石油炼制与化工
石油煉製與化工
석유련제여화공
PETROLEUM PROCESSING AND PETROCHEMICALS
2014年
7期
91-96
,共6页
张忠洋%李泽钦%李宇龙%李国庆
張忠洋%李澤欽%李宇龍%李國慶
장충양%리택흠%리우룡%리국경
催化裂化%反应-再生系统%神经网络%遗传算法%产率
催化裂化%反應-再生繫統%神經網絡%遺傳算法%產率
최화열화%반응-재생계통%신경망락%유전산법%산솔
FCC%reaction-regeneration system%neural network%genetic algorithm%yield
催化裂化反应-再生系统是一个高度非线性和强耦合的操作系统,用传统建模方法很难描述。鉴于人工神经网络(ANN)非线性预测和自学习自适应能力强,而遗传算法(GA)全局寻优能力强的特点,将两者结合,先通过 GA 寻得 BP 神经网络最优的权值和阈值初值,再赋予 BP,从而改善 BP 模型随机不确定选择初值的方法,提高其映射精度。以某炼油厂2.8 Mt/a MIP 装置反应-再生系统为研究对象,选取第一反应区温度、第二反应区温度、第一再生器温度、第二再生器温度、反应器压力、再生器压力等6个变量为神经网络的输入变量,汽油产率为输出变量,建立6-11-1的 BP 神经网络,并采用 GA 来对 BP 神经网络的权值和阈值进行优化。结果表明,未经 GA 优化时 BP 神经网络对催化裂化汽油产率的预测数据的均方误差为5.16,而经 GA 优化后预测数据的均方误差为4.92。
催化裂化反應-再生繫統是一箇高度非線性和彊耦閤的操作繫統,用傳統建模方法很難描述。鑒于人工神經網絡(ANN)非線性預測和自學習自適應能力彊,而遺傳算法(GA)全跼尋優能力彊的特點,將兩者結閤,先通過 GA 尋得 BP 神經網絡最優的權值和閾值初值,再賦予 BP,從而改善 BP 模型隨機不確定選擇初值的方法,提高其映射精度。以某煉油廠2.8 Mt/a MIP 裝置反應-再生繫統為研究對象,選取第一反應區溫度、第二反應區溫度、第一再生器溫度、第二再生器溫度、反應器壓力、再生器壓力等6箇變量為神經網絡的輸入變量,汽油產率為輸齣變量,建立6-11-1的 BP 神經網絡,併採用 GA 來對 BP 神經網絡的權值和閾值進行優化。結果錶明,未經 GA 優化時 BP 神經網絡對催化裂化汽油產率的預測數據的均方誤差為5.16,而經 GA 優化後預測數據的均方誤差為4.92。
최화열화반응-재생계통시일개고도비선성화강우합적조작계통,용전통건모방법흔난묘술。감우인공신경망락(ANN)비선성예측화자학습자괄응능력강,이유전산법(GA)전국심우능력강적특점,장량자결합,선통과 GA 심득 BP 신경망락최우적권치화역치초치,재부여 BP,종이개선 BP 모형수궤불학정선택초치적방법,제고기영사정도。이모련유엄2.8 Mt/a MIP 장치반응-재생계통위연구대상,선취제일반응구온도、제이반응구온도、제일재생기온도、제이재생기온도、반응기압력、재생기압력등6개변량위신경망락적수입변량,기유산솔위수출변량,건립6-11-1적 BP 신경망락,병채용 GA 래대 BP 신경망락적권치화역치진행우화。결과표명,미경 GA 우화시 BP 신경망락대최화열화기유산솔적예측수거적균방오차위5.16,이경 GA 우화후예측수거적균방오차위4.92。
The system of reaction and generation unit of RFCCU is a highly non-linear and strong coupled operation system and is too hard to be described by traditional model. The combination of the artificial neural network (ANN)with strong nonlinear prediction and self-learning ability and the genet-ic algorithm (GA)with global optimization ability provides a promising way to solve the problem. The optimal initial weights and threshold value are calculated by GA for the BP neural network firstly and feeded back to BP model to improve the method for random uncertain choice of initial value and the map-ping accuracy. In a practical application of this method for a 2.8 Mt/a MIP unit,a 6-11-1 type of BP neural network where the GA is used to optimize the weights and values of the BP network was estab-lished using the temperatures of two reaction zones and two regenerators along with the pressures of the reactor and regenerator as six input variables to predict the output variable gasoline yield. The results show that the predictive gasoline yield by BP neural network without GA has the mean squared error (MSE)of 5.16 while the one with GA optimization has the MSE of 4.92.