科研 Skill

相场与CALPHAD耦合建模

phase-field-calphad

Covers mesoscale microstructure evolution and CALPHAD thermodynamic phase-diagram modeling. Use for phase-field plus CALPHAD workflows.

发布者 Codar安全审核通过v0.1.0更新于 2026-11-18

SKILL.md

Phase-Field Modeling and CALPHAD

Overview

CALPHAD (CALculation of PHAse Diagrams) computes temperature- and composition-dependent phase equilibria from thermodynamic databases — the standard tool for multi-component phase diagrams and alloy design. Phase-field modeling simulates how microstructure (grain growth, phase transformation, precipitation) evolves over time, often using CALPHAD-derived thermodynamic data as input. Both sit at the mesoscale, between atomistic simulation (computational-materials) and continuum mechanics (finite-element-modeling).

When to use this skill

Activate when the request involves:

  • CALPHAD, phase diagram (temperature-dependent, not the 0K DFT hull in materials-structure-analysis), phase-field simulation, microstructure evolution, grain growth, precipitation, Thermo-Calc, pycalphad
  • Terms: Gibbs free energy, thermodynamic database (.tdb), order parameter, interface mobility
  • "Compute this alloy's phase diagram at temperature X," "simulate grain growth/precipitation"

Core usage

CALPHAD phase equilibrium calculation (pycalphad)

from pycalphad import Database, equilibrium, variables as v

db = Database("alloy_system.tdb")   # thermodynamic database — see pitfalls on database provenance
components = ["FE", "C", "VA"]     # VA = vacancy, a CALPHAD convention for sublattice models
phases = list(db.phases.keys())

eq = equilibrium(db, components, phases, {v.X("C"): 0.01, v.T: 1000, v.P: 101325, v.N: 1})
print(eq.GM.values)   # Gibbs energy of the equilibrium state

Binary phase diagram construction

import numpy as np
from pycalphad import binplot

binplot(
    db, components, phases,
    {v.X("C"): (0, 0.05, 0.001), v.T: (300, 1300, 10), v.P: 101325, v.N: 1},
)

Phase-field simulation (conceptual — MOOSE/similar is input-file/C++-driven, not primarily scripted in Python)

# A phase-field model tracks one or more order parameters (phase fraction,
# composition) evolving under a free-energy functional (often CALPHAD-derived)
# via the Allen-Cahn (non-conserved order parameter, e.g. grain boundaries) or
# Cahn-Hilliard (conserved, e.g. composition) equations. MOOSE's phase-field
# module handles this via input files defining the free energy, mobility, and
# interface parameters — see MOOSE documentation for the current input syntax,
# which is substantial and version-specific.

Validation & Pitfalls

Canonical references: Saunders & Miodownik, CALPHAD: Calculation of Phase Diagrams (1998), for CALPHAD methodology; Chen (2002), "Phase-field models for microstructure evolution," Annual Review of Materials Research, for phase-field modeling.

  • CALPHAD results are only as good as the thermodynamic database used, and databases vary in quality, scope, and the systems they were assessed for — a database extrapolated outside the composition/temperature range it was fitted to can give a plausible-looking but unreliable phase diagram; check the database's documented valid range before trusting results outside it.
  • Metastable phases and phases outside the database's assessed system are invisible to a CALPHAD calculation by construction — a CALPHAD phase diagram shows equilibrium among the phases included in the database, not necessarily every phase that could actually form (including kinetically stabilized ones); this is a different limitation than the 0K-only nature of the DFT-hull approach in materials-structure-analysis, but similarly requires stating what the calculation does and doesn't capture.
  • Phase-field simulation results depend heavily on interface energy/mobility parameters, which are often uncertain or fit rather than independently measured — treat quantitative kinetic predictions (transformation timescales) with more skepticism than qualitative morphology predictions unless the parameters were specifically validated against experimental kinetics for the system in question.
  • Grid resolution and interface width in phase-field simulations are numerical parameters, not physical ones, and must be checked for convergence — an under-resolved interface (too few grid points across the diffuse interface width) produces numerically unstable or spuriously fast dynamics; verify convergence with respect to grid resolution before trusting simulated kinetics.
  • CALPHAD equilibrium calculations assume full thermodynamic equilibrium is reached — real processes (rapid cooling/quenching, short-duration heat treatments) can leave a material in a non-equilibrium state a CALPHAD diagram doesn't directly predict; this is where the phase-field (kinetic) approach complements the CALPHAD (equilibrium) one, not a substitute for it.

示例提示词

我想把CALPHAD热力学数据库耦合进相场模拟来预测微观组织演化。

所属插件

材料表征、性能与材料信息学面向实验与数据驱动的材料研究:XRD/电镜/谱学/热分析/力学测试等表征,合成与加工,电化学与半导体表征,相场、CALPHAD 与连续介质力学,材料数据库、特征工程、贝叶斯优化与主动学习,选材与生命周期评价。

安装该插件会一并获得同领域的 42 个 Skill / MCP。

同插件的其他科研 Skill

常见问题

相场与CALPHAD耦合建模能做什么?

Covers mesoscale microstructure evolution and CALPHAD thermodynamic phase-diagram modeling. Use for phase-field plus CALPHAD workflows。

如何安装相场与CALPHAD耦合建模?

在科搭 Codar 左侧点击「插件」,在「组件」页签搜索「相场与CALPHAD耦合建模」并安装;在 Claude Code 或 Codex 中可以发送本页的安装提示词,或下载 Skill 包后解压到 ~/.claude/skills 或 ~/.codex/skills。

相场与CALPHAD耦合建模需要哪些权限?

该科研 Skill没有声明额外权限;执行具体操作时仍由 Agent 按自身规则请求确认。

相场与CALPHAD耦合建模经过官方审核吗?

是。SpeedSkills 收录的每个版本都会记录内容摘要与权限声明,当前版本已通过安全审核。