基于模型融合的 PBX 等静压成型相对密度全过程智能预测

Given the difficulty in obtaining the number of samples and the complexity of the forming state in polymer bonded explosives (PBX) isostatic pressing forming, the idea of constructing an isostatic pressing prediction model through model fusion was proposed. Firstly, the forming state parameters of PBX grains were obtained using an isostatic pressing discrete element model. A PBX isostatic pressing staged forming state prediction model based on an improved Stacking model fusion algorithm was established through staged processing of the isostatic pressing forming process. Afterwards, the PBX isostatic pressing stage forming state prediction model was correlated for the boosting, holding and unloading stages, and a PBX isostatic pressing forming state prediction model for entire process was constructed. At the same time, the forming state prediction model for entire process was improved based on historical data error correction method. The constructed model was applied to predict and analyze the changing trend of relative density and relative density difference, fully verifying the reliability of using the Stacking model fusion algorithm to construct the PBX isostatic pressure entire process prediction model.