摘要
旋转填料床可高效强化气液吸收过程,气液传质系数的影响因素有转子尺寸、填料特性等结构参数和转子转速、气液流量、气液浓度、操作温度等操作参数。为准确预测气液传质系数,针对旋转填料床中NaOH溶液吸收CO2的过程,提出了其传质的无因次影响变量,然后基于最小二乘支持向量机(LS-SVM)的方法建立了旋转填料床吸收CO2气液传质性能预测模型。结合基于模拟退火算法的网格搜索法与10折交叉验证法获得模型参数最优值为:γ=3.833 4×104,σ2=0.717 6。对57组数据进行预测。结果表明预测相对误差为±15%以内,均方差为4.01×10-5,确定系数R2为0.984 2。模型预测精度较高,且泛化性能较好。
Rotating packed beds can efficiently intensify the gas-liquid mass transfer process. As one of the key design parameters,the mass transfer efficiency is hard to predict accurately due to the complex factors including structural parameters( rotor dimensions,packing characteristics,etc.) and operating parameters( rotational speed,gas and liquid flowrates,gas and liquid concentrations and operating temperature,etc.). In this work,dimensionless variables that affect the mass transfer coefficients for CO2absorption into NaOH solutions with rotating packed beds were introduced.Therefore,a model based on least squares support vector machine( LS-SVM) for prediction of gas-liquid mass transfer performance of CO2absorption by NaOH solutions in rotating packed beds was proposed. Optimal model parameters ofγ = 3. 833 4×104,σ2= 0. 717 6 were obtained by grid search method based on simulated annealing combined with 10-fold cross-validation method. The validity of the model was demonstrated by prediction of 57 sets of data. The model predictions are in agreement with experimental data within ±15%; mean square error is 4. 01×10-5and correlation coefficients( R2) is 0. 984 2. The present model shows a good accuracy and generalization.
出处
《煤炭学报》
EI
CAS
CSCD
北大核心
2013年第12期2253-2258,共6页
Journal of China Coal Society
基金
国家自然科学基金资助项目(50806049)
上海市自然科学基金资助项目(08ZR1415100)
国家留学基金资助项目(201208310168)
关键词
最小二乘支持向量机
旋转填料床
CO2吸收
传质系数
least squares support vector machine
rotating packed bed
CO2absorption
mass transfer coefficient