石油化工自动化
石油化工自動化
석유화공자동화
AUTOMATION IN PETRO-CHEMICAL INDUSTRY
2012年
6期
40-42,56
,共4页
轻汽油醚化BP神经网络LM算法预测控制
輕汽油醚化BP神經網絡LM算法預測控製
경기유미화BP신경망락LM산법예측공제
light gasoline etherification%BP neural network%LM algorithm%predictive control
针对催化裂化轻汽油(Fcc轻汽油)醚化的过程提出了BP神经网络的模型预测控制,通过控制Fcc轻汽油的流速,来实现重油量浓度指标的控制。应用BP神经网络建立该过程的预测模型,并采用迭代优化的控制算法,根据相应的性能指标,不断地修正神经网络的权值,从而整定下一批次的控制信号。通过Matlab里的神经网络工具箱,建立一个有参考模型的神经网络预测控制器,观测最终的实际输出。
針對催化裂化輕汽油(Fcc輕汽油)醚化的過程提齣瞭BP神經網絡的模型預測控製,通過控製Fcc輕汽油的流速,來實現重油量濃度指標的控製。應用BP神經網絡建立該過程的預測模型,併採用迭代優化的控製算法,根據相應的性能指標,不斷地脩正神經網絡的權值,從而整定下一批次的控製信號。通過Matlab裏的神經網絡工具箱,建立一箇有參攷模型的神經網絡預測控製器,觀測最終的實際輸齣。
침대최화열화경기유(Fcc경기유)미화적과정제출료BP신경망락적모형예측공제,통과공제Fcc경기유적류속,래실현중유량농도지표적공제。응용BP신경망락건립해과정적예측모형,병채용질대우화적공제산법,근거상응적성능지표,불단지수정신경망락적권치,종이정정하일비차적공제신호。통과Matlab리적신경망락공구상,건립일개유삼고모형적신경망락예측공제기,관측최종적실제수출。
Predictive control with BP Neural Network model for etherification of Fcc light petrol is proposed. The control of heavy fuel oil concentration is realized by controlling the flow rate of Fcc light petrol. The prediction model for the process is built up with BP Neural Network with adopting iterative optimization algorithm, and the weights of neural network is corrected continuously based on the performance indicators to determine the next batch controlling signal. Neural network predictive controller is built with a reference model by using Neural Network toolbox in matlab, and observes the actual output.