CAJ | 학술논문

为准确识别数字化油库输转作业模式，在分析了油库典型输转作业模式基础上，提出了一种基于人工神经网络的输转作业模式识别方法。以阀门状态反馈信息作为网络输入，输转作业模式为输出构建立了神经网络结构，进而选取实际数据进行离线训练确定网络权值，最后通过仿真实验测试了该方法的识别效果，结果表明基于RBF神经网络的方法能快速、准确识别油库输转作业模式，对提高油库监控系统的智能化水平具有重要意义。
위준학식별수자화유고수전작업모식，재분석료유고전형수전작업모식기출상，제출료일충기우인공신경망락적수전작업모식식별방법。이벌문상태반궤신식작위망락수입，수전작업모식위수출구건립료신경망락결구，진이선취실제수거진행리선훈련학정망락권치，최후통과방진실험측시료해방법적식별효과，결과표명기우RBF신경망락적방법능쾌속、준학식별유고수전작업모식，대제고유고감공계통적지능화수평구유중요의의。
In order to recognize oil transportation pattern correctly in digital oil depot, on the basis of analyzing typical oil trans-portation pattern a method is presented based on artificial neural network. Firstly this algorithm uses feedback of valves as input sig-nal and uses oil transportation pattern as the network output to establish a neural network. Then real data are chose for training the neural network parameters offline to get network weight. Lastly the recognition effectiveness was tested by simulation. The simula-tion experimental results show that the method based on RBF neural network can recognize operation pattern accurately and quickly. Study results are significant for system intelligence improvement in oil depot.