The invention discloses a zero sample prediction and numerical
phase diagram generation method and
system for a
creep behavior of a
metal material, and belongs to the technical field of digital design and
performance prediction of materials. The method comprises the following steps: firstly, constructing a cross-material generality rule mapping relation
library, and establishing quantitative correlation among material macroscopic attributes,
crystal structures,
phase change temperatures, physical mechanism
weight distribution and
model parameters; for a to-be-predicted target
metal material, a preset model and parameters can be called from the relation
library only by inputting the
crystal structure and the
melting point of the to-be-predicted target
metal material, a high-precision physical mechanism
weighted network prediction model is automatically assembled, and accurate prediction of the
creep behavior of the to-be-predicted target metal material in the full stress-temperature domain is achieved. Through continuous calculation and
visualization, a quantitative and continuous
creep numerical
phase diagram with physical
interpretability is generated, and competition and evolution of different creep mechanisms are visually revealed. According to the method, normal form transformation from experience fitting to physical prediction is realized, and a core tool is provided for rapid evaluation and digital design of
new materials.