ML for High Entropy Alloys in AM

Machine learning accelerates composition design and process optimization for multi-principal element alloys in additive manufacturing

Topic Survey
Area Advanced Alloys + ML
Papers (250+ cit) 1,200+
Total Citations 650,000+
Typical Systems CoCrFeNi, AlCoCrFeNi, TiZrNbTa
AM Processes LPBF, DED, WAAM
Composition Space 10^6+ possible alloys

High entropy alloys (HEAs) contain 5+ principal elements in near-equiatomic ratios, creating vast composition spaces impossible to explore experimentally. Machine learning combined with additive manufacturing enables rapid discovery and production of novel HEAs with exceptional properties - high strength, corrosion resistance, and high-temperature performance.

Contents
  1. Overview
  2. ML for Composition Design
  3. Property Prediction
  4. AM Processing of HEAs
  5. Applications
  6. Key Papers

Overview

HEAs represent a paradigm shift from dilute alloys to concentrated solid solutions. The combination of ML and AM is transformative: ML navigates the astronomical composition space, while AM enables rapid fabrication of complex geometries with site-specific compositions.

1,200+
Capstone Papers
10^6+
Possible Compositions
>1 GPa
Typical Yield Strength
R2>0.90
ML Prediction Acc.

ML for Composition Design

Design Parameters

ML models predict phase formation and properties from composition-derived features:

Phase Prediction

Phase ML Method Key Features Accuracy
Single-phase FCC Random Forest VEC > 8, low delta 92%
Single-phase BCC SVM VEC < 6.87 89%
Dual phase Neural Network Intermediate VEC 85%
Intermetallic Gradient Boosting High delta, low omega 87%

Inverse Design

Given target properties, ML finds optimal compositions:

Property Prediction

Mechanical Properties

Property Best Model Input Features R2 Score
Yield strength XGBoost Composition, VEC, delta, process 0.91
Ultimate tensile strength Neural Network Composition + microstructure 0.89
Elongation Gradient Boosting Composition, phase fraction 0.82
Hardness Random Forest Composition, delta 0.93

High-Temperature Properties

Corrosion Resistance

AM Processing of HEAs

LPBF (Laser Powder Bed Fusion)

Most common AM process for HEAs:

DED (Directed Energy Deposition)

Process-Property Relationships

HEA System Optimal Energy Density Key Challenge ML Solution
CoCrFeMnNi 60-100 J/mm3 Mn evaporation Composition compensation
AlCoCrFeNi 80-120 J/mm3 Al segregation Scan strategy optimization
CoCrFeNiTi 70-110 J/mm3 Intermetallic precipitation Cooling rate control
TiZrNbTaMo 100-150 J/mm3 High melting point Preheat optimization

Applications

Aerospace

Nuclear

Tooling & Wear

Key Papers

High-entropy alloys: A critical review
Miracle & Senkov, Acta Materialia, 2017 | Foundational review | 5,500+ citations
Machine learning in materials science for HEA design
ML frameworks for composition prediction | 1,200+ citations
Additive manufacturing of high-entropy alloys: A review
AM processing of HEAs | 800+ citations
Data-driven design of HEAs with targeted properties
Active learning for alloy discovery | 500+ citations

View all HEA papers →