基于过程数据的空分系统参数估计

Parameters estimation of air separation system based on processed data

根据空分系统模型特点确定模拟计算中待估计参数,包含热力学参数以及设备参数。文献中通常采用常规浓度的实验气液平衡数据回归热力学参数,由于实验浓度达不到高纯度范围,热力学参数用于高纯度体系模拟时误差较大,故采用过程数据估计。考虑到待估参数较多,提出了通过分析参数全局灵敏度来判断关键参数,以减小参数估计的规模,估计参数由原来的10个减少到2个关键参数。基于过程数据采用最小二乘法对关键参数进行参数估计。将优化后的关键参数值应用于模拟计算,计算结果与5个实际工况测量值进行对比,可以看出模拟结果与实际工况吻合较好;将优化热力学参数代入热力学方程进行相平衡计算,并与实验气液平衡数据以及未优化的参数计算的结果进行比较,优化后参数的计算结果更加逼近于实验相平衡数据。

Thermodynamic parameters and equipment parameters are determined on the basis of characteristics of air separation system model.Thermodynamic parameters are regressed by experimental vapor-liquid equilibrium data.Since the concentration of the experimental data is far from the high purity of the process,the parameters need to be estimated based on processed data to minimize the simulation error.In order to reduce parameters estimation scale,parameters global sensitivity analysis is adopted to identify the key parameters.The number of the parameters finally reduced from 10 to 2.Two key parameters are estimated by least square method based on processed data.Apply the optimized parameters to simulate calculation.Compare the calculate result with the 5 actual operating modes,it＇s shown that the simulate results tally well to the measured data.Substitute the optimized thermodynamic parameter into thermodynamic equation to calculate phase equilibrium.Compare the result with experimental vapor-liquid equilibrium data and adopt un-optimized parameter in thermodynamic equation calculate result,it＇s shown that the thermodynamic model using optimized parameter can meet the experimental vapor-liquid equilibrium data better.