OBJECTIVE To optimize the preparation of Xiakucao Kaiyin Granules(XKG) and control the quality of its intermediates.

METHODS The physical characteristics of spray dry powder of XKG were determined by powder evaluation method, and the physical fingerprint composed of 9 secondary physical quality indexes, such as bulk density, tap density, angle of repose and Hausner ratio, was established to determine the stability of the previous process and the quality consistency of spray dry powder.Taking the particle forming rate, dissolution rate, moisture absorption rate and angle of repose as evaluation indexes, dry granulation was carried out, and the auxiliary materials of XKG were screened.The mixture design experiment combined with G1-entropy weight method and neural network method were used to optimize the proportion of the selected excipients, and the best preparation technology of XKG was determined by comparing the two methods. The physical fingerprint of particles was established to evaluate the consistency of particle quality among different batches.

RESULTS The similarity of physical fingerprints of 9 batches of spray dry powder was greater than 0.970, and the physical properties were stable. The comprehensive score of the best proportion of auxiliary materials obtained through the analysis of mixture design was higher than that obtained by PSO-BP neural network modeling and optimization, so it was finally determined that the proportion of medicine and auxiliary materials of XKG was 7:3 and 29% maltodextrin and 71% lactose were added for dry granulation.The similarity of physical fingerprints of five batches of granule was greater than 0.994.

CONCLUSION The established physical fingerprint of intermediates can be used to control the quality process of XKG, and the optimized prescription of XKG can improve the physical properties of granules and improve the consistency of granule quality.