Then,the optimal structure of the GA-BP model of deformation resistance was determined and an integration model of deformation resistance was built.Finally,a novel rolling force model is obtained by substituting the traditional fitted deformation resistance into the Sims model with the integration model of the deformation resistance.The results proves that the introduction of virtual factors can increase the hit rate of±5%from 75.8%to 78%and can reduce the root mean square error from 4.72%to 4.48%.Besides,it is found that the mean relative error of the traditional fitted deformation resistance is 0.142,while that of the modified deformation resistance is only 0.03.In addition,the mean relative error in the original rolling force model is 0.145,while that of the present model is only 0.03.
基金supported by EPSRC(Engineering and Physical Sciences Research Council,United Kingdom)(Grant numbers:EP/W002299/1,EP/Y024605/1).
文摘Machine learning has started to be used in engine research to optimize combustion and predict fuel spray characteristics.This paper presents the development of a machine learning model using a Genetic Algorithm-Backpropagation(GA-BP)neural network to predict spray penetration.The GA-BP neural network was selected for its ability to optimize neural network weights and thresholds,thereby improving model convergence and avoiding local minima,which are common challenges in complex,non-linear problems such as spray prediction.The model was trained using experimental data from diesel injector spray tests,and its accuracy was evaluated through parametric sensitivity analysis,examining the influence of various input factors.A comparison between the machine learning model and the traditional empirical formulas of spray penetration revealed that the machine learning model achieved greater accuracy.In terms of the sensitivity to inputs,it is interesting to find that the cognition of machines is different from that of humans.When an input parameter does not have any functional relationship with other input parameters,the absence of this input parameter will lead to a significant decrease in the accuracy of the output result.The results demonstrate that the machine learning approach offers higher accuracy and better generalizability compared to traditional empirical methods.This study recommends the ways to get better results of penetration prediction with BP neural networks,which is efficient in training and utilizing Artificial Neural Networks(ANNs).
