Core Thermal Properties in Casting Simulation
A deep dive into the physics, mathematics, and computational tools that transform raw thermal data into precision casting outcomes — from phase change fundamentals to defect prediction at scale.
The Physics of
Phase Change
At the heart of every casting simulation lies a deceptively complex physical reality: molten metal must journey through three distinct states — liquid, semi-solid (the mushy zone), and fully solid — each with radically different thermal behavior. Understanding and accurately representing this transition is the cornerstone of any high-fidelity casting model.
Three Distinct Material States
Every State Behaves Differently
Thermal Conductivity
Thermal conductivity determines the rate at which heat flows through a material. In cast metals, this value changes significantly between liquid and solid phases.
Liquid aluminum, for instance, conducts heat very differently than its solid counterpart, and capturing this variation is critical for accurate simulation outputs. Errors in conductivity assumptions propagate directly into temperature gradient predictions.
Conductivity Changes Across Solidification
Specific Heat & Density
Specific heat governs how much energy a unit mass of metal can store per degree of temperature change, while density defines the mass distribution across the casting volume.
Both properties are strongly temperature-dependent and must be supplied as full thermal curves rather than single-point values. Together, they define the volumetric heat capacity that determines how quickly or slowly a section heats and cools.
Two Properties Work Together
Use Full Thermal Curves
Specific heat and density evolve continuously as temperature changes. High-fidelity simulations therefore require complete temperature-dependent property curves rather than constant values.
The Stefan Condition
The Stefan condition governs the motion of the liquid-solid interface by enforcing an energy balance across the moving boundary.
Because the interface velocity depends on the thermal gradient on both sides of the front, the problem is inherently non-linear.
Energy Balance Controls Interface Velocity
Because the position of the interface changes while simultaneously altering the thermal field that drives its motion, the Stefan problem is fundamentally non-linear.
Solve Energy Using Enthalpy
Modern solvers track total enthalpy — including latent heat contribution — as the primary solution variable rather than temperature alone.
Extremely Fine Time Steps
The moving interface and latent heat release create stability constraints that often require very small temporal increments.
Industry Standard
Implicit enthalpy formulations efficiently accommodate latent heat and moving boundaries in production-grade casting software.
Not Temperature Alone
Phase-change simulation succeeds when conductivity, heat capacity, density, latent heat, and moving-boundary physics are treated as a coupled energy problem rather than an isolated temperature calculation.
From Properties to Solidification
Phase Change Is Not
a Temperature Problem.
It Is an Energy Problem.
High-fidelity casting simulation requires accurate representation of the three-state journey from liquid to mushy zone to solid. Thermal conductivity changes substantially between phases and governs heat-flow predictions, while temperature-dependent specific heat and density collectively define volumetric heat capacity and cooling response. The moving liquid-solid interface is governed by the Stefan condition, which enforces energy conservation across the phase boundary and introduces strong non-linearity into the problem. Because latent heat release is inseparable from interface motion, modern casting solvers employ implicit enthalpy formulations that track total energy rather than temperature alone. This approach provides the numerical robustness and physical fidelity necessary for production-grade solidification prediction.
Accurate solidification modeling goes far beyond knowing when a casting "freezes." Engineers must characterize the entire thermal history of the alloy—from first nucleation through complete solidity—and encode that behavior into material databases that the solver can query at every node, at every time-step.
Solidification curves map the fraction of solid as a function of temperature, defining the mushy zone that exists between liquidus and solidus temperatures. Shrinkage curves complement this by quantifying the volumetric contraction that occurs during phase change. Together, they allow the simulation engine to predict where and when feeding becomes insufficient—the root cause of porosity and shrinkage defects.
The latent heat of fusion represents the energy released as the metal transitions from liquid to solid at the phase boundary. It is not a temperature rise—it is pure energy released at roughly constant temperature—and it must be explicitly accounted for in the energy balance.
