About Blend modul which is used to predict the Flory-Huggins parameter (chi)

In general, Blend modul is used to calculate the chi parameter as the input of DPD simulation. However, the accuracy of the results from Blend embedded in MS significantly depend on the forcefield. It is diffcult to obtain the good accordance with the experimental data. Recently, I used Blend modul to repeat the chi results of Reference [see accessory], but I have never succeeded. Who can help me to calculate the chi parameters?