Blacksmith Forge Sparks: Hot Steel, Hammer Blows and Flying Scale

Heat a bar in the fire, thin it on the anvil with real hot flow-stress physics and watch hammer-scale flakes fly, cool and fade out below the Draper point.

Switch to 2D

Presets

Playback

Where is the bar?

Parameters

Camera

Layers

Quality

Model check

The science

What is modelled

A smith heats a steel bar to a bright orange-yellow in the fire, carries it to the anvil and thins it with hammer blows while it is still soft. Each blow squeezes the hot steel, throws off flakes of hammer scale (iron oxide that formed on the surface in the fire) and the bar keeps cooling. The page puts numbers on three linked processes: the heat balance of the bar, the plastic work of a blow, and the flight and glow of the flakes.

Equations

  • Bar heat balance (6 lumped segments of 40 mm): rho c(T) V dT/dt = -h A (T - T_env) - eps sigma (T^4 - Ta^4) A - h_b A_b (T - Ta), with eps = 0.8 for oxidised steel; the last term is conduction into the anvil. In the fire the environment is the gas temperature with an effective coefficient of 250 W/(m2 K), in the tub 3000 W/(m2 K).
  • Hot flow stress: sigma_f(T) = 60 MPa exp[(Q/R)(1/T - 1/1473)] with Q/R = 4340 K, so steel is about three times stiffer at 800 C than at 1200 C.
  • Blow: E = m v^2 / 2; the bar absorbs W = eta E = p A_c dh with the slab-upsetting pressure p = sigma_f (1 + mu w / 3h), mu = 0.3, and the volume is conserved (h w L = constant), so a thinner segment spreads wider.
  • Plastic heating: dT = 0.9 W / (rho c V). It is only a fraction of a kelvin per blow, so forging is a race against cooling, not a heating process.
  • Flake flight: m dv/dt = m g - (1/2) rho_air C_d(Re) A |v| v with the Schiller-Naumann drag law, integrated implicitly in the drag so tiny flakes stay stable.
  • Flake temperature: m c dT/dt = -A [h (T - Ta) + eps sigma (T^4 - Ta^4)], h from Nu = 2 + 0.6 Re^(1/2) Pr^(1/3). A flake is visible while it is hotter than the Draper point (798 K); colour comes from a blackbody fit.

Checks run in the page

  • Radiative cooling of a flake in vacuum against the exact T^-3 law.
  • Vacuum projectile range against v^2 sin(2 theta) / g.
  • Terminal speed of a flake with constant drag.
  • Bar cooling by convection against the lumped exponential.
  • Blender geometry, and the energy and volume bookkeeping of 200 blows.

Honest simplifications

Each segment is one lumped body (the Biot number is roughly 0.2 on the anvil, so the lumped assumption is only approximate there); there is no conduction between segments, no scale insulation, no Curie-point peak of the specific heat and no strain-rate or work-hardening dependence of the flow stress. The flow-stress constants, the 3.5 flakes per joule and the flake size distribution are illustrative, not measured. Real blows localise under the hammer face, whereas here a blow thins the whole 40 mm segment. The flakes are hammer scale cooling in air; they do not burn, unlike the sparks of a grinder or of forge-welding flux, and in reality they bounce and tumble (here they stop on the first surface they hit). The heat clock slider speeds up only the heat flow of the bar, not the hammer or the flakes. Educational model, not a forging process design.

What to try

  • Leave the bar on the anvil and watch the coldest segment drop below the 850 C line, then carry it to the fire.
  • Raise the fire to 1800 K: the bar overheats and the page warns that steel burns.
  • Compare 0.15 mm and 1.2 mm flakes in the third graph: small flakes cool faster and die sooner.
Loading model...
Loading Blender model...
Bar temperature and the forging window
Thickness of each bar segment
Flake glow time and range against size
Made with 🟠 Blender · Three.js