HIGH-PRECISION DUAL LAYER NEURAL NETWORK ALGORITHM USED FOR CALCULATING ENERGY OF ORGANIC MOLECULAR CRYSTAL STRUCTURE

摘要

A high-precision dual layer neural network algorithm used for calculating the energy of an organic molecular crystal structure, related to the technical field of organic molecular crystal structures. The invention comprises: performing a first round of regular crystal structure prediction; extracting from the existing crystal all molecular conformations and calculating the energy thereof; extracting all molecular dimers within the range of the van der Waals radius of the central unit cell, and calculating the interaction energy between molecules; performing single molecular conformation analysis and building a convolutional neural network of single molecular conformational energy; building a convolutional neural network corrected for the molecular dimer energy; calculating the total crystal energy. The method improves the precision of energy calculation during medication molecular crystal structure prediction while maintaining the calculation speed. The rapid and precise energy calculation guides CSP to rapidly find truly stable crystal forms on the correct potential energy surface.