Batch distillation is a preferred separation technique for handling mixtures with widely varying compositions and containing different species. It provides the flexibility required to adjust the operating conditions best suited to achieve the desired degree of separation. Often, determining the best operating conditions for a given mixture requires the use of reliable computer simulation models in conjunction with accurate thermodynamic properties. BIP-Opt is a tool that uses batch column history data and the BDIST-SimOpt simulator to adjust the NRTL binary interaction parameters (BIPs), such that the simulated column dynamics are close to the actual column dynamics.
Most mixtures used in practice exhibit nonidealities due to interactions between the individual components. The interactions between components are best measured experimentally. However, accurate experimental data are difficult to obtain. Furthermore, even when reliable binary interaction pairs (BIPs) are used, the simulation results deviate from column operation data for certain composition ranges. For example, when studying an eight-component system containing THF (Tetrahydrofuran), Heptane and Methylene chloride, for most compositions the agreement was very good between the column dynamics predicted by BDIST-SimOpt and the column operation data. However, when the concentration of Heptane and Methylene Chloride was high, the column dynamics predicted by simulation significantly deviated from actual column operation. By changing the BIPs for one of the pairs, THF-Methylene Chloride, it was observed that the simulation of the column dynamics could be influenced to reduce the deviation.
Modern day batch control systems, which only implement the pre-determined operating policies, do not solve the problem completely. The best operating policies are typically determined based on human judgement and historical data. Model based operation using BDIST-SimOpt has been successfully implemented in solvent recovery systems. A reliable simulation model is crucial for model-based operation of batch distillation columns. One of the reasons for deviations between simulated and actual column is approximate thermodynamic properties. BIP-Opt provides the ability to fine-tune thermodynamic parameters so that the difference between simulated and actual column is within acceptable range. The fine tuned thermodynamic parameters can then be used with BDIST-SimOpt for optimizing operating policies implemented during operation.