当前基于高速膨胀的气体液化研究仍处于理论探索阶段,所用成核模型多为内部一致经典成核理论(ICCT),不过该理论应用于低温真实气体时偏差较大。针对理论推导过程的忽略的因素,作如下修正:采用逸度求解化学势差,考虑液滴曲率的表面张力修正优选Tolman模型,状态方程优选SRK、低温表面张力推荐LD公式。将模型修正前后的理论结果与水、烷烃的实验数据对比发现,低温时ICCT预测值偏高,最大偏差3~4个量级;而修正后偏差可降低1~2个数量级,预测偏差降至2个量级以内,特别地对烷烃类偏差均在1个量级以下。修正后的模型能获得较好的预测结果,可作为天然气高速膨胀液化研究的成核基准公式。The study of condensation for high-speed expansion gas is still on the stage of exploration, and the popular nucleation model may be Internally Consistent Classical Theory. However, it can’t gain satisfied results for real-gas in low temperature. As far as theory defect, the model was modified as follows: computing chemical potential difference with fugacity and using Tolman model to solve the relation between droplet curves and surface tension; as for state equation and surface tension in low temperature the paper recommends for SRK and LD formula. Comparing the modified model and ICCT model with the experiment data, we found that ICCT predictive value higher than experiment value about 3~4 order of magnitude while modified model can reduce deviation to 1~2 order of magnitude, especially for alkane the deviation under 1 order of magnitude. For modified model can gain more satisfied results, the paper recommends it as standard formula for the study of nucleation. At present, the research on gas liquefaction based on high-speed expansion is still in the stage of theoretical exploration. Most of the nucleation models used are internally consistent classical nucleation theory (ICCT), but the theory is applied to low-temperature real gas with large deviation. In order to neglect the theoretical derivation process, the following corrections are made: the Tolman model is adopted to solve the surface tension correction with fugacity, the SRK is the equation of state, and the LD formula is recommended for the low temperature surface tension. Comparing the theoretical results before and after the model modification with the experimental data of water and alkane, it is found that the predicted value of ICCT is high at low temperature, the maximum deviation is 3 ~ 4 orders of magnitude, while the corrected deviation can be reduced by 1 ~ 2 orders of magnitude, Within 2 orders of magnitude, especially for alkanes deviation of less than one order of magnitude. The modified model can get better prediction results and can be used as a nucleation standard formula for the research of natural gas high-speed expansion liquefaction. The study of condensation for high-speed expansion gas is still on the stage of exploration, and the popular nucleation model may be Internally Consistent Classical Theory. However, it can not gain satisfied results for real-gas in low temperature. As far as theory defect, the model was modified as follows: computing chemical potential difference with fugacity and using Tolman model to solve the relation between droplet curves and surface tension; as for state equation and surface tension in low temperature the paper recommends for SRK and LD formula. Comparing the modified model and ICCT model with the experiment data, we found that ICCT predictive value higher than experiment value about 3 ~ 4 order of magnitude while modified model can reduce deviation to 1 ~ 2 order of magnitude, especially for alkane the deviation under 1 order of magnitude. For modified model can gain more satisfied results, the paper recommends it as standard formula for the study of nucleation.