【摘要】 用饱和浸渍法把ＦｅＳＯ４ 直接担载于山东兖州和山西汾西两种烟煤上． 在实验的基础上，结合人工神经网络模型考察了ＦｅＳＯ４ 浸渍量、反应温度和反应时间对烟煤液化行为的影响，并通过ＸＲＤ和热力学计算探索了ＦｅＳＯ４ 在煤直接液化反应中可能发生的化学形态变化． 结果表明，训练完全的人工神经网络不仅可较好地拟合实验结果，而且可较好地预报反应条件对ＦｅＳＯ４ 催化活性的影响．ＦｅＳＯ４ 在煤液化反应时存在着临界浸渍量，当铁含量大于１－０ ％ 时，其值对液化结果影响不大； 煤直接液化反应存在着最佳反应温度，兖州煤为４１０ ℃左右，而汾西煤难以裂解，反应最佳温度为４３０ ℃； 在ＦｅＳＯ４ 催化条件下，兖州煤在４００ ℃反应４０ ｍｉｎ 以前即可达到较大程度的加氢液化，而汾西煤在４５０ ℃反应１ ｈ 以后才能达到较好的加氢效果．ＸＲＤ和热力学计算结果表明：ＦｅＳＯ４ 只有在煤中各种形态硫的作用下，才能转变为Ｆｅ１ － ｘＳ， 起到催化作用．更多还原

【Abstract】 Two bituminous coal samples, YZ from Yanzhou of Shandong, and FX from Fenxi of Shanxi in China, were subjected to direct liquefaction with FeSO 4 as the catalyst precursor, which is directly loaded on the coal. After being trained by experimental data, an artificial neural network (ANN) was used to simulate the liquefaction process and to evaluate the reaction variables on catalytic performance of FeSO 4. XRD analysis and thermodynamic calculation were also carried out to probe chemical change of FeSO 4 during the liquefaction. The results showed that the optimum catalyst loading for the coal samples is about 1%, the optimum reaction condition is θ =410 ℃ and t =40 min for YZ, and θ =430 ℃ and t =1 h for FX respectively, and the FeSO 4 can only be transformed into pyrrhotite Fe 1- x S in the presence of H 2 and H 2S which is generated by sulfur compound reacted with H 2 in the coal.更多还原