Modern casting simulation platforms integrate thermophysical databases covering over 230 casting alloys, spanning grey iron, ductile iron, compacted graphite iron, carbon and alloy steels, aluminum–silicon systems, copper alloys, and more.
Together, solidification curves, shrinkage curves, and latent heat data provide the material foundation for accurate solidification modeling—enabling the solver to predict feeding adequacy, porosity risk, and hot-spot formation at every node and time-step.
Defining the Solidification Behavior
Map the mushy zone, predict feeding failure.
Energy released at the phase boundary.
Real-world behavior, not handbook approximations.
liquid, mushy, solid—into chemistry-specific databases
that the solver can query at every node, at every time-step.
Modern casting simulation transforms CAD geometry into a computational mesh of millions of discrete points, then solves the governing heat transfer partial differential equations simultaneously across the entire domain. Accuracy depends on both the numerical method selected and the quality of the boundary conditions applied at every material interface.
From computational mesh to casting performance, accuracy depends on the complete model. Appropriate numerical methods, dynamically modeled boundary conditions, and capable simulation platforms work together to improve defect prediction and support better-informed foundry decisions.
Simulation: From Nodes to Reality
The ultimate promise of casting simulation is not merely to replicate what happens in the foundry — it is to predict what will go wrong before a single pound of metal is poured. Criteria functions and iterative virtual design have made this promise a reality, transforming casting engineering from an empirical craft into a quantitative discipline.
The Niyama criterion, defined as the ratio of the local thermal gradient G to the square root of the cooling rate, provides a dimensionless indicator of micro-porosity risk. Regions where Niyama falls below a threshold value are flagged as high-risk for shrinkage porosity — a defect caused by insufficient liquid feeding during the final stages of solidification.
The G/v criterion (gradient-to-velocity ratio) extends this analysis to predict grain structure and segregation behavior, allowing engineers to anticipate not just the presence of defects but their type and severity.
Together, these functions map the entire casting surface and interior with a spatial resolution that no physical inspection method can match at the design stage. Engineers can color-map Niyama values over the full 3D geometry, instantly identifying isolated hot spots, thin sections at risk of premature cutoff, and regions where riser placement or geometry changes will have the greatest impact.
The economic impact of simulation-driven design is well-documented. Foundries adopting iterative virtual design workflows routinely achieve 50% or greater reductions in riser volume and feeding metal, directly translating to yield improvements and energy savings.
Costly shop-floor trials — where prototype tooling is cut, poured, sectioned, and inspected — are eliminated or reduced to a single confirmatory trial rather than a multi-iteration discovery process.
This shift represents a fundamental transformation in the foundry's role. Rather than a craft operation guided by experience and intuition, the modern simulation-equipped foundry operates as a precision engineering laboratory. Every alloy composition, every mold design change, and every process parameter variation can be evaluated computationally before committing physical resources.
The result is faster time-to-production, higher first-pass quality, and the ability to tackle geometrically complex, safety-critical castings with confidence.
Typical yield improvement from simulation-driven riser optimization versus trial-and-error methods.
Alloys covered in modern thermophysical databases for iron, steel, and aluminum systems.
Virtual iteration reduces physical trials from many iterations down to a single validation pour.
Heat transfer coefficient can drop by a factor of ten when an air gap forms between casting and mold wall.
Simulation has turned every well-equipped foundry into a predictive engineering laboratory. By combining criteria functions, thermophysical databases, solidification modeling, and virtual iteration, engineers can identify defects before production begins, optimize feeding systems before tooling is manufactured, and validate designs before committing metal, labor, and energy. The era of zero-defect casting design through simulation-driven prediction is no longer a future vision — it is an operational reality available today.
The Future of Defect Prediction
Criteria Functions: Niyama & G/v
Virtual Iteration & Foundry Transformation
Simulation Performance Metrics
The Predictive Foundry
What's Your Reaction?
