| A new model has been developed to predict the results of small scale coal liquefaction experiments. This model represents the first attempt to incorporate processing variables directly into the model calculations. The variables taken into account were: Temperature, gas phase pressure and composition, coal type, coal treatments, reactor type and size, solvent type and amount, mixing speed, heat-up time, and the available hydrogen. Incorporation was done not through changes in the kinetic parameters of the model, but through the effects which each processing variable has on the kinetics, thermodynamics, and mass transport within the liquefaction system. This same model is therefore applicable to liquefactions run in batch reactors at temperatures up to 500{dollar}spcirc{dollar}C, with hydrogen or inert atmospheres up to 4000 psi, with all types of coals, with both donor and non-donor solvents.; To make the model useable, it was built into an interactive coal liquefaction computer simulation program. The program asks the user for input regarding the processing conditions then carries out the kinetic, thermodynamic, and transport calculations, taking the conditions into account. This allows users unfamiliar with the inner workings of the model to still predict results of their experiments. The results, including pressure predictions and daf weight fractions of preasphaltenes, asphaltenes, oils, gases, and insoluble material, are printed to the screen and recorded in a file at user selected time points.; The parameters used within the program were determined through an optimization procedure which involved running simulations of actual experiments reported in the literature. A total of 503 data points from 214 independent experiments were used to optimize 27 model parameters. The data included liquefaction experiments run using 6 different coals, 7 different solvents, wide variations in gas pressure and heat-up length, temperatures from 300-480{dollar}spcirc{dollar}C, solvent to coal ratios from 1/1 to 8/1, and a range of coal particle sizes. A large number of simulations were run using the optimized parameters to show that the experimentally observed phenomena were reproduced by the model. |
