摘要
为实现随机工况下群车加油系统的仿真,基于改进的BP神经网络提出了一种加油枪流量仿真方法。首先对群车加油性能试验数据进行预处理,并构建样本训练集和测试集。随后,搭建了基于BP神经网络的群车加油系统仿真模型,在模型构建过程中,分析了隐含层神经元个数对模型结果的影响,确定了隐含层神经元个数。并对传统的BP神经网络进行改进,采用LM算法来解决模型收敛问题。最后分析模型预测结果,并与传统算法进行了比较。结果表明,传统BP神经网络的加油量仿真值与实际值平均相对误差为4.97%,改进的BP算法平均相对误差为0.93%,改进的BP神经网络预测精度更高。该方法解决了群车加油系统加油量的仿真问题,可为群车加油系统后续的智能控制提供数据依据。
In order to realize the simulation of the group vehicle refueling system under random working conditions,a fuel gun flow simulation method is proposed based on the improved BP neural network.Firstly,the refueling performance test data of the group vehicle are preprocessed,and the sample training set and test set are constructed.Secondly,a group vehicle refueling system simulation model based on BP neural network is built.During the process of model construction,the influence of the number of hid⁃den layer neurons on the model results is analyzed,and the number of hidden layer neurons is determined.The traditional BP neural network is improved,and the LM algorithm is used to solve the model convergence problem.Finally,the prediction results of the model are analyzed and compared with the traditional model.The results show that the relative error between the simulation value of the fuel quantity based on the traditional BP neural network and the actual value is 4.97%,and the relative error of the improved BP neural network based on the LM algorithm is 0.93%,which shows the improved BP neural network has a higher prediction accuracy.This method solves the simulation problem of the refueling volume of the group vehicle refueling system,and can provide data basis for the subsequent intelligent control of the group vehicle refueling system.
作者
郭杨
曾国栋
马振利
黄思宇
刘慧姝
方钢
GUO Yang;ZENG Guodong;MA Zhenli;HUANG Siyu;LIU Huishu;FANG Gang(Army Logistics Academy,Chongqing 400030)
出处
《舰船电子工程》
2024年第11期106-109,共4页
Ship Electronic Engineering
基金
国家自然科学基金项目“机动输油管线水顶油排空油水混合机理及流动特性研究”(编号:52272338)
重庆市教育委员会科学技术研究计划重大项目“机动输油管线水顶油排空水流携油机理研究”(编号:KJZD-M202212901)资助。
关键词
随机工况
改进的BP神经网络
仿真
智能控制
random working conditions
improved BP neural network
simulation
intelligent control
